EX PARTE ROBERSON III v. << (2025)

CONCURRING OPINION

Today, the Court stays Applicant Robert Leslie Roberson's execution and remands the case for further evidentiary development. I join the majority and write separately to explain why.

One year ago, Applicant's execution was scheduled for October 17, 2024. On October 9, 2024, this Court issued our opinion in Ex parte Roark, 707 S.W.3d 157 (Tex. Crim. App. 2024), which recognized significant legal issues when the State relied on “Shaken Baby Syndrome” in their prosecution of Roark. The issuance of our opinion in Roark just a week before Applicant's scheduled execution—in addition to the involvement of the Legislature's Texas Committee on Criminal Jurisprudence—resulted in series of motions, writs and mandamuses being filed that has seldom occurred at this Court. As a result, all three branches of the State government and both high courts became involved in this case.

Based on our decision in Roark, the Applicant filed one of several writs and motions in an attempt to stay the execution and enforce a subpoena issued by the Texas Committee on Criminal Jurisprudence. That subpoena ordered Applicant to appear before their committee at the Capitol on October 21, 2024—four days after the execution date. Obviously, that posed several legal and practical problems on many levels, specifically the execution date scheduled for October 17, 2024. Initially, Applicant obtained relief for the enforcement of that subpoena in a Travis County District Court, but just hours before the execution, this Court set aside that stay in a 5–4 decision. See In re Tex. Dept. Crim. Just. ex rel. Paxton, 707 S.W.3d 217, 220 (Tex. Crim. App. 2024) (Newell, J., dissenting) (arguing that the Court of Criminal Appeals should stay the execution to address the jurisdictional question raised after a Legislative subpoena conflicted with the authority of the Court.).

The Applicant then turned to the Texas Supreme Court for last minute relief of the execution. In perhaps a first for this State, the Texas Supreme Court stayed the execution noting:

The dispute before us has nothing to do with Roberson's guilt or innocence or with the propriety of his sentence, and we express no view on any of those criminal-law matters. Instead, we must decide the important but unresolved separation-of-powers question presented: whether the legislature's authority to compel testimony requires the other branches to yield in the face of a scheduled execution. That issue does not arise under the criminal law; it instead asks how the People of Texas have structured their government and to which governmental entities they have allocated specific kinds of authority. In other words, when two or more branches of government issue commands that cannot all be obeyed, which of them must yield?

Confronted with this novel separation-of-powers question, which could very well arise again if left unresolved, we temporarily enjoined impairment of Roberson's ability to comply with the subpoena. That order necessarily had the consequence of preventing the execution from proceeding as planned.

In re Tex. House of Rep., 702 S.W.3d 330, 334 (Tex. 2024).

Although their case held, “Legislative investigatory power, even at its maximum, is insufficient to forestall a long-scheduled execution,” they mentioned nothing about Applicant's claims in light of the Roark decision. Id. Thus, while Applicant's execution was not carried out, neither he nor the habeas court were given any authorized opportunity to further develop the evidentiary record in light of Roark.

As identified in Roark, our scientific understanding of what has become known as Shaken Baby Syndrome has significantly advanced. Because of this deeper understanding, certain assumptions and conclusions that were once thought to be true may not be. In this case, the State's theory of the case was inseparably intertwined with Shaken Baby Syndrome as it was medically understood more than 20 years ago. Shaken Baby Syndrome was referenced over 130 times in the State's case-in-chief and was discussed by both the State and defense. Although the State also presented evidence of blunt force trauma as the result of physical impact as an alternate or contributing cause of death, the record does not parse them so they can be evaluated separately. In fact, the record reflects, through the testimony of a medical expert, that “shaking” falls into the definition of blunt force. Thus, the term “blunt force” was presented in a way that permitted the jury to equate it to shaking. That can make a significant difference in a capital murder prosecution.

Accordingly, I agree that this Court's recent decision in Roark provides a new legal basis for Applicant to overcome the subsequent writ bar set out in Article 11.071, sec. 5. The facts Applicant alleges are at least minimally sufficient to bring him within the ambit of the new legal basis for relief that we announced in Roark. It is therefore correct to stay the execution and remand this case to the habeas court for Applicant to develop his claims in an evidentiary hearing.

This is not to say at the end of the day, Applicant is entitled to any relief. However, since the habeas court never had the benefit of our opinion in Roark, they are in a much better position on remand to develop a record on the validity of Applicant's claims in light of Roark, and the State's claims that Applicant is not entitled to any such relief.

There is a delicate balance and tension in our criminal justice system between the finality of judgment and its accuracy based on our ever-advancing scientific understanding. A death sentence is clearly final and, once carried out, hindsight is useless. Thus, when moving forward in such a way, we should require the highest standards of accuracy so that we can act with a reliable degree of certainty. That is not possible with the record we presently have before us. As noted by the United States Supreme Court:

[T]he penalty of death is qualitatively different from a sentence of imprisonment, however long. Death, in its finality, differs more from life imprisonment than a 100-year prison term differs from one of only a year or two. Because of that qualitative difference, there is a corresponding difference in the need for reliability in the determination that death is the appropriate punishment in a specific case.

Woodson v. North Carolina, 428 U.S. 280, 305, 96 S.Ct. 2978, 49 L.Ed.2d 944 (1976).

The Court is correct in remanding the case to develop the evidentiary record in light of the 20 years of medical scientific advancement acknowledged in Roark. To hold otherwise would risk a hasty judgment forever clouded by doubt. See e.g., Willingham v. State, 897 S.W.2d 351 (Tex. Crim. App. 1995). As often said of courts of last resort, “We are not final because we are infallible, but we are infallible only because we are final.” Brown v. Allen, 344 U.S. 443, 540, 73 S.Ct. 397, 97 L.Ed. 469 (1953) (Jackson, J., concurring). We owe it to ourselves to obtain a level of certainty to match the level of finality of whatever judgment this Court deems should come to pass.

The Court today has re-opened and remanded Applicant Robert Leslie Roberson, III's fifth application for a writ of habeas corpus “for resolution” of his Article 11.073 claim in light of Ex parte Roark, 707 S.W.3d 157 (Tex. Crim. App. 2024). The Court has dismissed his sixth and seventh application and denied his pending motions to reconsider. I concur in this decision to the extent it prevents what I see as Roberson's potentially unconstitutional execution in view of evidence undermining his conviction and sentence, infra at 11–12. I dissent to the remand, as this will likely only serve to delay the ultimate merits resolution to which both the State and Roberson are entitled.

While I would support further (and more prompt) proceedings in this Court, such as oral arguments in the matter before us, the merits of Roberson's claims and the cumulative effect of the evidence Roberson presents—in his fifth application as well as his previous and subsequent applications—would be more properly and more swiftly assessed at this point by a jury in a new trial. I believe the State is entitled to an opportunity to present this case on its merits to an actual jury without further delay or further speculation about what a jury might do when presented with all of (and only) the admissible, relevant evidence. My conclusion that Roberson has undermined confidence in the original judgment to the point that it can no longer support the weight of his execution simply pierces the judgment as a matter of law. Because that conclusion is necessarily premised on speculation about what a hypothetical jury might do if limited to proper evidence and given access to all of it, the law can only restore the presumption of innocence, leaving the parties to their right to a determination by an actual jury.

I largely pretermit the question of the application of the legislative writ bar and procedural barriers in connection with Roberson's applications arising under Articles 11.071 and 11.073, respectively, as I believe neither would (or could) prevent consideration of new evidence showing that a jury would no longer likely convict Roberson and sentence him to death. As I have detailed elsewhere, I believe the Texas Constitution necessarily controls that question and requires this Court, not the Legislature, to apply its operative provisions aimed at protecting against a potentially erroneous deprivation of life at the hands of the State itself. Ex parte Wood, No. WR-45,746-04, 2025 WL 2148862, at *11–16 (Tex. Crim. App. July 30, 2025) (Schenck, P.J., concurring in part and dissenting in part); Tex. Const. art. II. While the Legislature is free to add to those protections, it may not detract from them, suspend the writ, or deny access to this Court for that remedy. Id. art. I, §§ 12, 13. As a result, a new trial is necessary and mandated by our Constitution, irrespective of any statutory bar ¹ to the consideration of the writ.

As to dismissal of his seventh application, I believe Roberson's claims of judicial irregularities in the trial proceedings do not state a viable constitutional claim beyond the reach of the legislative writ bar and thus fail.

BACKGROUND

This is a difficult case made up of uncomfortable facts and a convoluted procedural history ² beginning in 2002 when Roberson brought his daughter, two-year-old Nikki Curtis, to the Palestine Regional Medical Center emergency room because she was not breathing. My colleagues have done an excellent job both of summarizing the salient facts and procedural history of this case and relating each to our past precedents and relevant statutory provisions. Their respective opinions, and the balance of the Court's work on this case, reflect precisely the level of attention and effort one would hope for in any case of this nature.³ As I have no material difference with the facts they set forth, I will move on to discuss the legal standards applicable to Applicant's claims and the reasons a remand is improper.

DISCUSSION

I. ROBERSON'S PROCESS CLAIM IS SUBJECT TO THE STATUTORY WRIT BAR

In his -07 application, Roberson alleges a due process violation with which he hopes to overcome subsection (2) of the legislative writ bar. I do not believe these allegations rise to the level of constitutional harm permitting them to proceed without regard to the statutory successive writ bar.

I have previously suggested potentially viable claims of actual innocence in a death case are ill-suited to resolution by procedural means,⁴ including defaults and bars to review of a successive application. Our Texas Constitution's due course of law clause and its prohibition on cruel and unusual punishment, in my view, prohibit the execution of a person who has produced new evidence demonstrating he likely would not be convicted of the underlying crime. Process complaints, such as the one lodged in Roberson's seventh successive writ, will not typically implicate these interests.

While judicial misconduct implicating a structural defect in the proceedings may be so well-established or so clear as to pierce the presumption of regularity and support collateral relief from a final judgment,—and perhaps regardless of procedural default—those circumstances are exceedingly rare. Our friends in the federal courts and in our own courts have identified certain extreme circumstances—all likely amounting to direct due process violations—in various other settings. In all of these cases, demonstrable judicial misconduct invited post-conviction correction despite finality interests that would otherwise preclude relief. Each case involved clear evidence directly impeaching the presumption of regularity on proof of conduct, such as criminality on the part of the judge in the form of habitually taking bribes, undisclosed romantic engagement with counsel for a party, or other misconduct so extreme as to bypass finality interests as means of upholding the integrity of the judiciary.⁵ I do not believe the alleged misconduct in Roberson's -07 application rises to the same level.⁶

A. A Freestanding Judicial Misconduct Allegation Must Pierce the Presumption of Regularity to Operate Free of Procedural Barriers

To be sure, a judge must recuse when an objective “probability of actual bias on the part of the judge or decisionmaker is too high to be constitutionally tolerable.” Withrow v. Larkin, 421 U.S. 35, 47 (1975). “[U]nder the Due Process Clause there is an impermissible risk of actual bias when a judge earlier had significant, personal involvement,”⁷ as, for example, where the judge previously served as a prosecutor in the case now before him on the bench. Williams v. Pennsylvania, 136 S. Ct. 1899, 1905 (2016).

Texas law similarly disqualifies judges from presiding over any criminal “case ․ where either of the parties may be connected with the judge, either by affinity or consanguinity, within such a degree as may be prescribed by law,” Tex. Code. Crim. Proc. art. 30.01, and requires recusal in “any proceeding in which ․ either of the parties may be related to the judge by affinity or consanguinity within the third degree,” including where such relative is the accused or the injured party, in addition to any other situation where the judge's impartiality might reasonably be questioned. Tex. R. Civ. P. 18b(a)(3); see Arnold v. State, 853 S.W.2d 543, 544 (Tex. Crim. App. 1993) (holding that procedure for recusal of judges set forth in Rule 18a of Texas Rules of Civil Procedure “applies to criminal cases absent ‘any explicit or implicit legislative intent indicating otherwise’ ”).

B. Roberson's Claims are Subject to the Writ Bar

Here, Roberson alleges Anderson County Judges Calhoon and Bentley violated his due process rights in being “at the very least, presumptively biased against him.” Appellant's Seventh Appl. for Writ of Habeas Corpus at 84. While I doubt the principles of Williams are limited to judges who acted as former prosecutors, none of the activities cited in Roberson's most recent application present evidence of a due process violation.

Roberson alleges the Anderson County judiciary approved of a custodial determination allowing Nikki's grandparents to authorize her withdrawal from life support. Specifically, he alleges Judge Calhoon was constitutionally and/or statutorily disqualified “when he stripped Roberson of his legal rights as a father,” and “by virtue of his son Mark Calhoon's role as Anderson County's Chief Felony Prosecutor.” Appellant's Seventh Appl. for Writ of Habeas Corpus at 87, 92. He provides no evidence of the alleged behind-the-scenes missteps by Judge Calhoon in removing Roberson as Nikki's legal guardian, nor any evidence his trial was tainted because Judge Calhoon appointed his attorney—a pre-trial activity—or because of his son's subsequent assignment as the State's trial counsel.

Judge Calhoon's role seems far removed from that involved in Williams—being the prosecutor who decided to seek the death penalty and later sitting as judge to review that penalty. Instead, from what we are shown, it would appear that Judge Calhoon, having appointed counsel to Applicant—as to whom Roberson has lodged no complaint as to effectiveness—simply stepped aside and had no further role.

Applicant also takes aim at the trial judge, Judge Bentley, alleging he should have known of the supposed misrepresentation of Nikki's legal guardianship allegedly conveyed to the hospital.⁸ However, Roberson admits that following Judge Bentley's death, it is impossible to prove exactly what the judge knew about the alleged communications of the hospital and Judge Calhoon.

Whatever else might be true of these claims, all of them are subject to the requirement of timely assertion and the successive writ bar. Either Roberson or his counsel, or both, were aware of the actions to terminate Nikki's life support, of the familial relationship between Judge Calhoon and the State's trial counsel, and of any statements made by Judge Bentley at trial or to the press long ago. The failure to pursue them with diligence or present them in any of Roberson's prior applications for habeas relief subjects them to foreclosure at this stage. Tex. Code Crim Proc. 11.071(5).

II. THE ACTUAL INNOCENCE CLAIM IN THIS CASE

In his -03 and -05 applications, Roberson asserts he is entitled to relief on grounds of actual innocence and under Article 11.073 as construed in our decision in Ex parte Roark, 707 S.W.3d 157 (Tex. Crim. App. 2024). As the dissents note, our Court's Roark decision discredited Shaken Baby Syndrome, but did not hold that shaking a small child could not cause injury to her, and did not involve any evidence of intentional blunt force trauma.

Likewise, Roark did not confront some of the statutory barriers to cognizability present here, including consideration of when, with reasonable diligence, new scientific evidence advancing Roberson's claim of innocence might have been marshaled and why it was not presented in his earlier writ applications. I pretermit that discussion for two reasons.

First, I largely agree with the dissenters that Roark is materially different in its facts, as there was no evidence of intentional blunt force trauma in that case to explain the injuries. Of course, that does not end the inquiry in this case. As detailed below, it begins it. It would leave us to answer the question of what a hypothetical jury would do in this case if that past causation evidence could fairly be untangled from the other evidence at trial and the other, new expert opinion evidence—all assuming Roberson could clear the hurdles of diligence and timeliness posed by the legislature in Articles 11.071 and 11.073.

Second, I believe the Texas Constitution's due course of law clause and prohibition on cruel and unusual punishment collectively forbid the execution of a person who has produced new evidence demonstrating a reasonable jury would likely not convict and impose that penalty. Ex parte Wood, 2025 WL 2148862, at *11–16. If that is so, legislation purporting to foreclose access to relief under the writ would violate the Constitution's suspension clause, open courts provision, and prohibition on cruel and unusual punishment. Id. at *6–8; see Tex. Const. art. I, §§ 12, 13.

As I noted in Wood, I do not read the Legislature to have even purported to foreclose resort to actual innocence under our Constitution, nor intimated that a litigant demonstrating actual innocence under it should be executed anyway on account of his failing to avail himself of an earlier opportunity to prove it. Ex parte Wood, 2025 WL 2148862, at *6. I will only reiterate here that the due course of law interest in life is explicitly distinct from the interests (and protections) afforded to liberty or property and, as a result, “there is a corresponding difference in the need for reliability in the determination that death is the appropriate punishment in a specific case.” Zant v. Stephens, 462 U.S. 862, 884–85 (1983) (emphasis added) (quoting Woodson v. North Carolina, 428 U.S. 280, 305 (1976)).

Roark involved a sentence of confinement. This one is death. And, “death,” as Justice Scalia observed, “is different,” and compels “protections that the [U.S.] Constitution nowhere else provides.” Harmelin v. Michigan, 501 U.S. 957, 994 (1991) (opinion of Scalia, J., joined by Rehnquist, C.J.). I view our Constitution as having a primary, not secondary, role in recognizing that distinction. See Ex parte Wood, 2025 WL 2148862, at *3–5.

A. The Claims Presented

To be sure, Roberson's claim of actual innocence is marketed differently from the facts of the case. That is to say, the argument he has somehow shown—by clear and convincing evidence ⁹ no less—that no juror could have found him guilty of any offense simply does not comport with the evidence of multiple injuries to Nikki, including bruising on multiple portions of her face and head.

I thus begin my review of the evidence by agreeing with my dissenting friends that the evidence adduced at trial was sufficient to establish to the satisfaction of a reasonable juror beyond a reasonable doubt that Roberson inflicted multiple physical injuries upon Nikki. No reasonable juror would be compelled to accept Applicant's explanation of a fall from a bed as the sole cause of her injuries. Likewise, Roberson's admission that he “lost it,” as adduced at the punishment phase of his trial, obviously travels with his new-evidence-attack on the original conviction.

Further, Roberson's new evidence is, much like the expert trial testimony, well-reasoned and probative of possible causes of the swelling of Nikki's brain in addition to,¹⁰ but not exclusive of, injury caused by shaking or blunt force impacts the jury might have inferred from the expert opinion testimony that Roberson inflicted. Samuel v. State, 477 S.W.2d 611 (Tex. Crim. App. 1972) (“[J]ury was warranted in concluding from the seriousness of the wound ․ appellant intended to kill.”). Yet he does not identify any single, alternative cause or constellation of causes for Nikki's death ¹¹ necessarily sufficient to cause death independent of the evidence of injury the jury could have found Nikki sustained as a result of Roberson's actions (i.e., shaking or striking her or some combination of both). See Tex. Penal Code § 6.04(a) (defendant remains culpable for results of criminal conduct unless a concurrent cause is “clearly sufficient” to produce the result and his own is “clearly insufficient”).

If we were obliged to consider all of the evidence available to evaluate Roberson's claim of actual innocence—and we are—we would have: (1) a basket of expert opinions on causation of Nikki's brain injuries, including the original opinions of Shaken Baby Syndrome and blunt force impacts that are all logically attributable to Roberson and (2) new opinions regarding other causes unattributable to Roberson—including those related to natural causes (pneumonia) and the toxic side effects of medication no longer advised for use in children of Nikki's age.

The first challenge for the Court, obviously, is in assessing the weight a hypothetical juror would give (or that the actual jury did give) to the evidence of Shaken Baby Syndrome as compared to blunt force—both of which were presented extensively to the jury by both sides. Because our holding in Roark undermines the evidence relative to Shaken Baby Syndrome, we must speculate as to what the jury did (or another jury would do) with the evidence presented of shaking, as compared to blunt force impact evidence. This challenge is then compounded by the expert opinion evidence of other possible causes for the brain swelling and death that the jury did not hear. In this posture, at least the target is set at a preponderance by the Legislature in Article 11.073, as opposed to the heightened burden of proof this Court set out in Elizondo.¹²

If the sole question in this case was causation and turned on the direct effect of the various expert opinion testimony—as augmented by the new evidence (and subtracting shaken baby opinion testimony)—I believe I would agree with Judge Keel, Judge Yeary, and Judge Parker's dissents. Any but-for causation attributable to Roberson in the mind of a hypothetically reasonable juror as an act of abuse would be at least potentially sufficient to sustain the conviction so long as the act would suffice to produce death on its own account. See Tex. Penal Code § 6.04(a). This would include any other possible explanation for Nikki's death, including the subsequent decision to terminate Nikki's life support highlighted in his process complaint. Felder v. State, 848 S.W.2d 85, 89–90 (Tex. Crim. App. 1992) (where victim's need of life support is attributable to defendant's actions, removal of it does not disrupt culpability).

B. Applicant's Conviction Hinges on the Necessary Mens Rea to Constitute the Offense and to Impose the Death Sentence

The evidence of Nikki's brain swelling did far more at trial, however, than simply demonstrate the actus reus and support a theory of at least concurrent causation of Nikki's death. While both the guilty act and causation were necessary to support any conviction, evidence of Roberson's culpable mental state is what made him eligible for any charge of murder,¹³ much less capital murder or the death penalty.

There was (and still is) no direct evidence Roberson intended to kill Nikki or had subjective awareness her death was likely to result from anything he did. Tex. Penal Code §§ 19.02(b)(1) (setting out scienter for murder as intentional or knowing infliction of death),19.03(a); see also id. § 6.03(b) (defining “knowingly” as aware[ness] that ․ conduct is reasonably certain to cause the result”). And, without any direct evidence of Roberson's mental state, the evidence of the extent of the swelling injuries to Nikki's brain and the available explanation of the causes of that swelling also carried the weight of providing proof—by means of inference from the force Roberson allegedly employed—that he at least knew her death was likely to result from his actions.

In Texas and elsewhere, where the State seeks to impose capital punishment, proof of either (1) the defendant's intent to kill or (2) the defendant's knowledge that death would likely result from the defendant's actions is essential. Tex. Penal Code § 19.03(a); Enmund v. Florida, 458 U.S. 782, 801 (1982). As discussed more fully below, the jury's assessment of Roberson's mental state is as much a part of the offense as any proof of any act or omission that would make his conduct a crime. See Tex. Penal Code § 6.02.

And, quite apart from the question of whether Roberson could be convicted of any crime based on any act or mental state relative to that act, this evidence was also integral to the jury's subsequent evaluation of whether Roberson deserved the death penalty. Tex. Code of Crim. Proc. art. 37.071 (directing jurors to determine “taking into consideration all of the evidence, including the circumstances of the offense” whether a sentence of life in prison or death is proper). In fact, the requirement of jurors to fully evaluate the defendant's moral culpability based on the circumstances of the case is incorporated into the Eighth Amendment of the U.S. Constitution. See Lockett v. Ohio, 438 U.S. 586, 605 (1978); Eddings v. Oklahoma, 455 U. S. 104, 111–125 (1982). It is not coincidental that the Penal Code's Chapter 6 defining the requisite mental states and requiring proof of same as part of the offense is entitled “culpability.”

The only evidence jurors had of Roberson's mental state was by derivative inference from the extent of Nikki's brain injury, which was said to have resulted from shaking, blunt force, or a fall from her bed. Jurors were free to reject the latter and accept any combination of the other two. The first two causes (shaking and blunt force) would both obviously support an inference as to Roberson's mental state; jurors could reasonably infer from the force of either he must have intended or had knowledge that Nikki's death would result. One of those two is now out of bounds, and the new evidence offers two (or three ¹⁴ ) further possible explanations for the same swelling that would not support any inference of the intent or knowledge necessary to support the conviction or the sentence of death.

Stated directly, there was and still is sufficient evidence to find a reasonable jury presented with all of the relevant evidence would more likely than not conclude Roberson “snapped” or “lost it” and inflicted physical injuries on Nikki. It may also still be likely a jury could credit some or all of the State's evidence on the cause of Nikki's brain swelling and further infer that Applicant striking Nikki was at least a but-for cause not only of her bruising, but of some or all of the swelling that led to her death. For purposes of causation, a menu of three possible causes—short fall, shaking and blunt force—would still compel a verdict of guilty unless the former was “clearly” insufficient and either or both of the latter were “clearly sufficient” to cause her death. Tex. Penal Code § 6.04. So, it may be likely that the jury would have attributed Nikki's brain swelling to the blunt force, found it “clearly sufficient,” and rejected the other two possible causes.¹⁵

But even those inferences as to the actus reus and causation would not support a conviction for capital murder and sentence of death. Jurors would need to further conclude beyond a reasonable doubt that the swelling and death were not evidence of the other available causes but were, in fact, evidence of the force Applicant used on Nikki and from which they might infer his intent to kill her or his knowledge that her death was reasonably certain to follow from his actions.¹⁶

That too is possible, but in my view unlikely. It is just as likely, if not more likely, that a jury would instead convict Applicant of any of the lesser-included felony offenses, including murder, or, far more likely, those more typically associated with the kind of “recklessness” or “negligence” associated with his having “snapped” or “los[t] it.” See Tex. Penal Code § 22.04 (felony injury to a child or elder). Indeed, these theories were submitted to the jury at Applicant's trial presumably for this reason, but without the present debate over Shaken Baby Syndrome or the other new expert opinion evidence said to account for the swelling that led to Nikki's death.

These changes would factor into a jury's sentencing decision as well. Even if a jury would likely still find some act by Roberson to be the cause in fact of Nikki's death and infer—despite the alternate evidence—that he intended or knew that result was likely, it would still be required to determine whether these inferences were clear and strong enough, in view of the alternate explanations, to warrant a sentence of death as opposed to life in prison. Cf. Sawyer v. Whitley, 112 S. Ct. 2514 (1992) (discussing eligibility for death penalty as an independent source of habeas inquiry).

C. Response to Concerns Over Role of Mental State

There may be concern among my colleagues as to an actual innocence claim under our state law and the relevance and role of a conviction's causation evidence to the jury's determination of guilt and its ultimate sentencing decision.

1. Guilt or Innocence is Crucial to Both the Statutory and Constitutional Inquiries—and Turns on the Definition of the Crime.

I would emphasize that our criminal jurisprudence treats an applicant as innocent of any offense for which the proof of his culpable mental state beyond a reasonable doubt is lacking. This was plain under the English common law we adopted with our first Constitution ¹⁷ and reflected in the ancient maxim actus non facit reum nisi mens sit rea (an act does not make a person guilty unless the mind is guilty). That understanding of the “offense” as constituting both the act and the requisite mental state has been carried forward into our current Penal Code. Its culpability chapter provides that “a person does not commit an offense unless ․ he acts [with the mental state] as the definition of the offense requires.” Tex. Penal Code § 6.02(a).

Moving forward to the offense here, our capital murder statute, Penal Code § 19.03, adopts the preceding section on ordinary murder, id. § 19.02, as the qualifying actus reus for pursuit of the death penalty, but is careful to borrow only its paragraph (b)(1) concerning the requisite mental state, which requires the defendant intentionally or knowingly caused the death of an individual. Section 19.02(b)(2), which is rejected in the elevation to capital murder, would permit conviction on proof the defendant intended to cause serious bodily injury and committed an act clearly dangerous to human life that caused the death of an individual. Id. Thus, Roberson's intent to cause harm or even his likelihood of causing serious bodily injury actually resulting in death was (and is) insufficient to support the offense for which he stands convicted and his punishment of death.

I am not aware of an affirmative holding from this Court that new evidence suggesting a habeas applicant did not act with the required mental state is not proof of innocence. On the contrary, this Court's decision in Ex parte Henderson aligns with the common law and the plain text of our Penal Code. See generally Ex parte Henderson, 384 S.W.3d 833 (Tex. Crim. App. 2012) (granting a new trial following a capital murder conviction where new evidence contradicted trial testimony that child's “injuries resulted from an intentional act of abuse or an accidental fall”). As Judge Cochran and others noted in that case: “The problem is that we do not know whether the jury would have found that applicant intentionally (as opposed to recklessly, negligently, or accidentally) caused Brandon's death absent Dr. Bayardo's expert scientific opinion.” Id. at 844 (Cochran, J., concurring); see also Wooten v. Norris, 578 F.3d 767, 781 (8th Cir. 2009) (entertaining, though ultimately rejecting, actual innocence claim based on lack of requisite mens rea).

An analogy may be useful to demonstrate several points here.

Suppose our habeas applicant, Mrs. Smith, was driving her car, accompanied by her unfaithful husband, when her vehicle jumps the curb and strikes a crowd of pedestrians in an outdoor mall, killing several and wounding others. Her husband, seeing a chance to secure a divorce on favorable terms, tells investigators and later testifies at trial that she was laughing maniacally and counting the victims at each impact, swerving to maximize the tally, as the horror unfolded. Mrs. Smith's attorney presents medical expert testimony at trial that she is, in fact, an epileptic and argues that while she was aware of her condition and her increasingly frequent blackouts, she likely passed out and did not intend to steer her vehicle into the crowd or to kill anyone.

That assertion, if true, would relate only to her mental state, not to the actus reus. Driving with knowledge of such a disability might constitute recklessness (a conscious disregard of “substantial and unjustifiable risk” of causing death) or negligence (actor “ought to be aware of a substantial and unjustifiable risk” that the “result will occur”) in the eyes of a juror adhering to our statutory definitions of culpability. E.g., State v. Skaggs, 185 Ohio App.3d 752 (2010) (employing our Mode Penal Code based definitions of culpable mental states).

Following Mrs. Smith's conviction and sentence to death, her former husband marries his former paramour, prompting Mrs. Smith to file her first habeas application pointing to his apparent infidelity and motive to mislead the jury. That application predictably fails. Shortly thereafter, her former husband confesses his perjury to his therapist and also comments on his ex-wife's having forgotten about the dash camera that would have likely captured proof that she was in fact unconscious at the time of the event. The therapist, having inadvertently recorded the conversation on his phone, ultimately decides to surrender the recording to the ex-wife's lawyer, who then obtains the dash camera video from the secured police impound facility where it has remained undisturbed under continuous recorded observation.

Mrs. Smith's decision to drive with awareness of her medical condition coupled with the act of causing injuries and death would still amount to a crime, albeit homicide or manslaughter. But suppose the crime for which she is convicted and sentenced is capital murder, defined as requiring the intent to kill, or, in the least, the knowledge of her conduct's causation of another's death. Tex. Penal Code § 6.02(a) (“a person does not commit an offense unless he intentionally, knowingly, recklessly, or with criminal negligence engages in conduct as the definition of the offense requires”) (emphasis added).

As with Roberson (and Henderson), the actus reus remains, as does some crime—namely of negligence or recklessness. But would our Constitution permit Mrs. Smith's the execution for that crime? And, would we reject the application—and permit the execution to proceed—because Mrs. Smith's dash camera video could have been located with diligence earlier and presented in the original habeas application? Or has her inadequate “bite” at the apple barred her access to this Court? I would think the answer to both questions is obvious. Even the federal courts, who generally prefer not to interfere in our domain, would likely correct us. See, e.g., House v Bell, 547 U.S. 518, 538 (2006) (actual innocence claim viable despite procedural default); McQuiggin v. Perkins, 569 U.S.383, 399 (2013) (actual innocence claim presented after limitations period); Schlup v. Delo, 513 U.S. 298, 323-324 (1995) (setting standard to permit review of state prisoner's claim of actual innocence despite successive nature of writ); United States v. Olano, 507 U.S. 725, 736 (1985) (citing actual innocence as most obvious form of miscarriage of justice that generally precludes procedural default objections).

While I am confident that neither our Penal Code nor our Constitution permit the death penalty for an offense anchored in negligence or any level of culpability apart from the intent to kill (or “knowingly” committing an action reasonably certain to kill), that does not end the statutory analysis applicable to Roberson's Article 11.071 and 11.073 claims. First, I am compelled to acknowledge a textual basis in the operative portion of Article 11.073 (“the person would not have been convicted”) that aligns with a view of “the offense” that would ignore the role of mens rea I detailed above. Likewise, the statute poses antecedent technical questions concerning the state of science ¹⁸ and Roberson's diligence in “discovering” it that may well align with a denial of relief under the statute.

For both reasons, I see development of the constitutional question as necessary to answer whether our Constitution requires us to treat Roberson as entitled to a new trial as to his guilt of capital murder. See infra at 35 (“yes”). And, while I again agree with my dissenting colleagues that our actual innocence doctrine may not yet expressly recognize the requirement of mental state as an integral part of the crime, my concern is not with the dissents, but our caselaw to this point.

2. The Jury's Sentencing Decision Requires the Ability to Act on Proof of Culpability for the Offense Itself

Of course, the evidence in Roberson's case and Mrs. Smith's hypothetical would go not only to the jury's decision to convict Roberson of capital murder, but to the separate, critical decision of his sentence. To be clear, I do not suggest that evidence tending to mitigate the accused's moral culpability apart from that arising from the circumstances of the offense itself could rise to a claim of “innocence” for purposes of the decision to apply the death penalty under our special issues inquiry in Article 37.071(2)(b). In fact, I might see the question differently if the evidence at issue did not simultaneously go directly to culpability. But, in this case it does, and leaves little room for doubting a jury must be able to act on it in deciding the appropriate sentence.

The U.S. Supreme Court has not yet spoken directly to us on this point. It has made clear to Ohio in Eddings that the federal Constitution requires it (and us) to permit jurors to hear and act on “any of the circumstances of the offense that the defendant proffers as a basis for a sentence less than death,” to Oklahoma that the sentence must be individualized as in Eddings; and to this State in Penry,¹⁹ that our former special issue framework was unconstitutional and failed to permit jurors to give effect to relevant mitigating evidence so clearly that it did not require recognition of a new rule in order to reverse us. The Supreme Court has also clearly stressed the “ ‘acute need’ for reliable decision making when the death penalty is at issue.” Deck v. Missouri, 544 U.S. 622, 632 (2005) (quoting Monge v. California, 524 U.S. 721, 732 (1998)). And, while the Supreme Court's Sawyer decision does not speak to this question precisely, it made clear that the federal writ will reach to those facts making one ineligible for pursuit of the death penalty in the first instance.

If, in the face of all of this, were we now to adopt a rule treating proof of lack of the requisite mens rea of capital murder or even murder as something other than actual innocence under our writ—thus permitting the execution of Mrs. Smith (or Mr. Roberson) on account of their being potentially guilty of negligent homicide or recklessness amounting to manslaughter—the U.S. Supreme Court might then speak to us on the scope of the federal Constitution as to the guilt/innocence or sentencing questions.

I simply do not see why that would be necessary. As I noted in Wood, we have our own Constitution and our own place under it. As I read it, that document is sufficient to pretermit all of these questions. I will therefore move on to summarize why it answers our challenge in this and other like cases and hope to leave the U.S. Supreme Court to develop federal constitutional law for some other state less committed to the organic constitutional design. See U.S. Const. amend. X.

III. WE SHOULD ANCHOR OUR ACTUAL INNOCENCE JURISPRUDENCE WHERE IT BELONGS

More than three decades ago this Court found that the execution of a habeas applicant who had demonstrated actual innocence “would surely constitute a violation of [the federal] constitutional due process right.” Ex parte Holmes, 885 S.W.2d 389, 397 (1994). We observed, “[w]ith this sound ․ principle of jurisprudence we cannot disagree.” Id. I will go further and note that the recognition of such rights in that provision of the federal constitution necessarily equates to the conclusion that the right is so obvious as to be “implicit in the concept of ordered liberty” and “so rooted in the traditions and conscience of our people as to be ranked as fundamental.” Palko v. Connecticut, 302 U.S. 319 (1937).

This raises two questions. First, why is it that the U.S. Constitution, cabined as it is by federalism constraints, would compel this result without a holding (if not decision) from the U.S. Supreme Court to that effect, without an inquiry into the text of our own Constitution?²⁰ Second, if this conclusion is so obvious and fundamental as a matter of the federal constitution to be (in our view anyway) applicable to all states, wouldn't it likely have occurred to the framers of our own? Stated another way, how is it possible that our own Constitution does not touch on the question when it, unlike its federal counterpart, contains: (1) two due course of law clauses, one of which commands that courts be open to provide remedies to “any person” for “injury done”; (2) demands an effective and speedy writ of habeas corpus free from legislative or executive suspension; and (3) provides a protection from prohibition on cruel or (not “and”) unusual punishment?

If the prospect of executing a citizen despite post-evidence proof of innocence is so obviously abhorrent that judicial recourse to prevent it is ranked as fundamental across all fifty states, it should be obvious our framers would have understood this when they crafted these constitutional provisions with greater substantive and procedural reach. Putting aside the myriad of other due course of law or open courts concerns with the “when or whether” questions pertaining to how an applicant might present his proof of innocence, it seems obvious to me that our framers would have regarded the execution of a person innocent of the offense for which they were sentenced to be disproportionate to the point of being cruel. And, because I reject the notion that the states (and particularly this one) are forgotten appendages of the federal government, I see the need to answer that question as a matter of state constitutional law as just as obvious.

The only question is the standard by which the applicant should have to show his innocence so as to be entitled to relief.

I will pose the following question to my colleagues in dissent and concurrence alike: a state's rights-oriented citizen of ours is petitioning this Court for relief from a sentence of death under our Constitution (waiving any like federal claim) on account of post-judgment evidence of innocence. Is he entitled to be heard and obtain relief on:

a) Conclusive and certain proof of his innocence to the point instructed verdict would be appropriate if the case were being tried to a jury at common law 21 and we were sitting as a trial court?;

b) Clear and convincing evidence that a jury would acquit him based on his newly discovered evidence?;

c) Proof to a preponderance, that is, evidence showing it is more likely than not a jury would not convict or sentence him to death?; or

d) None of above?

Despite the implications of all of our past decisions, at present our answer appears to be “d,” none of the above. And, while we have held as a matter of federal constitutional law that “b” would suffice, that answer is derived from a series of U.S. Supreme Court decisions that do not so hold and are laden with the heavy deference federal law poses on interference with our own role under our state constitution and laws.

And, with that in mind, I will then ask how we can possibly explain that even a showing of conclusive proof of innocence under “a” would not suffice. Despite our applicant's love for Texas and his certain innocence, he is doomed unless he bends the knee toward Washington, D.C. Why are we relying on a federal constitutional standard the U.S. Supreme Court has repeatedly refused to recognize? Why are we not applying our own Constitution and resolving close and difficult cases involving the heightened interest it states in life pursuant to its provisions? As the only Texan to serve on the Supreme Court observed more than six decades ago, “[n]othing can destroy a government more quickly than its failure to observe its own laws, or worse, its disregard of the charter of its own existence.” Mapp v. Ohio, 367 U.S. 643 (1961) (Op. of T. Clark, J.).

As I said in Wood, I would answer the question under our own Constitution and settle on a preponderance standard, leaving the State free to retry cases plagued with the serious doubt cast by new evidence and removing the risk inherent in the act of proceeding to execute a person without an actual testing of the evidence before a fact finder actually empowered to settle the question.

IV. A REMAND IS INAPPROPRIATE AND DOES NOT FURTHER THE INTEREST OF JUSTICE

This Court has now remanded Roberson's -05 application to the trial court “for resolution” of his claims involving Article 11.073 and Ex parte Roark. The burden here is on Applicant, who he has filed no fewer than seven successive habeas applications describing his factual argument in detail. All that remains are legal questions over the appropriate standard, statutory or constitutional, and whether he has met them. This Court not only has the acuity to parse those matters on its own, but the authority to order a retrial where necessary and appropriate. While I think the Court (and the parties) may benefit from hearing further from counsel, I do not believe the delay associated with a remand is proper.

Between the abundance of material Roberson has submitted as evidence with each of his successive applications and the number of attempts made within this and other opinions to determine what—if any—impact it would have on a jury weighing it against the evidence presented at trial, there is no better place for that determination to be made than before actual jurors. Sending these legal questions to the trial court for their eventual return here serves only to delay a merits resolution and deprive the State of its right to carry its burden of persuasion before an actual jury, which might—despite my prognostication—find against Roberson. But, Roberson's claims of new evidence have lingered long enough and deserve an answer by an actual factfinder empowered to resolve them. Tex. Code Crim Proc. 35.11 (permitting state or defendant to demand jury).

CONCLUSION

While Roberson's new evidence does not clearly establish his innocence of any crime, it does undermine what scant proof of mental state the State offered to support its theory Roberson intended to kill Nikki, or acted knowing her death was likely to result from his actions. Because that proof was essential to support his conviction for capital murder and to the jury's sentencing decision, and because the evidence before us persuades me that a rational jury would not likely find that intent beyond reasonable doubt, I would order a new trial.

Because the majority does not, I dissent. To the extent the majority's decision to remand this case prevents Applicant from being executed when the State, at trial, may not have proven beyond a reasonable doubt he maintained the mens rea necessary for a capital sentence, I concur. However, this Court should take initiative to order a new trial so this, and other issues, may be properly determined by a jury.

Like Judge Parker, I concur in the Court's order insofar as it appropriately dismisses Applicant's last two subsequent writ applications (-06 and -07) and refuses to entertain suggestions to reconsider two of his previous subsequent writ applications (-03 and -04). I dissent, however, to the Court's order to the extent that it reopens Applicant's subsequent writ application in Cause Number WR-63,081-05 and grants his motion for stay of execution.

To reopen a subsequent writ application that has already been disposed of—if it is ever appropriate—may be justified only under the most extraordinary of circumstances. Ex parte Moreno, 245 S.W.3d 419, 429 (Tex. Crim. App. 2008). That a reasonable, good faith argument can be made that our prior decision to dismiss the -05 subsequent writ application was, in hindsight, “incorrect” does not count as an appropriate basis to entertain a suggest for reconsideration. Id. In any event, the Court offers no such argument today. Once again, the Court is essentially ruling by fiat. See Ex parte Wood, No. WR-45,746-04, ––– S.W.3d ––––, 2025 WL 2148862, at *21 (Tex. Crim. App. July 30, 2025) (Yeary, J., dissenting) (complaining of the Court “exercising its brute ability (if not authority) to order a remand [in a subsequent capital writ application] notwithstanding the strictures of Section 5 [of Article 11.071, Texas Code of Criminal Procedure]”).

I also take this occasion to reiterate my view that, for the reasons I gave recently in Ex parte Horvath, 2025 WL 1699335 (Tex. Crim. App. June 18, 2025) (not designated for publication), to the extent that Ex parte Roark, 707 S.W.3d 157 (Tex. Crim. App. 2024), has been read to suggest that Shaken Baby Syndrome (SBS) is necessarily junk science or that shaking alone cannot cause a serious brain injury to a child, it should be disavowed.

In Ex parte Roark, a majority of this Court granted an applicant habeas relief under Article 11.073 of the Texas Code of Criminal Procedure. 707 S.W.3d at 188. Throughout the Court's opinion in Roark, the majority seemed to suggest that shaking alone might be insufficient to cause serious injury or death to a child. Id. at 185−88. It suggested that, “as early as 2004, the Journal of Neurosurgery published an article stating [that] the terminology of ‘shaking’ should be avoided.” Id. at 185. It cited a 2015 study suggesting that shaking alone, without an associated impact, could not create a subdural hematoma. Id. It mentioned research related to mechanical dolls and animal abuse that “has yet to bridge the gap between theory and reproduceable results which the scientific method demands.” Id. “Essentially,” the majority wrote, “science has evolved to a degree that has removed ‘Shaken’ from ‘Shaken Baby Syndrome.’ ” Id. Why? Because, the Court concluded, there was a “need to vague the terms to ‘Impact Syndrome’ and then to ‘Abusive Head Trauma’ [(AHT)].” Id.

But arguably credible and reliable scientific evidence still exists to suggest that shaking a child can cause serious injury or death. The majority in Roark suggested that a shift away from the use of the term “shaking” began in 2004. Id. at 185. But even if there is any truth at all to that claim, the shift is still nowhere near to constituting a scientific consensus.

Indeed, in 2016, a study published in the Journal of Pediatrics showed that “shaking with or without impact (in contradistinction to several other alternative theories) is [still] generally accepted to be a dangerous form of child physical abuse and capable of producing [subdural hematoma], [retinal hemorrhages], and coma or death.” Sandeep K. Narang et al., Acceptance of Shaken Baby Syndrome and Abusive Head Trauma as Medical Diagnoses, 177 J. PEDIATRICS 273, 277 (2016), https://www.jpeds.com/action/showPdf?pii=S0022-3476% 2816% 2930402-4. There, an observational study surveyed 628 multidisciplinary physicians from the ten leading children's hospitals within the specialty departments most involved in suspected AHT cases, as well as from coroners’ offices and medical examiners’ offices. Of those 628, 607 characterized SBS or AHT as a valid diagnosis. The results “demonstrate[d] that physicians, irrespective of specialty, viewed the risks of shaking, with or without impact, to be similar to a high-velocity [motor vehicle collision] and dissimilar to a very short fall.” Id. And this study was published after the 2015 study that the Roark majority cited for the proposition that “shaking alone, without an associated impact, could not create a subdural hematoma.”

Moreover, the biomechanical evidence that the Court discussed in Roark, which had relied on research related to mechanical dolls and animal abuse to assert that shaking alone cannot cause death, is insufficient to show a scientific consensus. As recently as 2020, the American Academy of Pediatrics (AAP) issued a “Policy Statement,” stating that “[a]ll biomechanical models, whether physical or computer generated, have limitations and fall short of a precise representation of the complex pathophysiology of the human infant.” Sandeep K. Narang et al., Abusive Head Trauma in Infants and Children, 145 PEDIATRICS 1, 4 (2020) (Policy Statement of the Am. Acad. of Pediatrics), https://publications.aap.org/pediatrics/article/145/4/e20200203/36936/Abusive-Head-Trauma-in-Infants-and-Children. And as recently as March of 2025, the AAP issued a “Technical Report,” stating: “Findings from biomechanical studies investigating shaking are mixed as to whether shaking alone can cause severe brain injury or fatality in infants.” Sandeep K. Narang et al., Abusive Head Trauma in Infants and Children: Technical Report, 155 PEDIATRICS 1, 20 (2025) (Technical Report of the Am. Acad. of Pediatrics) (emphasis added), https://publications.aap.org/pediatrics/article/155/3/e2024070457/201049/Abusive-Head-Trauma-in-Infants-and-Children. So, while biomechanical studies may not have “bridge[d] the gap between theory and reproduceable results[,]” Roark, 707 S.W.3d at 185, they certainly have not established a consensus that shaking a child cannot cause serious injury or death.

Perhaps the most flawed argument suggested by the Roark majority, though, was the notion that the change from the term SBS to AHT served as proof that the use of the term “shaking” was no longer scientifically sound. Roark, 707 S.W.3d at 185. At an evidentiary hearing in Michigan, Dr. Cindy Christian—a past chair of the American Academy of Pediatrics Child Abuse and Neglect Committee—explained that the Academy advocated for that change in order to fully capture all forms of abuse, not to remove “shaken” from “Shaken Baby Syndrome[,]” as the Roark majority suggested. People v. Lemons, 514 Mich. 485, 553, 22 N.W.3d 42, 81 (2024) (Zahra, J., dissenting). As Justice Zahra of the Supreme Court of Michigan aptly noted: “Plainly, this [change] did not mean that violently shaking a baby is not a serious health risk that can potentially cause death and severe injury in small children.” Id. All of this credibly suggests that babies, though often quite resilient, can be abused, and injured, and even killed, in a multitude of ways—including by shaking—and the change from SBS to AHT actually reflects just that.

It is far from clear that the scientific evidence relating to SBS or AHT has changed to reflect a consensus that shaking alone cannot cause serious injury or death to a child. To the extent that Roark might have suggested as much, I conclude that it erred. Credible and reliable scientific evidence suggests that shaking a child can indeed cause either serious bodily injury or death.

Moreover, this case has never been one in which the State relied for Applicant's conviction, exclusively or even principally, on the scientific theory underlying SBS. Instead, the State's predominant theory was that Applicant's young daughter was killed by the infliction of blunt force trauma. It was the defense who injected SBS as a theory of what happened in this case. So, the claim that Applicant's conviction turns on the scientific validity of SBS is manifestly false.¹

With these additional observations, I dissent to the Court's order insofar as it reopens Applicant's -05 writ application and stays his execution.

CONCURRING & DISSENTING OPINION

Before the Court today are Applicant's suggestions to reconsider the Court's denial of Applicant's -03 habeas application and dismissal of Applicant's -04 and -05 habeas applications, as well as Applicant's newly filed -06 and -07 habeas applications.

The Court is correct to dismiss Applicant's first and second claims in his -07 habeas application because they fail to satisfy the requirements of Code of Criminal Procedure Article 11.071, Section 5. I also agree with the Court's decision to deny Applicant's suggestions to reconsider his -03 and -04 habeas applications. Lastly, I agree with the Court's ultimate decision to grant Applicant's suggestion to reconsider on his -05 habeas application, reopen the habeas application, and remand for evidentiary development in light of Ex parte Roark, 707 S.W.3d 157 (Tex. Crim. App. 2024). My disagreement with the Court is rather limited: Unlike the Court, I would also remand Applicant's actual innocence claims in his -06 and -07 habeas applications. I see little distinction between the bases for Applicant's Article 11.073 claims and Applicant's actual innocence claims. Consequently, I join the Court's remand order but dissent insofar as the Court declines to remand Applicant's actual innocence claims.

I. Background

a. Procedural History

Applicant was convicted of capital murder in Anderson County in February 2003 for the murder of his two-year-old daughter, Nikki Curtis. Roberson v. State, No. AP-74,671, 2002 WL 34217382, at *1 (Tex. Crim. App. June 20, 2007) (not designated for publication). On direct appeal, this Court affirmed. Id. at *12. We also denied relief on Applicant's initial post-conviction habeas application and dismissed as a subsequent application a pro se document titled “Notice of Desire to Raise Additional Habeas Corpus Claims.” Ex parte Roberson, No. WR-63,081-01, 2009 WL 2959738, at *1 (Tex. Crim. App. Sept. 16, 2009).

After Applicant's federal habeas proceedings ended adversely to him,¹ the trial court set an execution date of June 21, 2016. On June 8, 2016, Applicant filed his second subsequent habeas application (the instant -03 writ) in the convicting court. He raised four claims relating to changes in the science underlying Shaken Baby Syndrome (SBS). After a remand, we denied relief. Ex parte Roberson, No. WR-63,081-03, 2023 WL 151908, at *1 (Tex. Crim. App. Jan. 11, 2023). A new execution date was set for October 17, 2024.

On April 24, 2024, Applicant filed in this Court a “Suggestion to Reconsider on Court's Own Initiative and Motion to Hold for Adjudication of Ex Parte Roark.” He asserted three reasons for reconsidering his -03 writ application: (1) this Court had filed and set Roark, which raised an Article 11.073 claim related to SBS; (2) Texas law (the amendment of Family Code Section 261.3017) “now recognize[d] the danger of miscarriages of justice when caregivers can be accused of abuse based on the hastily formulated opinions of ‘child abuse specialists’ ”; and (3) the habeas court's FFCLs were “wholly unreliable.” The Court denied the Suggestion on September 4, 2024.

On August 1, 2024, Applicant filed his third subsequent habeas application (the instant -04 writ) in the convicting court, raising five claims. These claims again related primarily to SBS. We dismissed the application on September 11, 2024. Ex parte Roberson, No. WR-63,081-04, 2024 WL 4143552, at *2 (Tex. Crim. App. Sept. 11, 2024).

On October 9, 2024, eight days before Applicant's execution date, this Court decided Ex parte Roark, No. WR-56,380-03, 707 S.W.3d 157 (Tex. Crim. App. Oct. 9, 2024). We recognized that the changes in the scientific knowledge underlying SBS would have resulted in different expert opinions and, as relevant to Roark's case, would have resulted in an acquittal. Id. at *23, *26.

On October 15, 2024, two days before his execution, Applicant filed his fourth subsequent habeas application (the instant -05 writ), raising a single claim that the decision in Roark provided a new legal basis for raising his Article 11.073 new science claim. The Court dismissed the application without addressing its merits and denied a stay of execution. Ex parte Roberson, No. WR-63,081-05, 2024 WL 4504434, at *1 (Tex. Crim. App. Oct. 16, 2024). Judges Richardson, Newell, Walker, and McClure dissented.

Enter the Texas House of Representatives. The House Committee on Criminal Jurisprudence sought to subpoena Applicant for testimony relating to proposed legislative changes to Article 11.073. A district court in Travis County granted a temporary restraining order the afternoon of Applicant's execution. The Attorney General sought mandamus relief, which we conditionally granted, vacating the district court's temporary restraining order, which permitted the execution to proceed. In re Tex. Dep't of Crim. Just. ex rel. Paxton, No. WR-96,121-01, 2024 WL 4512269, at *2 (Tex. Crim. App. Oct. 17, 2024). The Supreme Court of Texas intervened and stayed the execution, In re Tex. House of Representatives, No. 24-0884, 2024 WL 4521051, at *1 (Tex. Oct. 17, 2024), and then ordered briefing, In re Tex. House of Representatives, No. 24-0884, 2024 WL 4530038, at *1 (Tex. Oct. 20, 2024). After briefing, on November 15, 2024, the Supreme Court concluded that the State's separation-of-powers jurisprudence does not “permit judicial enforcement of a legislative subpoena that would require canceling a long-scheduled execution.” In re Tex. House of Representatives, 702 S.W.3d 330, 346 (Tex. 2024).

On February 24, 2025, Applicant filed his fifth subsequent habeas application (the -06 writ). In the -06 writ, Applicant seeks relief based on (1) actual innocence, (2) Article 11.073, and (3) deprivation of a liberty interest under federal law. Applicant also filed suggestions that the Court reconsider its rulings on Applicant's -03, -04, and -05 writs. On July 16, 2025, the district court scheduled Applicant's execution date for October 16, 2025. Applicant filed a motion to stay his execution and a motion for oral argument. On August 20, 2025, Applicant filed his sixth subsequent habeas application (the -07 writ). He again asserts his actual innocence, and now for the first time claims that the trial court was biased and should have been disqualified or recused.

II. Applicable Law

a. Article 11.073

Article 11.073 of the Texas Code of Criminal Procedure permits an applicant to obtain post-conviction habeas relief based on a change in science relied on by the State at trial if certain requirements are met. Tex. Code Crim. Proc. art. 11.073. Those requirements include, inter alia:

(a) This article applies to relevant scientific evidence that:

(1) was not available to be offered by a convicted person at the convicted person's trial; or

(2) contradicts scientific evidence relied on by the state at trial.

(b) A court may grant a convicted person relief on an application for a writ of habeas corpus if:

(1) the convicted person files an application, in the manner provided by Article 11.07, 11.071, or 11.072, containing specific facts indicating that:

(A) relevant scientific evidence is currently available and was not available at the time of the convicted person's trial because the evidence was not ascertainable through the exercise of reasonable diligence by the convicted person before the date of or during the convicted person's trial; and

(B) the scientific evidence would be admissible under the Texas Rules of Evidence at a trial held on the date of the application; and

(2) the court makes the findings described by Subdivisions (1)(A) and (B) and also finds that, had the scientific evidence been presented at trial, on the preponderance of the evidence the person would not have been convicted.

Id. art. 11.073(a)–(b).

Article 11.073 also includes its own writ bar. Subsection (c) provides:

For purposes of ․ Section 5(a)(1), Article 11.071, ․ a claim or issue could not have been presented previously in an original application or in a previously considered application if the claim or issue is based on relevant scientific evidence that was not ascertainable through the exercise of reasonable diligence by the convicted person on or before the date on which the original application or a previously considered application, as applicable, was filed.

Id. art. 11.073(c).

To make this determination, we consider:

[W]hether the field of scientific knowledge, a testifying expert's scientific knowledge, or a scientific method on which the relevant scientific evidence is based has changed since:

(1) the applicable trial date or dates, for a determination made with respect to an original application; or

(2) the date on which the original application or a previously considered application, as applicable, was filed, for a determination made with respect to a subsequent application.

Id. art. 11.073(d).

“Scientific method is defined as ‘[t]he process of generating hypotheses and testing them through experimentation, publication, and republication.’ ” Ex parte Robbins, 478 S.W.3d 678, 691 (Tex. Crim. App. 2014). “ ‘Scientific knowledge’ includes a change in the body of science (e.g., the field has been discredited or evolved) and when an expert's opinion changes due to a change in their scientific knowledge (e.g., an expert who, upon further study and acquisition of additional scientific knowledge, would have given a different opinion at trial).” Ex parte Chaney, 563 S.W.3d 239, 255 (Tex. Crim. App. 2018).

I agree with Judge Richardson's bottom-line conclusion: Our decision in Roark “provides a new legal basis for Applicant to overcome the subsequent writ bar.” Ante at 4 (Richardson, J., concurring).

III. Analysis

a. Is this a shaken baby case?

I start by addressing the overarching question that hangs over this case: Was Applicant tried on a shaken baby theory? Based on the State's arguments at trial, coupled with the evidence admitted to the jury, it is unquestionable that the State's underlying theory was a shaken baby death.² The transcript of Applicant's trial reveals at least 79 references to shaking or tangential uses of the term in voir dire, 130 references in guilt/innocence, and 15 references in punishment. For ease of reading, I will list some of the shaken baby references but describe Dr. Squires’ and Dr. Urban's testimonies separately:

1) In the State's opening argument, the prosecutor stated that Applicant “grabb[ed Nikki's] face” and “shakes her face” after finding her unconscious.

2) In opening, the prosecutor also described an unrelated incident where Applicant was allegedly observed “violently shaking [Nikki] back and forth, screaming at her to shut up,” and how Nikki's head “popp[ed] back and forth as he was shaking her.

3) Lastly, in opening, the State told the jury that Dr. Squires, the treating physician, would conclude that “Nikki died or rather was the victim of child physical abuse consistent with the picture of what they call shaken impact syndrome.”

4) In Applicant's opening statement, Applicant's trial counsel told the jury that this case was “not a capital case and the evidence will not support it. This is, however, unfortunately a shaken baby case.”

5) Applicant's trial counsel also stated that “[t]he evidence will show that Nikki did suffer injuries that are totally consistent with those applied by rotational forces more commonly known as shaken baby syndrome.”

6) Lastly, referencing voir dire, Applicant's trial counsel argued that “[e]very one of you related that you had heard the term shaken baby, that it was an act of basically a lack of control of emotion. It's a bad thing, but it's not something that rises to the level of capital murder.”

7) Courtney Berryhill testified that she would from time to time go spend the night at Applicant's house. She testified that she observed Applicant, on one occasion, shake Nikki when “[h]e picked her up” on her shoulders and “shook her to make her stop crying.” Applicant “shook her hard.”

8) Rachel Cox testified that she, Courtney, and Applicant lived together for some time. Rachel testified that she saw Applicant discipline and spank Nikki by “pick[ing] her up and then shak[ing] her and spank[ing] her.” Rachel said that Applicant did this about 10 times. She demonstrated the shaking to the jury using a teddy bear and testified that Applicant would shake Nikki for “[a] few seconds.”3

9) Teddie Cox, Applicant's girlfriend, testified that she observed on one occasion Applicant “picking [Nikki] up and shaking her” before throwing her back on the bed. The shaking lasted for “a few seconds.”

Dr. Squires testified as follows:

Q: Okay. So let's talk about that. What you wrote is when you saw her, she wasn't going to live, and your diagnosis was massive brain injury and your only explanation was trauma. And medical findings is a picture of shaken impact syndrome. All right. It's a pretty significant diagnosis, doctor. Can you explain to us then what shaken impact syndrome is?

A: There's a very well known, well described entity in children and it goes by several terms. Most of the lay public knows term shaken baby syndrome. And what, and if I may just for a minute, explain shaken baby. When one human being is much smaller than-- Let me say it this way. Children are uniquely at risk that if you take a child and you shake them, their head will go back and forth very forcefully and you know that you can cause major brain injury doing that. And one of the features is that you might not be able to see anything on the outside and have all these significant brain injury. And the reason babies are so prone to that, there's lots of reasons, but mainly it's because they're so small compared to how big whoever it is shaking them. In addition, their heads are big compared to their bodies, their neck muscles are weak, and they don't-- They're not conscious enough to protect their neck. In addition their brains have higher water content. So for all those reasons, shaken baby has been a well described entity. Now, some people think that with shaken baby that the most part of the damage is that they're often shaken and then thrown against something. And at the time when the head is moving back and forth very, very vigorously and then all of a sudden it stops against something; that at that moment is probably when a lot of the damage is being done because these shearing forces actually go through the brain itself. There are some experts that think that you cannot kill a child by just shaking alone, but you have to-- And they call it shaken impact. So the term is about the same. I will say that most, when I would consider most of the experts do think that shaking alone, if done vigorously, will kill a child, but most children are shaken and then thrown against something. And it's in the whole context of the head being vigorously shaken back and forth and then slammed against-- It can be a mattress, so that maybe there's no signs of trauma at all and yet as that head is moving and then suddenly stops, those shear forces go through it and cause tremendous damage to the brain, deep in the brain.

Q: And in Nikki's case you did have-- you had dramatic evidence of an impact to the back of her head?

A: I would like to say, you know, one possibility is that the impact happened at a different time. I mean, you know, I can't, a hundred percent. What I know is that there was an impact because it was swollen. Clearly, the most likely thing was that there was an impact that had to-- But the actual brain injury, we do not feel is explained by a simple impact.

Q: All right. And the items we talked about, the subdural hemorrhages, the retinal hemorrhages, and the brain swelling; what are they indicative of?

A: Well, it is my opinion, my estimation after a consultation with all that there was some component of shaking that happened to explain all the deep brain injury out of proportion, I would say, to the injury to the skull and the back of the head. There had to have been something more than just impact. We see children fall out of windows and all sorts of things and we know what an impact injury looks like and when you see this much damage deep to the brain, then you see subdural blood. The reason subdural blood is so important is there are little blood vessels that go between the bone and the dura. And when you shake a baby those blood vessels break and you get blood over the top of the brain. So whenever we see lots of subdural blood, I don't mean localized right under a fracture, but all over, usually that's indicative of this shaking. And then the retinal hemorrhages are just further-- It's one more thing that really lets you know that those eyes were being shaken and that the blood vessels broke.

Q: And then you've got some additional findings there. As far as the onset of symptoms with a child that's hurt this badly, is it a prolonged thing where it just develops hours and hours or how does that happen?

A: It's a spectrum. Some shaken babies are very mild and people might not even realize it. Other children, if you shake them hard enough and you hurt them bad enough, they stop breathing immediately. So anything in-between. It is my assessment in this child that after the event that caused all this deep brain injury she would hot have been normal. And any reasonable person would know that she wasn't normal. However, she could live for several hours and might not totally stop breathing long enough-- She certainly could live for hours after the event, but she would never have talked, walked, and been thought to be normal by anybody.

* * *

Q: Okay. All right. Let's talk then about that a little bit. The amount of force that is necessary to cause these damages, all right, what kind of force are we talking about? Are we talking about just a little shaking or what?

A: No. It's a very violent forceful act. It is not something that ever happens accidentally. It is not something that you see in normal children who are cared for by reasonable adults. It's a very violent act.

Later on cross-examination, Dr. Squires testified:

Q: Okay. In talking with you before, you told me that it's your best feeling that the brain injury occurred by virtue of the rotational force of the shaking; is that correct?

A: Yes.

* * *

Q: Okay. You also used the term frequently called, ‘reasonable parent’. Are you intending that to mean someone of average intelligence, experience, and bonded with their child?

A: I'm not even sure it's average. I think a person who is capable of caring for a child. A person who has enough intelligence to know how to handle a child I think would recognize violent shaking for the violence of which it is.

* * *

Q: You also went into that the reaction of a child after that incident goes all across the spectrum; is that correct?

A: No, I'm sorry, if I--

Q: Okay.

A: In just, in the act of shaking, in this case, with this severe injury, it's my strong assessment that after the event that caused that, after she was shaken, this child would have not have been normal. She would not have talked. She would not have walked. Anybody that would have had any interaction would have known that this child was in trouble.

Q: You saw no other indication of traumatic injuries to the child; is that correct?

A: That's correct.

Q: No other significant bruises along the upper thighs, legs, buttocks, abdomen, arms?

A: That's correct. I saw no bruising.

Q: Okay. And the skeletal survey showed no orthopedic injury?

A: That's correct. There were no fractures.

Q: No old injury either?

A: No old fractures.

Q: No scar tissue? You know, sometimes you see scar tissue?

A: Sometimes we--

Q: Not necessarily orthopedic, but soft tissue.

A: No, I did not detect any old scar tissue that I thought were signs of trauma.

And on re-direct examination, Dr. Squires testified:

Q: All right. So let's talk about the violence one more time; the degree of violence that's necessary to do this to a child. Can you characterize the level of violence that's required?

A: It is not a-- You don't just take a kid and sling them once or twice. You really have to shake them really hard back and forth and then you typically slam them against something. It's an out of control, angry, violent adult.

Dr. Urban, the State's medical examiner, testified:

Q: And then looking at the evidence of injury, you say that those injuries are caused by blunt force. And it was hard to look at the pictures. So just tell us what you mean by blunt force. What causes blunt force?

A: Typically in a-- Especially in a child this age, blunt force can be caused both by-- well, by an impact to the head, so being struck with something or being struck against something. Shaking also falls into this definition of blunt force and when enough-- And although it doesn't seem like, you know, shaking is not necessarily striking a child, when you are-- When a child is say, shaken hard enough, the brain is actually moving back and forth within, again, within the skull, impacting the skull itself and that motion is enough to actually damage the brain.

Dr. Urban also testified repeatedly about shaking injuries, including the injuries typically associated with shaken baby cases. As to the cause of death, Dr. Urban testified:

Q: And then you conclude that she died as a result of blunt force head injuries?

A: Yes.

Q: And then is there any way to tell, to say I believe this much of her death was caused by the shaking and this much of her death was caused by the battering she took?

A: No.

Q: There's no way to segregate it out?

A: No.

Q: Just all those things put together killed Nikki?

A: Yes.

After both sides rested and closed, the parties discussed objections to the trial court's proposed jury charge. Appellant objected to the inclusion of a deadly weapon instruction. The State argued that “there's evidence from the medical examiner, the forensic scientist, that it's shaken impact syndrome.”

And in closing arguments, the following references were made:

1) The State quoted Applicant's trial counsel and stated that “this is a shaken baby case. It is not accidental.”

2) The State referenced the testimony of Courtney, Rachel, and Teddie, and the prior instances of Applicant shaking Nikki.

3) The State discussed Dr. Squires’ testimony and told the jury that Dr. Squires had testified that “these were inflicted injuries consistent with not just shaken baby syndrome, this is not a child that was just shaken out of frustration, but shaken impact syndrome.” And the State recalled the testimony of Dr. Urban, stating that Dr. Urban had testified that Nikki was “[n]ot just shaken,” but had suffered “blunt force injuries ․ multiple blows to the head.”

4) Applicant's trial counsel argued that “this is a shaken baby case, but no, this is not a murder case.”

5) Applicant's trial counsel stated “Dr. Squires was asked by the State and by us whether this case was a classic shaken baby case. And it is.”

6) Applicant's trial counsel argued that there “is no evidence of intent or knowledge to commit the specific act of murder,” because “shaken baby syndrome is what applies.”

7) The State in rebuttal argued that Applicant did not “just shake [Nikki].”

8) Referencing Dr. Urban's testimony, the State argued that shaking Nikki is “like turning off a light switch. Shake and it scrambles the brain and they're rendered in a state of unconsciousness․”

9) The State argued that, “the last thing that [Nikki] saw before he killed her was the hate in her dad's eye when he was shaking her to death is what she saw.”

In sum, the answer to the preliminary, overarching question for this case—whether this is a shaken baby case—is obvious. To say that this case is not a shaken baby case is to ignore large swaths of trial testimony and repeated references in the arguments of not just the defense, but the State as well.⁴

b. Applicant's New Evidence

To get Article 11.073 relief, Applicant must show:

(A) relevant scientific evidence is currently available and was not available at the time of the convicted person's trial because the evidence was not ascertainable through the exercise of reasonable diligence by the convicted person before the date of or during the convicted person's trial; and

(B) the scientific evidence would be admissible under the Texas Rules of Evidence at a trial held on the date of the application; and

(2) ․ had the scientific evidence been presented at trial, on the preponderance of the evidence the person would not have been convicted.

Tex. Code Crim. Proc. art. 11.073(b). “Relevant scientific evidence” is evidence that “(1) was not available to be offered by a convicted person at the convicted person's trial; or (2) contradicts scientific evidence relied on by the state at trial.” Id. art. 11.073(a). For the foregoing reasons, Applicant has met his pleading burden for a remand in light of Roark.

i. Applicant's New Evidence

To support his Article 11.073 claim, Applicant sought and received opinions from a litany of experts. All of Applicant's experts uniformly concluded that the evidence was insufficient to support Applicant's conviction and that the scientific knowledge underlying the State's theory at trial had changed such that no expert would testify today that Applicant caused Nikki's death.

Dr. Harry J. Bonnell. (-03 writ, Exhibit A). Dr. Bonnell is a medical doctor and expert in forensic pathology. Dr. Bonnell noted that “[n]either medical facility” that Nikki was admitted to “performed a lumbar puncture (to obtain cerebrospinal fluid for evidence of meningitis or bleeding),” but instead “started the patient on antibiotics which would mask any meningitis or sepsis (blood infection) present.” Dr. Bonnell also noted the mis-intubation of Nikki which “prevented half of her lung from receiving air until the mistake was identified on a later chest x-ray film and the tube re-positioned.”

Based on his review of the records, Dr. Bonnell concluded that “the diagnosis of intracerebral bleed due to central vein or sagittal sinus thrombosis due to meningitis due to middle ear infection cannot be ruled out as the cause of [Nikki's] death with a reasonable degree of medical certainty.” Furthermore, “[t]he findings on x-ray imaging are consistent with the diagnosis that the intracerebral bleed was due to central vein or sagittal sinus thrombosis due to meningitis due to middle ear infection.” Lastly, Dr. Bonnell opined that “fatal injuries can occur from short falls of less than three feet in height,” and such a fall “cannot be ruled out as the cause of [Nikki's] death.”

Like Dr. Auer, as discussed infra, Dr. Bonnell emphasized that “the medical personnel in this case seemed to adopt the presumption that child abuse was the root cause from the outset, a presumption that appears to have imposed a cognitive bias that prevented them from investigating all reasonable possible causes of death either before or after her death.” This “cognitive bias engendered by the presumption of child abuse and the certainty within the general medical community at the time that the triad of symptoms ․ amounted to a medical diagnosis of child abuse,” led authorities to ignore “multiple possible causes” of death.

Dr. Janet Ophoven. (-03 writ, Exhibit B). Dr. Ophoven is a forensic pathologist with special training and over thirty years of experience in the evaluation, investigation, and interpretation of childhood injuries. Dr. Ophoven stated that the scientific understanding of shaken baby syndrome and its implications had changed considerably. In 2002, the diagnosis was often applied to cases of children who had little to no signs of physical trauma to the body or brain substance. However, since 2002, key research has prompted changes in the medical consensus concerning shaken baby syndrome. Dr. Ophoven explained that because of this fundamental change in the understanding of the diagnosis, courts could no longer rely on opinions and testimony regarding infant head injuries from the early 2000s.

Dr. Ophoven further opined that Dr. Urban's testimony failed to address any of the other circumstances or conditions now known to cause retinal hemorrhages, including but not limited to increased intracranial pressure caused by infection, coagulation abnormalities, chest compression, and hypoxia, all of which were present in Nikki's case. Dr. Ophoven also stated that the vast majority of head injuries in children Nikki's age are in fact caused by falls and that Dr. Urban's testimony lacked any scientific basis for the determination that the mechanism of Nikki's head injury had to have been deliberately inflicted and thus the result of homicide. Discussing Applicant's story—the fall off the bed—Dr. Ophoven concluded that “the injury potential from the fall, as described, should have been biomechanically evaluated before a fall could be dismissed as a possible cause or contributing factor to [Nikki's] death.”

Dr. John Plunkett. (-03 writ, Exhibit C). Dr. Plunkett is a general and forensic pathologist with a specialized concentration in infant head injury.⁵ Dr. Plunkett opined that neuroresearchers now understand that the treatment a child presenting with head injuries receives can be what causes the condition that is later interpreted as evidence that the injuries were inflicted. Dr. Plunkett also highlighted medical conditions that cause the “triad” of symptoms associated with shaken baby syndrome, including a variety of infections caused by bacteria and viruses or breakthrough bleeding associated with thrombosis. Furthermore, Dr. Plunkett opined that new scientific studies indicate that “shaking” capable of causing actual brain damage is likely associated with significant structural neck damage, injuries not found in Nikki.⁶

Dr. Plunkett also explained numerous changes in the scientific understanding behind infant deaths like Nikki's. For example:

• A biomechanical analysis of a potential injury mechanism is now mandatory, because it is now known that the findings historically associated with shaking may be caused by household accidents and low-velocity falls.

• A short fall can cause the injuries at issue in this case. At trial, the jury was told it was “impossible” for a short fall to cause those injuries.

Dr. Ken Monson. (-03 writ, Exhibit D). Dr. Monson is a professor of mechanical engineering who specializes in the study of the brain's behavior at a cellular and vascular level during and after trauma. Dr. Monson's research has established that the accelerations measured in some short-fall scenarios approach, and even exceed, the estimated thresholds for brain injury. Dr. Monson noted that scientific research since Applicant's trial has concluded that children can be seriously injured from short falls. He critiqued the investigation in this case, particularly that no measurements were taken of the makeshift bed on the cinder blocks.⁷ Estimating the height of the bed, Dr. Monson opined that “accelerations calculated assuming a worst[-]case scenario would be expected to exceed the lower range of injury thresholds.”

Dr. Monson also criticized Dr. Urban's trial testimony, citing post-2002 research demonstrating that shaking is an unlikely cause of brain injury. Like Dr. Plunkett, Dr. Monson opined that “the lack of neck injury” thoroughly undermines the shaken baby theory.

Dr. Francis Green. (-04 writ, Exhibit 5). Dr. Green is an expert in lung pathology who has specialized in the field of lung diseases since 1978. Based on his review of Nikki's medical files, Dr. Green concluded that “Nikki Curtis died of a chronic viral interstitial pneumonia and its complications, and not as the result of any trauma, either inflicted or accidental.” This conclusion was based on Nikki's “clinical history of recurrent infections, indicative of immune deficiency, her history of a recent acute infection that apparently began about a week or more before her death, the associated medical records, and importantly her autopsy results including the associated pathology slides.”

Dr. Green emphasized that “Nikki's lung tissue pathology slides revealed several important abnormalities,” including “evidence of ongoing injury involving the large airways (tracheitis) and small airways (chronic lymphocytic bronchiolitis), and chronic interstitial pneumonia—all features characteristic of a chronic viral infection.” Additionally, Dr. Green found evidence of “terminal necrotizing pneumonia, consistent with a secondary bacterial infection.” Dr. Green concluded that his “pathologic examination of Nikki's lungs showed changes of a viral infection that had proceeded through the stages ․ leading to her death,” and that “[t]he viral infection was likely present for at least a week and possibly several weeks before death.” Importantly, “[b]oth the interstitial pneumonia and the bronchopneumonia would have caused severe hypoxemia (low blood oxygen), leading to cardio/respiratory arrest, admission to hospital and death.”

Discussing Dr. Urban's autopsy report, Dr. Green stated that he “disagree[d]” with the cause of death listed on the autopsy. Furthermore, Dr. Green opined that “Nikki's death may have been hastened by use of a combination of drugs that suppress respiration.” Dr. Green asserted that “[t]he subdural blood observed during the autopsy is explained by the oxygen deprivation that [Nikki] experienced because of her pneumonia.” He noted that, “[o]xygen deprivation can cause vessels in the dural membrane to leak. If oxygen deprivation persists, the subdural blood can accumulate and cause encephalopathy or brain swelling. This condition is not specific to trauma.”⁸

Dr. Julie Mack. (-04 writ, Exhibit 6). Dr. Mack is a board-certified radiologist published in the field of pediatric radiology. She was asked to review Nikki's Computed Tomography (CT) scan images, as well as review Dr. Green's findings and conclusions.⁹

As to the CT scans, Dr. Mack concluded that the “imaging findings show definitive evidence of an impact related insult to the right side of the head. There is soft tissue swelling of the scalp. There is no skull fracture.” Dr. Mack opined that “[t]he dural based bleeding is consistent with an impact.” Dr. Mack also stated that “[s]ubdural hemorrhage with associated edema can occur after short falls as a direct result of impact.”

As to the lung scans and Dr. Green's finding, Dr. Mack agreed with Dr. Green and concluded that “the cause of the increasing opacification seen over the course of several chest x-rays is not simply edema related to resuscitation efforts”—in other words, that the increasing opacification was caused by “disseminated intravascular coagulation [(DIC)], ․ most likely related to sepsis and a complication of [Nikki's] long standing lung disease.”

Dr. Kennan Bora. (-04 writ, Exhibit 7). Dr. Bora, a medical toxicologist and emergency medicine physician, reviewed Nikki's medical files to determine “what role, if any, the medications prescribed and found in [Nikki's] system may have played.” Dr. Bora concluded that “[t]he post-mortem toxicology results and the medication history are significant and may have contributed to her death.”

Dr. Bora first noted the “perplexing” presence of Phenytoin in Nikki's blood. Phenytoin is a medication that is primarily used to treat seizures. Yet, “there [was] no documentation for this medication having been given” to Nikki. There was also no evidence that Applicant administered the Phenytoin. Next, Dr. Bora noted that Nikki had a “post-mortem level of promethazine [that] was 0.4 mg/L.” Based on the half-life of the drug, Dr. Bora concluded that this was “clear evidence that [Nikki] was given too much promethazine,” the clinical impact of which “would be respiratory depression (not breathing as much) as well as potentially ataxia (unsteady gate) making it more likely for her to fall over.” Based on his observations, Dr. Bora concluded:

The patient had a febrile illness, severe sepsis, or septic shock, due to either a bacterial or viral infection. She had evidence of a markedly elevated promethazine level, significantly outside of therapeutic concentrations, which may have contributed to respiratory depression and her death. She had multiple drugs which would have decreased her seizure threshold (making seizures more likely) as well as drugs which would have caused her to be uncoordinated and more likely to fall and hit her head. Even without trauma, patient who have bacterial or viral infections that lead to severe sepsis or septic shock are more likely to bruise, bleed, and hemorrhage by a variety of pathways.10

Dr. Ronald Auer. (-04 writ, Exhibit 8). Dr. Auer is a research scientist, a medical doctor certified in neuropathology, and the author of a leading neuropathology treatise, Forensic Neuropathology and Associated Neurology. Dr. Auer reviewed Nikki's medical files, as well as Dr. Urban's autopsy findings, and concluded:

This 2-year-old had viral pneumonia with breathing arrest and fever, multiple visits to the doctor since age 7 days, culminating in a predictable cardiac arrest. Epinephrine drip superimposed upon cardiac arrest-induced disseminated intravascular coagulation led to recirculation detour hemorrhages around a non-perfused brain, which were misinterpreted as head trauma. Non-perfused brain, or brain death, results in detour hemorrhage around the brain, necessarily because the brain is non-perfusable, when circulation is restored.

This is not a fatal head injury. That mistaken notion arises from misreading a goose egg from when the child fell off the bed due to hypoxia, and a misinterpretation [of] the pathogenesis of subdural hematoma and retinal hemorrhage. The treating doctors are not to be faulted for failing to recognize the sequence of events that resulted in the internal head injuries that they observed. They were dealing with a young child in the midst of an irreversible health crisis and what they could observe in the ER plainly indicated a damaged brain. Nor did the treating doctors, under those circumstances, have time to stop and examine Nikki's medical history to make a complete differential diagnosis.

Dr. Auer first began with an assessment of the timeline behind Nikki's death. He noted that when Nikki arrived at the hospital unconscious and blue, this likely showed “a full cardiac arrest. If longer than 12 minutes or so, it is equivalent to brain death clinically or non-perfused brain pathologically, due to permanent closure of the brain microcirculation.” This was confirmed by the 15:02 entry in Nikki's medical records on January 31, 2002, which stated: “[p]upils remain fixed and dilated,” indicating she was already brain dead that afternoon. Furthermore, that Nikki was intubated (and later re-intubated because the breathing tube had been inserted incorrectly), not only “worsened hypoxia,” but “confirm[ed] that she was brain dead, meaning her brain microcirculation had closed.”

Dr. Auer then discussed the medication that was given to Nikki. First, Nikki was given epinephrine, a medication that “causes a hyperdynamic circulation around non-perfused brain, which together with DIC, gives copious bleeding around the non-perfused brain.” Second, Nikki was given vasopressin, a medication that “increases the blood pressure further exacerbating the bleeding tendency.” Third, Nikki was given dopamine, a medication that “increases the force of contraction of the heart (termed an inotropic agent) further exacerbating the bleeding tendency.” Fourth, Nikki was given heparin, which “causes bleeding, by inhibiting factor X, and thrombin, in the blood clotting cascade-pathway.” Based on this, Dr. Auer concluded that Nikki's “significant internal bleeding within the skull but virtually no external injuries” was a result of a “non-perfused brain as flow during resuscitation is diverted around the brain in a detour manner, as blood is unable to go through the brain.”¹¹

Dr. Auer also opined on the injury on the rear of Nikki's head. Judge Parker asserts that “[t]he ‘mushy spot’ on the back of the head was a group of bruises, and was the type of injury that was ‘really only seen in impacts to the head.’ ” Post at 2 (Parker, J., concurring and dissenting); see also post at 19 (Keel, J., dissenting) (discussing impact site to the back of Nikki's head). But Dr. Auer sees it quite differently:

The scalp “goose-egg” is not a fatal head injury, and is common among children, especially those woozy with hypoxia due to pneumonia. The goose egg does not comprise a fatal head injury by virtue of not bruising the brain i.e. not causing a contusion. Conflating this with fatal head trauma in this case is medical error. A fatal head injury is used to attempt to explain the reperfusion detour hemorrhages around non-perfused brain.

Dr. Auer stated that subarachnoid bleeding, the second type of bleeding identified by Judge Parker, id., is a “detour hemorrhage” that is “forced around the non-perfused brain because the blood cannot pass through the brain due to microcirculatory closure,” particularly given all the medications given to Nikki.

To the extent that bruising may be an indicator of assault, see, e.g., post at 2 (Parker, J., concurring and dissenting) (discussing bruising on Nikki's body), Dr. Auer opines that “there is no way hospital staff can turn a child in bed or touch the skin firmly during insertion of intravenous lines or during intubation, without causing bruising, when [DIC] is present.”¹²

Judge Parker also asserts that Nikki's “lacerated frenulum” indicates abuse. Id.; post at 12 (Keel, J., dissenting) (discussing Dr. Squires’ testimony that included “the frenulum laceration”). But as Dr. Auer notes, “[i]t is necessary to do a finger sweep prior to insertion of an orotracheal tube, or in fact inserting anything into the mouth. When the finger sweep occurs, the frenulum can be damaged. But here, we see the oral effect of disseminated intravascular coagulation.” Lest we also forget that the original intubation was botched and had to be reperformed.¹³ Moreover, Dr. Laposata, as discussed infra, noted that no medical records indicated that a torn frenulum was observed when Nikki arrived at the emergency room in Palestine. It can only follow that the frenulum injury—one of the injuries that Judge Parker relies on to support Applicant's guilt—was not caused by Applicant but rather the emergency treatment at the hospital.

In summary, Dr. Auer concluded that Nikki, “a sickly child from birth,” with “repeated episodes of ‘breathing apnea’ when she would stop breathing and turn blue,” was “quite ill ․ [f]or a week before [her] death.” She suffered from interstitial pneumonia which “led to cardiac arrest due to the profound effect of deficient blood oxygenation on the blood pressure.”

Dr. Auer then discussed why the misdiagnosis occurred. Dr. Auer opined that “in 2002, shaken impact syndrome was believed to be a unitary entity and was regularly diagnosed.” However, “[b]y 2009, shaking went mostly out the window,” and by 2017, “shaking was debunked.” Moreover, Dr. Auer noted that “many of the events ․ have nothing to do with [Applicant] and he could not conceivably be held responsible for events carried out in [the] hospital, not for the long infectious history of the child.”

Dr. Michael Laposata. (-06 writ, Exhibit 1). Dr. Laposata is a board-certified clinical pathologist who specializes in coagulation. Dr. Laposata reviewed Nikki's medical records and concluded that Nikki “clearly had a severe infection during the last week of her life,” as well as “elevated liver enzymes indicating abnormalities in her liver functioning.” The liver abnormalities meant that “Nikki's liver was unable to make new proteins ․ required to form a clot,” which was “contributory to bleeding throughout the body.” Moreover, Dr. Laposata noted that Nikki was treated with medication for the infections in the days before her death. The medication included Amoxicillin, Cefzil, Rocephin, Bactrim, and Omnicef—all antibiotics that would have “further reduced the clotting function of Nikki's platelets.”

Dr. Laposata also opined that Nikki suffered from “moderate” DIC during her hospitalization prior to her death. Infections—like the ones Nikki suffered from—can be stimuli that cause DIC. Dr. Laposata stated that “[i]n DIC patients like Nikki [ ] who are forming clots when there is no need to plug the hole in a bleeding vessel, the platelets and the coagulation factors are used up. It is for this reason that patients can spontaneously bleed with DIC and bruise easily with normal handling because they do not have enough functional platelets and coagulation factors.” Based on this, Dr. Laposata opined that “the presence of hemorrhages in the brain and behind the eye, and bruises, are all commonly observed in patients with DIC. All of the findings in Nikki's case associated with bleeding are clearly reflective of the presence of DIC.” This is because “[w]hen a person experiences DIC, their body rapidly uses up all of the clotting factors. Then, after their clotting factors are fully depleted, they can bleed anywhere and everywhere in their body ․ [and] a patient in DIC can develop bleeding under the skin in the form of bruises.” Furthermore, “DIC can also follow small amounts of damage to the brain from the small bleeds, such as Nikki's subdural hemorrhage, which is consistent with her reported short fall from the bed.” To the extent that there were other minor injuries, “Nikki's minor bruises on her face captured in the hospital photos and scalp are entirely consistent with [Applicant's] efforts to wake her, the life-saving efforts to resuscitate her upon her arrival at the Palestine ER, and then subsequent treatment by her medical providers during the course of her hospitalization.”¹⁴

Dr. Laposata also addressed the discrepancy in the amount of blood between the initial CT scan in Palestine and the autopsy. He concluded that “[t]his discrepancy is entirely consistent with bleeding that developed after Nikki arrived in the hospital caused by DIC.”¹⁵ Dr. Laposata also explained that the misdiagnosis of Nikki's condition was likely caused by an “unaware[ness] of chronic DIC.” In chronic DIC cases, like in Nikki's, “bone marrow works to compensate for low platelet count by making platelets more quickly.” Dr. Laposata opined that “[m]inor trauma in such a patient with chronic DIC can lead to major bleeding, especially if the infection is worsening into something more life-threatening like pneumonia.”¹⁶

Addressing the SBS triad, Dr. Laposata opined that at the time of Nikki's death and Applicant's trial, an SBS diagnosis “was presumed whenever a child, like Nikki, had intracranial bleeding, brain swelling, and retinal hemorrhages.” But, since then, “that presumption has been repudiated.” Moreover, “substantial research has demonstrated ․ that numerous medical and genetic conditions, including severe infections like Nikki had, as well as accidental short falls with head impact, as [Applicant] reported she seemingly sustained, can produce the same constellation of conditions formally exclusively attributed to SBS.” (Emphasis added).¹⁷

ii. “By a preponderance of the evidence”

The State and the dissenting opinions point to various pieces of evidence to argue that the disavowal of SBS would not have resulted in a different outcome in this case. Judge Parker argues that Applicant's “demeanor and statements shows more than mere nervousness and displays a consciousness of guilt.” Post at 4 (Parker, J., concurring and dissenting); see also post at 19 (Keel, J., dissenting). This argument falls flat for several reasons. First, to the extent that Judge Parker relies on Applicant's confession to Dr. Kelly Goodness, that confession was—as Judge Parker herself concedes—limited to shaking the baby after Applicant “lost it.”¹⁸ At no point did Dr. Goodness ever assert that Applicant had confessed to any conduct other than shaking Nikki. Second, Applicant's behavior after Nikki's death is attributable to his undiagnosed autism. For example, Detective Brian Wharton, the lead detective in Nikki's death, signed an affidavit stating:

The lack of emotion he displayed and the seeming lack of understanding about what was happening to him did not seem right. His demeanor did not seem normal, and that something seemed wrong with him is something that was discussed. Witnesses we interviewed at the hospital also mentioned how Mr. Roberson seemed odd. He was not behaving as expected either for a distraught parent, or a suspect. I learned only years later that he has since been diagnosed with Autism Spectrum Disorder.

* * *

I had doubts then, but today I am certain that nothing here supports a continued effort by the State to execute Mr. Roberson.

But Applicant's undiagnosed autism was never submitted to the jury, and instead the State argued—without response—that Applicant was callous, cold, and unfeeling. Third, Judge Parker argues that Applicant changed his story: “Applicant told Sergeant Wharton that Nikki had simply fallen of [sic] the bed, but later told his girlfriend Teddie Cox that she hit her head on a brick.” Post at 4 (Parker, J., concurring and dissenting); see also post at 19 (Keel, J., dissenting) (discussing Applicant's statements). Judge Parker conveniently ignores the fact that the bed—a makeshift mattress and box springs—was propped up on cinder blocks. So, when Applicant told Teddie that Nikki had hit her head on a brick, Applicant was referring to the cinder blocks that held up the bed Nikki had just fallen off from.

Our opinion in Ex parte Roark resolves the remainder of the other evidence of injuries that Judge Keel and Judge Parker argue indicates physical abuse. In Roark, we held “that scientific knowledge has evolved regarding [shaken baby syndrome] and its application in [Roark's] case. Additionally, we find that given further study, the experts would have given a different opinion on several issues at a trial today.” 707 S.W.3d at 185.

Judge Parker asserts that Roark “is readily distinguishable.” Post at 5 (Parker, J., concurring and dissenting); see also post at 18 (Keel, J., dissenting) (“Roark is factually distinguishable”). To start, as Judge Parker concedes, there are ample similarities between the two cases.¹⁹ See id. But the similarities do not end with Judge Parker's recitation:

• Both Roark and Applicant were the victim's sole caretaker during the period it was alleged that they shook the victim.

• There were no eyewitnesses to what transpired in each case.

• Each victim was diagnosed with shaken baby syndrome based on a subdural hematoma, retinal hemorrhaging, and brain swelling.

• The State in each case pursued a shaking-and-impact theory at trial.

• In both cases, the State relied primarily on Dr. Squires to diagnose shaken baby syndrome based on the so-called “triad” symptoms and to refute the defense's arguments. Indeed, the habeas court in Roark found that “Dr. Squires’[ ] testimony refuting the possibility of a rebleed was crucial to the State's case.” In granting relief, this Court concluded that her testimony was rendered unreliable by new scientific knowledge. Dr. Squires testified again in Applicant's trial and again invoked the triad as conclusive evidence of abuse. Dr. Squires’ testimony could not sustain the conviction in Roark, and it cannot sustain one here.

In Roark, the conviction rested on the “triad” symptoms (subdural hematoma, retinal hemorrhaging, and brain swelling) that are “very, very classically seen in the Shaken Baby Syndrome.” 707 S.W.3d at 162. This Court found that the case against Roark was “circumstantial”; that “there were no eyewitnesses to the events that transpired”; and that “the most persuasive evidence in trial was medical testimony.” Applicant's case rested on the same triad symptoms; the evidence against him was circumstantial; there were no eyewitnesses to the events that transpired; and the most persuasive evidence in trial was medical testimony. When the sole basis of conviction is forensic testimony later discredited, confidence in the verdict cannot be maintained. See id. at 187 (“[T]here would likely be no statements that it is ‘almost certainly’ or ‘classically aligns with’ a particular mechanism. This leaves the jury with little help.”).

Judge Parker reemphasizes that “Applicant's suspicious conduct and incriminating admissions go well beyond the nervousness and possible lack of empathy in Roark.” Post at 5 (Parker, J., concurring and dissenting). However, similarities between both cases abound. Roark's girlfriend's stepmother testified that the victim would cry every time Roark held her, and that Roark was “real quiet” at the hospital and had “no reaction” to what was going on. The State used this testimony to characterize Roark as abusive and cruel. In granting relief, this Court apparently did not find this evidence against Roark persuasive: Roark's behavior is not mentioned beyond factual assertions. Here, Applicant was accused by three witnesses at trial of having disciplined or shaken Nikki in the past and of having no reaction or a delayed reaction when transporting Nikki to the hospital and speaking with doctors. It is unclear why the Court now finds such evidence persuasive against Applicant, especially when we now know these reactions were a result of his undiagnosed autism, and no similar alternative explanation was offered for Roark.

This Court in Roark found that “scientific knowledge has evolved regarding SBS and its application in [Roark's] case,” 707 S.W.3d at 185, and further that “there would be a marked shift in the testimony today concerning the effect of a short-distance fall to a child, the effect of shaking a child, ․ retinal hemorrhaging, and SBS in general as applied to [the victim's] injuries,” id. This Court also relied on “multiple similar cases in which the relevant medical areas were assessed with new science and have come to the same conclusion as this Court in [Roark's] case.” Id. at 187.²⁰

The underlying principle between Roark and this case is the same: when the science relied upon by the State in obtaining a conviction is called into question by the medical community, such that the applicant would not be convicted using the relevant evidence today, the applicant is entitled to habeas relief. Applicant's contravening evidence establishes that Nikki died from a combination of undiagnosed chronic interstitial viral pneumonia, prescription medications that suppressed her breathing, and her infection that developed into sepsis, which caused DIC, a severe and often fatal bleeding disorder.

A jury today would not have heard the same conclusions from Dr. Urban as the jury in 2002 did. As we noted in Roark, Dr. Urban “would be confronted with twenty years of reputable scientific studies and publications that, if graphed, continually point away from their stated positions. If [Dr. Urban] were to experience the ostrich effect and wish to bury ․ her head in the sand, then [she] would have to bear the brunt of a grueling cross-examination.” Roark, 707 S.W.3d at 187. Instead, a jury would be presented with expert testimony indicating that Applicant's story—a fall from the bed—was consistent with Nikki's injuries. Applicant's 2003 trial, as well as the habeas court, did not have the benefit of our decision in Roark. I would remand Applicant's Article 11.073 and actual innocence claims to the habeas court for an evidentiary hearing in light of Roark. Applicant has alleged facts that are at least minimally sufficient to bring him within the ambit of the new legal basis for relief that we announced in Roark.

IV. Actual Innocence

In Ex parte Elizondo, 947 S.W.2d 202 (Tex. Crim. App. 1996), we recognized that an actual innocence applicant must prove by clear and convincing evidence that no reasonable juror would have convicted him based on the newly discovered evidence. Id. at 210. This is a Herculean burden. Ex parte Brown, 205 S.W.3d 538, 655–45 (Tex. Crim. App. 2006).

Applicant's actual innocence claims in his -06 and -07 habeas applications are based on exactly the same evidence as his Article 11.073 claims. I see no reason why this Court should remand one claim for evidentiary development but dismiss the other. In my view, the medical evidence developed by Applicant is sufficient to meet the subsequent writ bar under Section 5(a)(1). I would remand both Applicant's Article 11.073 and actual innocence claims and let the habeas court, in a speedy and just manner, develop the evidentiary record.

V. Conclusion

“[T]he penalty of death is qualitatively different from any other sentence.” Lockett v. Ohio, 438 U.S. 586, 604 (1978) (internal quotations omitted). This “qualitative difference between death and other penalties calls for a greater degree of reliability when the death sentence is imposed.” Id.

Several judges on this Court have recognized that Article 11.073 embodies the Legislature's decision to emphasize accuracy over finality in the context of post-conviction habeas relief. See, e.g., Ex parte Robbins, 478 S.W.3d at 704 (Cochran, J., concurring) (“In Robbins, this Court chose finality over accuracy; in Henderson we did the opposite, and in 2013, the Texas Legislature also chose accuracy over finality by enacting Article 11.073”); Ex parte Robbins, 560 S.W.3d 130, 161 (Tex. Crim. App. 2016) (Newell, J., concurring) (“By enacting Article 11.073 without any express limitation on what constitutes ‘scientific knowledge,’ the Legislature tipped the scales in favor of accuracy perhaps at the expense of finality.”); see also ante at 4 (Richardson, J., concurring) (“There is a delicate balance and tension in our criminal justice system between the finality of judgment and its accuracy based on our ever-advancing scientific understanding. A death sentence is clearly very final and, once carried out, hindsight is useless. Thus, when moving forward in such a way, we should require the highest standards of accuracy so that we can act with a reliable degree of certainty.”).

The Court today correctly chooses accuracy over finality. I would go further and also remand Applicant's actual innocence claims—which are based on nearly identical facts as his Article 11.073 claims—for further evidentiary development in light of our decision in Roark. With these thoughts, I join the Court's remand order but also respectfully dissent insofar as Applicant's actual innocence claims are not also remanded for further evidentiary development.²¹

Appendix A

CONCURRING AND DISSENTING OPINION

One of Applicant's main and persistent claims is that the scientific theory of “shaken baby syndrome” has been discredited. But we need not concern ourselves with the status of that scientific theory here. Applicant's case is not a “shaken baby” case. It is not even a so-called “shaken impact” case, in which shaking is stopped by an impact on a pillow or other soft object, which then purportedly causes the baby's brain to crash into the inside of the skull. This case is a “blunt-force trauma” case. A shaking or “soft” impact cannot explain the “boggy” spot on the two-year-old child's head, the two different kinds of bleeding outside the skull, the bruising on the face and ears, and the bleeding from a tear in the mouth. One of Applicant's other repeated contentions is that the child could have died from a short fall. But when one catalogues the extensive injuries suffered by the child, including two different types of fresh, inside-the-skull, bleeding, the obvious conclusion to draw is that the child was beaten to death.

Nikki, the decedent in this case, was not a “baby.” She was over two years old. She was far less vulnerable to injury than an actual infant, or even the 13-month old child in Ex parte Roark.¹ Nikki had extensive injuries to the face and the back of the head,² which were described as “massive” and “very extensive.” She had a bruise on her chin, bruising along the left side of her jaw and cheek, a bruise on the right side of the forehead, a scrape next to the left eye, and bruising on the back of the head that was severe enough to cause bleeding underneath the scalp, resulting in a swollen, “mushy” or “boggy,” soft spot. The “mushy spot” on the back of the head was a group of bruises, and was the type of injury that was “really only seen in impacts to the head.” The bruising on Nikki's face looked like a hand print. She had a lacerated frenulum (connective tissue in the mouth between the upper gums and the upper lip) and bruising just inside each ear.

Nikki suffered from five different types of bleeding—two inside the skill, two outside the skull, and one in the mouth. Inside the skull was subdural and subarachnoid bleeding. Outside the skull, but underneath the scalp, was subscalpular and subgaleal bleeding, which were sources of the “mushy” spot discussed earlier. And the bleeding from the mouth was likely from the torn frenulum. All of the bleeding was fresh blood, indicating recent injuries.

The emergency room physician testified that there was “extensive bleeding all throughout the head.” Consequently, he concluded that a single head impact from a fall from a bed would be extremely implausible, if not impossible. The pathologist testified that the hemorrhage beneath the scalp (outside the skull) was “very dense” and in “separate” locations. From this observation, she concluded that there were “multiple impacts over the entirety of the head.” She further concluded:

In particular, with this much hemorrhage it's very difficult to elucidate exactly [how] much blows there were. But I'm confident with these separate areas of dense hemorrhage and separate areas on the head that there were multiple blows to different points on the head.

The pathologist said that the injuries to the child's brain would have been “immediate” and that the child would have “immediately suffered what we call a change in the level of consciousness.”

The pathologist also testified that Nikki had a bruise on the back of her right shoulder, a scrape on the left arm, a scrape on the left pinkie toe, a hemorrhage in the muscles of the back, and a very small amount of hemorrhage in the muscles of the neck on the back side at the base of the skull. Unsurprisingly, the manner of death listed in the autopsy was “homicide.”

The presence and severity of external injuries—that is, injuries outside the skull—debunk any notion that the homicide finding in the present case depends on the vitality of the scientific theory of shaken baby syndrome. Producing these external injuries required forceful contact with a hard surface. And the testimony was that there had to have been multiple forceful impacts to the head. The severity of the internal injuries and the multiple types of internal bleeding (inside the skull), when viewed with the external injuries, lend further support to the notion that a great deal of force was exerted against the child's head.

In fact, “shaken baby syndrome” was Applicant's defense. He tried to convince the jury that the death was inadvertent by relying on the idea that the child could have been killed due to shaking alone. Casting doubt on “shaken baby syndrome” actually hurts his case because it would make a jury even less likely to believe that the child's death was unintentional.

Moreover, Applicant's demeanor and statements show more than mere nervousness and display a consciousness of guilt. He called his girlfriend, Teddie Cox, to tell her that he was taking Nikki to the hospital because she was not breathing and would not wake up. But when Teddie called back, Applicant still had not left the house, and she urged him to do so. After the child had been taken to the hospital, Applicant actively prevented a nurse from talking to the child's maternal grandparents. To this same nurse, Applicant admitted to waking the child up every hour, which suggested he was aware of a head injury. He also said to her, “You know, I love my little girl. I would never mean to hurt her”—suggesting that he knew he had hurt her. Applicant told Sergeant Wharton that Nikki had simply fallen off the bed, but he later told his girlfriend Teddie Cox that she hit her head on a brick.

At the jail, Teddie directly asked Applicant if he had killed Nikki. Applicant responded that if he did do it, he did not remember and that he might have “snapped” but didn't remember doing so. And at the punishment phase of trial, Dr. Kelly Goodness, Applicant's expert, testified that Applicant admitted that he “lost it” and shook Nikki. He also admitted to her that he had lied to the authorities.

And Applicant had a previous history of abusive conduct toward the child. Courtney Berryhill, Teddie's eleven-year-old niece, saw Applicant shake Nikki by the arms in an attempt to make her stop crying. Rachel Cox, Teddie's ten-year-old daughter, testified that Applicant had a “bad temper” and that she had seen him shake and spank Nikki when she was crying. Rachel had seen this conduct occur about ten times. She also saw Applicant threaten to kill Nikki. Teddie testified that Applicant once yelled at Nikki, “If you don't shut up, I'm going to beat your ass.” She also saw Applicant hit Nikki with his hand and once with a paddle. Teddie also saw Applicant, while angry with Nikki, pick her up off the bed, shake her for a few seconds, and throw her back on the bed.

To the extent Applicant relies upon our recent decision in Roark, that case is readily distinguishable. It is true that the children in both cases had brain swelling with a mid-line shift. Both children had retinal hemorrhaging that was large enough to be seen by the naked eye. Neither child had a skull fracture. Both children had some injuries that might be indicative of sexual assault but ultimately could not medically be ruled as such. And in both cases, the defendant at the hospital seemed overly nervous and did not display emotions that would be expected of a close caregiver of a child whose life was in danger.³

But the child in Roark had no recent external injuries to the head,⁴ and she suffered from only one type of bleeding—subdural bleeding—which was present only inside the skull.⁵ Moreover, some of the blood inside the child's skull in Roark was from old blood, which suggested the possibility that the fresh blood could be a rebleed.⁶ This contrasts with Nikki suffering multiple external injuries to the head, five different types of bleeding, and all of the bleeding being fresh blood. It is also worth noting that the child in Roark did not die,⁷ but Nikki arrived at the emergency room “lifeless” and died of her injuries. And Applicant's suspicious conduct and incriminating admissions go well beyond the nervousness and possible lack of empathy in Roark.

And to the extent Applicant relies on experts who claim that the child's various bruises and bleeding could be the result of pneumonia or “disseminated intravascular coagulation,” this illness hypothesis is contradicted by the child's clear lungs, clear chest x-rays, and normal labs.

In summary, the evidence that the child died from multiple “hard” impacts to the head was strong. The story about the child falling out of the bed lacked credibility and was inconsistent with the evidence. Any evidence undermining “shaken baby” syndrome would not have helped Applicant, but would have made the jury more confident that Applicant intentionally and knowingly caused the child's death.

“Methinks he (Applicant) doth protest too much.” What can Applicant say about his confession to his own expert witness that he was in a rage? Or his confession to the mother of the child that he lost it and snapped? Or the testimony that Applicant had previously abused the child on a number of occasions? Or his various lies to the police? Or the expert diagnosis that the cause of the child's injuries and subsequent death was blunt force trauma? Or the autopsy finding of homicide?

Applicant has filed seven habeas applications—his seventh filed before we have even disposed of his sixth. And he has filed multiple “suggestions for reconsideration” in applications three through five, with this being his fourth suggestion to reconsider his third application, his third suggestion to reconsider his fourth application, and his second suggestion to reconsider his fifth application. Applicant has engaged in a course of conduct of relitigating past claims, repackaging earlier claims in ways that attempt to convey a change in substance but really don't, and adding new claims he could have raised long ago. I agree with the Court that his sixth and seventh applications do not meet an exception to the bar against subsequent applications,⁸ but I would also hold that all of his suggestions for reconsideration are without merit.⁹

In Ex Parte Torres, this Court observed that the subsequent-writ provisions in Articles 11.07 and 11.071 were meant to generally limit a convicted person to “one bite of the apple.”¹⁰ Applicant has been given numerous bites of the apple for approximately fifteen years on his journey to seek relief from his conviction. He has, in fact, had an entire apple pie.

The law does not afford him the relief he is seeking for the capital murder of his own daughter, whom he killed when he was in a rage because she was crying. Applicant's own trial attorney characterized Applicant's conduct as “the act of an angry, out of control parent” and essentially conceded that Applicant's short-fall story was a lie. From the grave, Nikki's voice calls out for justice.

I concur in the Court's order dismissing Applicant's sixth and seventh applications, but I dissent to the Court's decision to remand on his fifth application.

DISSENTING OPINION

The Court rewards Applicant's abuse of the writ with an unmerited and pointless remand of his perennial challenge to cause-of-death testimony offered at his trial. The remand is unmerited because Applicant has not met Article 11.073’s unavailability or materiality requirements or overcome Article 11.071’s subsequent writ bar. See Tex. Code Crim. P. arts. 11.071 and 11.073. And it is pointless to order factual development of a legal question, namely, the meaning of Ex parte Roark, 707 S.W.3d 157 (Tex. Crim. App. 2024). Instead of rewarding Applicant's abuse of the writ, the Court should—and in fact must—uphold—again—the trial court's judgment and sentence of death. See Tex. Code Crim. P. art. 11.071 § 5(a)(1) (forbidding consideration of a subsequent writ unless it establishes that its current claim rests on a previously unavailable factual or legal basis).

I. Overview

Applicant spent only one night alone with his 27-month-old daughter, Nikki Curtis. The morning after, he brought her to the emergency room where she was found to be limp, lifeless, blue, and bruised. Her pupils were fixed and dilated, and the back of her head exhibited a boggy, mushy, bump. Applicant gave inconsistent accounts for Nikki's condition, centering around a possible fall from a bed— an explanation that the medical professionals found at odds with her condition. She died days later after life support was withdrawn.

At Applicant's trial for capital murder, the prosecution's medical and forensic testimony showed that he had beaten and shaken her to death, and the defense did not dispute it. Instead, the defense embraced it and claimed that he had “lost it” but had not meant to kill her—a strategy that pursued a lesser offense with a lower culpable mental state than is required for capital murder of a child. See Tex. Penal Code §§ 19.02, 19.03. The jury rejected the lesser options, found him guilty as charged, and answered the punishment-phase special issues in favor of the death penalty. See Tex. Code Crim. P. art. 37.071 (b)(1) and (d)(1).

Applicant's conviction and sentence were affirmed on direct appeal, and he was denied relief on his first writ. His second writ—a pro se effort—was dismissed. His second subsequent writ—our -03—presented his first challenge to shaken baby syndrome (“SBS”) as “junk science” under Article 11.073—then a new legal basis for relief. See Tex. Code Crim. P. art. 11.071 § 5(a)(1). We stayed his execution and remanded his application to the habeas court for a hearing. After hearing the evidence, the habeas court entered findings, recommended against relief, and sent the case back to us. We made our own review, agreed with the habeas court, and denied relief.

Since then, Applicant has made the same 11.073 claim in subsequent applications and suggested that we reconsider our dispositions of them. In support of those filings, he has resubmitted the evidence he first offered in support of his -03 claim and added more expert opinions and articles rendered or published since then that agree with and amplify his original claim under Article 11.073—namely, that testimony about SBS is scientifically unsound. And until today, we have dismissed the repeated claims and refused the suggestions, and rightly so: he has never shown his evidence to have been previously unavailable or material as required by Article 11.073, nor has he overcome the subsequent writ bar.

II. Article 11.073

Article 11.073 applies to “relevant scientific evidence” that was not available at trial or that “contradicts scientific evidence relied on by the state at trial.” Tex. Code Crim. P. art. 11.073 (a). It offers relief via habeas if the application shows that the relevant scientific evidence would have been admissible, was previously unavailable, and would have been material at trial. Id. art. 11.073 (b). Two of these requirements are at issue here—unavailability and materiality.

Unavailability depends on timing and the exercise of reasonable diligence. Id. The applicant must show that “currently available” relevant scientific evidence was not available at trial “because the evidence was not ascertainable through the exercise of reasonable diligence by the convicted person” before trial. Id. art. 11.073 (b)(1). Materiality requires a finding that, “had the scientific evidence been presented at trial, on the preponderance of the evidence the person would not have been convicted.” Id. art. 11.073 (b)(2).

To decide “whether relevant scientific evidence was not ascertainable through the exercise of reasonable diligence on or before a specific date, the court shall consider whether the field of scientific knowledge, a testifying expert's scientific knowledge, or a scientific method on which the relevant scientific evidence is based has changed since” trial or since the filing of a previous writ if the claim is made in a subsequent application. Id. art. 11.073 (d).¹

In short: Article 11.073 requires (1) a previously unavailable change in scientific evidence that (2) would have likely yielded an acquittal if evidence of that change had been offered at trial. Applicant's claim fails the unavailability requirement because relevant scientific evidence challenging SBS was available before his trial and before the filing of each of his subsequent writs. And it fails the materiality requirement because his habeas evidence would have been incompatible with his defense strategy at trial.

A. Evidence Challenging SBS Was Available Before Applicant's Trial

Applicant's trial started in late 2002 and ended in early 2003. The record shows that his -03 evidence challenging SBS was available before trial because (1) his -03 habeas experts said so, (2) such evidence was offered in Roark's trial in 2000, (3) Applicant's trial attorney talked about the SBS controversy before trial with the medical examiner, Dr. Jill Urban, and (4) one of Nikki's treating physicians, Dr. Janet Squires, testified during trial about some aspects of the controversy.

In his -03 writ, Applicant offered testimony or declarations from the following experts: forensic pathologists Drs. John Plunkett, Harry J. Bonnell, Janice J. Ophoven, and Carl Wigren; forensic neuropathologist Dr. Roland Auer; and biomechanical engineer Dr. Kenneth Monson. Collectively, they challenged SBS by testifying that subdural hematoma (“SDH”), retinal hemorrhages (“RH”), and brain edema—the so-called triad symptoms—are not diagnostic of head injury, such findings can result from other causes, shaking alone cannot cause SDH or RH, and lucid intervals can follow a fatal head injury. They offered other explanations for Nikki's death, including a short fall, pneumonia, drug toxicity, a bleeding disorder, or a combination of these. They disputed that she suffered more than one impact to her head, argued that her condition could have been caused by a short fall, and maintained that many of her injuries—e.g., bruises, bleeding under the scalp, torn frenulum—were the result of medical intervention.

The State countered with Urban, the medical examiner who performed Nikki's autopsy and testified at Applicant's trial, and Dr. James Downs, an anatomic, clinical, and forensic pathologist who specializes in research of child head injuries. Urban and Downs maintained that Nikki suffered multiple head impacts, her injuries and condition were inconsistent with a short fall, and she did not suffer from or die of a disease process, drug toxicity, or a bleeding disorder.

1. Applicant's Experts Demonstrated Evidence Was Available before Trial

Plunkett published his articles relating to “short falls, lucid interval, SDH, and RH” in 1999 and 2001; he stated that his 2001 study “disproved” several assertions.² He referenced multiple articles published before Applicant's trial to support his opinions.³ He noted that the literature and understanding have been “consistently evolving” and that “these views” were in the minority so “experts willing to examine cases and testify in a manner consistent with this scientific knowledge were few and far between at that time.” He did not say such experts were unavailable.

Monson described the first biomechanical study of shaking as a mechanism of injury conducted “well before Applicant's trial in 2002-2003” which suggested that shaking is an unlikely cause of brain injury, though “this work was dismissed by many.” Monson also cited a “review article discussing shaking in the context of biomechanical principles” published by a neurosurgeon in June 2002, an article also cited by Plunkett.⁴

Three of the four articles that Monson cited as support for the fatality of short falls were published before Applicant's trial.⁵ In addition, Monson published two articles before trial relating to his research interests, which he described as “Biomechanics of injury, particularly traumatic brain injury: pediatric head injury, accidental and intentional”; “Role of cerebral vessels in the structural response of the brain; vascular mechanotransduction and remodeling”; and “Physical experimentation; computational modeling for investigation of soft tissue mechanics; head injury.”⁶

Bonnell declared that “it was known in the medical literature at the time of [Nikki's] death that increased intracranial pressure (as was present in this case), could cause retinal hemorrhages.” He also said he would have been available to testify about his opinion at Applicant's trial.

Ophoven testified that, in 2002 and 2003, “a debate was beginning in the literature” about RH. But she also testified that the debate began before that. She disagreed that the change began with Plunkett in 2001, saying,

There have been a number of individuals starting, I believe, in Bexar County, Texas. Dr. DiMaio used to say that there hadn't been a case of shaking in his jurisdiction for over 25 years because the folks in that office didn't -- in San Antonio -- realized many, many years ago that the science didn't support the theory.

When asked, “Now, we've known for a long time that those three things don't exclusively determine it to be from a shaking incident; haven't we?” Ophoven said that “it depends on who you ask.” And she confirmed that it has been “known for a long time that these things, when looked at separately, can be caused by other things.” She also agreed, “We've known for a long time that [RH] can be a result of something else.” She said that her opinions about various causes of RH “are opinions that have been repeated and substantiated in the literature for many, many years.”

Auer's report stated that “[b]eliefs in 2002 were that shaken baby syndrome had some problems revealed by physics, therefore proponents of SBS in the medical community thought it best to invoke impact as well as shaking.” He confirmed that “a very small number of pioneers” were questioning diagnosis by the triad in 2002. He asserted, “It was not until 2017 that shaking was debunked.”

2. Roark Demonstrated Availability before Applicant's Trial

Roark was tried in March 2000. He defended himself on grounds that challenged SBS. At trial, neurosurgeon John Dietze testified that he treated a chronic SDH when he drained blood from the complainant's subdural space, implicating lucidity after head injury and rebleed. Roark, 707 S.W.3d at 170. Dr. Allen Marengo-Rowe, a pathologist and hematologist, testified that people are susceptible to bruising or rebleeding when the fibrinolytic system is engaged in dissolving clots, that diffuse axonal injuries (“DAI”) can cause rapid loss of consciousness whereas an SDH might not exhibits symptoms for weeks, and that RH is common with SBS but can result from other causes too. Id. at 168. Bexar County's Chief Medical Examiner, Dr. Robert Bux, did not believe a person could shake a child hard enough to cause DAI, and DAI caused a rapid onset of obvious neurological deficits. Id. He said that rebleeds can happen spontaneously or because of minor trauma like falling in a bathtub and can commence up to 72 hours before loss of consciousness. Id. at 169.

Roark's mother, a registered nurse, seemed attuned to the debate about SBS in 1997 when the complainant was injured. Id. at 170. She testified that she made a point of telling Dietze at the hospital that the complainant had fallen and hit her head two weeks before, and she claimed that she asked Squires if Squires would investigate possible causes of injury other than SBS but was rebuffed. Id.

3. Applicant's Trial Attorney Questioned SBS Before Trial

One of Applicant's trial attorneys raised questions about SBS in a pretrial phone call with Urban. The attorney expressed his understanding that “pathologists and forensic physicians don't like the term ‘shaken-baby syndrome’ ” and that he had been “reading some of the literature with respect to rotational forces and the application of tort [sic] force ․ that's developing.” When asked for an example of what a massive, passive rotational force would be, Urban answered, “Again, being struck with something or struck against something.” Defense counsel referenced the literature he had read and asked about hemorrhaging in the eyes being “[t]he other indicator[ ] ․ of impact like this.” The attorney expressed his understanding that “some of the other indications of rotational force are injuries to the neck, cervical, bones, things of that nature.”

At one point, Urban told him that “there's a lot of argument within the forensic community as to whether it's—purely shaking can cause these injuries or whether you actually have to have an impact.” She explained that the term shaken impact “kind of includes a spectrum of both shaking injuries and a blunt force impact injury directly to the child's head.” She told him that injuries like those suffered by Nikki are not “typically” seen from short falls but did not rule out the possibility.

4. Trial Testimony Raised SBS Debate

Squires testified at trial about the debate surrounding SBS. When the prosecutor asked about the onset of symptoms, she said “it's a spectrum.” She testified about the disagreement as to shaking alone as a mechanism of injury. When describing the shearing forces that impact the brain through shaking or shaking with impact, Squires testified that “[t]here are some experts that think that you cannot kill a child by shaking alone.” She said that she “would consider most of the experts do think that shaking alone, if done vigorously, will kill a child, but most children are shaken and then thrown against something.”

Evidence challenging SBS was available before trial, so Applicant failed the first requirement for relief on an Article 11.073 claim—previous unavailability.

B. SBS-Challenging Evidence Would Not Have Been Material

Applicant raised seven specific complaints about SBS, and he argued the materiality of them all. His complaints were as follows: (1) diagnosis of SBS based on a triad of symptoms is illegitimate; (2) shaking alone is insufficient to cause pediatric head injury; (3) fatal shaking causes neck injuries, and Nikki had none;⁷ (4) short falls may be lethal; (5) Nikki suffered a single impact to her head; (6) lucid intervals after head injury are possible; and (7) there are other more likely causes of death. But only previously unavailable evidence counts in an Article 11.073 materiality analysis.

Of his seven complaints, the first—the legitimacy of diagnosing SBS based on a triad—is the only one that might have been previously unavailable. But Nikki was not diagnosed that way, and no one testified at trial that the three symptoms—RH, SDH, and brain swelling—were caused only by shaking or that their opinions about inflicted trauma were based on the triad. In other words, evidence disproving the legitimacy of the triad-based diagnosis of SBS would not have improved his chance of acquittal because the triad was not an issue at his trial.

Squires based her opinion of “non-accidental inflicted trauma” on the brain findings, the bruising, and the torn frenulum, collectively. Urban's conclusion about blunt force injuries as the cause of Nikki's death was based on the multiple and distinct impact injuries that she identified. Dr. John Ross, the pediatrician called to the ER, said that the injuries he saw on Nikki were inconsistent with a fall, and he believed they were intentionally inflicted. Dr. Thomas Konjoyan, the ER physician, testified that her condition did not “fit the picture of a virus that went bad quickly” and was inconsistent with a fall.

Furthermore, Applicant's habeas evidence questioning SBS did not account for the impact injury to the back of Nikki's head. His experts rejected shaking alone as a mechanism of injury, but they did not address shaking with impact. On the contrary, Ophoven conceded that shaking alone could not cause an impact site to the back of the head: “Well, not unless something else happened to cause an impact. Obviously, shaking doesn't cause bruising to the scalp.” She also testified, “An impact death is the consequence of injury to the brain that either immediately causes death or instigates complications that are incompatible with life, which is what occurred in this case, severe brain swelling with increased intracranial pressure.” She just attributed Nikki's impact to a fall rather than Applicant “losing it.”

Similarly, Wigren asserted that “it's literally impossible to break those little, you know, bitty veins with shaking” but acknowledged that “you can break those -- those bitty veins when there is an impact.” He suggested the impact could be “from like, let's say, a short fall, and in rare instances that can occur” instead of an act of “an angry, out of control parent” but also stated, “Usually, it's blunt force injury that will cause those veins to rupture.”

Even Auer, who dismissed the bump on Nikki's head as a “boo-boo” or “goose egg,” agreed that impact can be lethal.

In short, Applicant's experts did not dispute the fact that Nikki suffered an impact injury or that an impact to the head can be fatal. They disputed only its cause—a fall versus inflicted trauma. But the defense at trial conceded that the fall story had been a fabrication and that Applicant “lost it” with Nikki. He contested his mental state but not his acts of shaking and beating Nikki. Given that defense, evidence about the potential lethality of short falls would not have improved his chance of acquittal, nor would the remainder of evidence about the SBS debate.

Applicant has argued, however, that we should look “beyond the impact of the evidence reflected in the trial record and instead [consider] how the absence of the new evidence ‘affected the preparation and presentation’ of the case,” citing Thomas v. State, 841 S.W.2d 399, 406 (Tex. Crim. App. 1992). But the state of the evidence at trial—not an applicant's case preparation or case presentation—is Article 11.073’s materiality focus. And even if it were not, it would be impossible to assess the effect of this evidence on trial preparation because we have not heard from Applicant's trial attorneys—and their absence from this litigation is conspicuous.

This is not a case where a change in a field of science discredited the scientific evidence presented at trial. See, e.g., Ex parte Chaney, 563 S.W.3d 239 (Tex. Crim. App. 2018) (holding that the field of scientific study regarding human-bitemark identification had evolved since trial almost completely discrediting the State's bitemark evidence, which was an essential part of its case). At most, the habeas evidence discredited science suggesting that shaking alone causes fatal pediatric head injury. But Applicant's conviction rested on both shaking and multiple impacts—“shaking” and “beating.” No habeas evidence contradicted the lethality of a blunt-force-impact head injury or the fact that Nikki suffered a head injury.

Nor is this a case where a testifying expert later changed his opinion. See, e.g., Ex parte Robbins, 478 S.W.3d 678 (Tex. Crim. App. 2014) (discussing medical examiner's change of opinion about the cause of death); Ex parte Henderson, 384 S.W.3d 833, 834 (Tex. Crim. App. 2012) (discussing medical examiner's change of opinion about the lethality of short-distance falls and the manner of death). Urban did not recant or contradict her trial testimony but reaffirmed it at the habeas hearing. Even after reviewing the reports by Applicant's experts and considering the evidence offered at the habeas hearing, Urban was steadfast. She would still classify Nikki's death as a homicide due to blunt force head injuries.

Accordingly, Applicant's -03 habeas evidence was not material under the terms of Article 11.073.

C. Later Applications Challenging SBS

Applicant raised his SBS claim in other writ applications after the -03. They recycled the -03 evidence and sometimes attached affidavits, articles, letters, transcripts, etc., as exhibits in a further effort to challenge SBS. To the extent that the exhibits were scientific evidence—and many were not any kind of evidence—they suffered the same defects as his -03 evidence. First, they did not show unavailability but repeated or amplified the scientific evidence Applicant developed in his -03 writ. See Tex. Code Crim. P. art. 11.073 (d). Second, they were immaterial under the terms of the statute because they were incompatible with the defense theory at trial. See id. art. 11.073 (b)(2).

In his -04 writ, Applicant presented three affidavits; one each from radiologist Dr. Julie Mack, pathologist Dr. Francis H. Y. Green, and medical toxicologist and emergency-medicine physician Dr. Keenan Bora. Mack's affidavit repeated findings and opinions she offered in her earlier report admitted in the -03 hearing. And Green and Bora shared opinions already given by Auer and Wigren in the -03. Applicant did not show that Mack's, Green's, or Bora's opinions were based on changes in the scientific evidence since the -03 writ was filed, a requirement for a subsequent writ raising an 11.073 claim. Tex. Code Crim. P. art. 11.073 (c) (referencing Article 11.071’s subsequent writ bar and requiring unavailability as of date on which an original or previously considered application was filed).

Like the -04, Applicant's -06 writ repeated his previously raised Article 11.073 claim and added more expert opinions and articles but made no showing that the science had changed or was unavailable when the -05 was filed or that the evidence would have been material to the outcome of his trial. He also relied on last year's display of political grandstanding in support of his cause, but that is not evidence.

That brings us to the -05. It offered no additional evidence but cited the evidence from the -03 and -04 writs to argue that his case was identical to Roark's. The majority seems to adopt that position, citing the possibility that Roark is a new legal basis for supporting Applicant's claim. But it is not a new legal basis; it did not forbid testimony about SBS; it applied the law to the facts before it, and those facts are different than the facts we face here.

III. Roark Was Not A New Legal Basis

First, Roark did not forbid testimony about SBS. SBS is a matter of debate—a “very acrimonious” and “very vitriolic” debate according to Ophoven—pushed by “a highly vocal minority” according to Downs. But it is still an accepted diagnosis according to both Monson and Downs. And Monson admitted that the potential lethality of shaking alone would be a moot point here because of the impact to Nikki's head. Universal agreement among scientists is not a requirement for admission of scientific evidence. Wolfe v. State, 509 S.W.3d 325, 340-41 (Tex. Crim. App. 2017). And it is not this Court's job to settle such debates.

Second, Roark’s materiality analysis was founded on agreed findings after the State's experts changed their opinions about the cause of injury to the complainant. “The jury heard dueling experts at [Roark's] trial. Today, the jury would hear consensus on primary issues (such as short-distance falls, shaking causing injury, retinal hemorrhaging, lucid intervals, and chronic rebleeds).” Roark, 707 S.W.3d at 187. That analysis is inapplicable here; neither Urban nor Squires changed her opinion about Nikki's injuries after Applicant's trial.

On the contrary, Urban reiterated her trial testimony at the -03 habeas hearing and explained why she disagreed with Applicant's experts about the cause of death, the cause of Nikki's brain bleeding, and the number of impacts to Nikki's head. Her understanding of the potential lethality of a short fall had not changed since 2002, and she discounted a short-fall explanation for Nikki's death given the multiple impact sites that she saw. She still believed that the motion of the brain within the skull can tear the bridging veins and result in subdural bleeding. She agreed that attributing injury to shaking alone is controversial because it cannot be proven or disproven. But she maintained that Nikki suffered blunt impacts and that shaking could have been a component to her injuries, too. Even after hearing the habeas evidence and what she had learned in the intervening years, she would not change her opinion: Nikki's cause of death was blunt force injuries.

Third, Roark is factually distinguishable from this case, and those factual distinctions limit the reach of its holding:

• The complainant in Roark (“BD”), had no sign of a recent head impact, but Nikki had several.

• BD's CT scan showed chronic SDH, but Nikki's didn't.

• Rebleeding was found to be “crucial” in Roark's trial, Roark, 707 S.W.3d at 175, but it was not an issue here.

• DAI featured prominently in Roark's trial but not at all in Applicant's.

• Regarding RH, Squires testified that BD's hemorrhages were “big blobs of blood,” “classically” associated with SBS, and the habeas court found that the State presented evidence that RH “is almost exclusively a component of child abuse injuries.” But not so here; Squires testified that Nikki's RH was “one more thing that really lets you know that those eyes were being shaken and that the blood vessels broke.” Urban told the jury that “areas of hemorrhage surrounding the optic nerves ․ is something that is typically seen in a blunt force or shaking type of injury.”

• The shaking-versus-impact testimony in Roark's trial was inverse to Applicant's: there the experts testified to shaking and possible impact, but here they testified to impact and possible shaking.

• Roark presented eyewitness evidence of several falls, but Applicant didn't, and his defense conceded that the short-fall idea was a fabrication. Furthermore, the short-fall idea here was refuted by other evidence, including the below-the-hatband impact site on the back of Nikki's head.

• The supposed impossibility of a lucid interval was used against Roark to time BD's injury, but Squires would not be pinned down about timing at Applicant's trial. Urban would not commit either; she said the injuries—not symptoms—would have been immediate, though Nikki would have immediately exhibited a change ranging from “possibly unconscious” to something more minor like muffled cries.

• The new medical opinions in Roark refuted the State's evidence of a shaking injury and directly supported Roark's trial defense that BD's injuries were from an accidental fall several weeks earlier that rebled. But evolving opinions about the SBS diagnostic triad would have had no effect on the State's evidence in Applicant's case because Nikki was not diagnosed that way, and she suffered an impact injury to the head.

• Finally, Applicant's jury heard evidence of an admission: Teddie Cox, his girlfriend, testified that Applicant told her that if he killed Nikki, he didn't remember it and that he “snapped.” But no admissions were offered against Roark.

Roark was decided on the facts presented there, and its reach is limited accordingly. The majority errs to treat it as a new legal basis because its holding was fact-driven.⁸

V. Conclusion

Applicant claims that SBS has been “debunked,” but he is wrong; SBS is a matter of ongoing and acrimonious debate. Contrary to his lament, his Article 11.073 evidence has been repeatedly considered by this Court. Instead of showing the unavailability and materiality of his evidence, however, he tallies the appearance of the word “shake” in the trial record, a tedious and specious irrelevancy. He has never met Article 11.073’s requirements, and his post-03 writs failed Article 11.071’s subsequent writ bar. Roark did not change that reality. We should dismiss his applications, deny his motions, and uphold the trial court's judgment and sentence of death. The Court errs to do otherwise, and I respectfully dissent.

FOOTNOTES

1.   See Tex. Code. Crim. Proc. art. 11.071, § 5(a); id. art. 11.073.

2.   Reference within this opinion to each of Applicant's successive writ applications will be made in numerical order, so as to read “the -03 application,” “the -04 application,” “the -05 application,” “the -06 application,” and “the -07 application.”

3.   In noting this I do not imply that this level of effort and care is unusual in any way, but, rather, that they often go unshown given the challenging timelines associated with matters of this nature and the proclivity to file the kind of late and successive applications the Legislature intended to shield us from with Article 11.071’s successive writ provision.

4.   Ex parte Wood, No. WR-45,746-04, 2025 WL 2148862, at *15 (Tex. Crim. App. July 30, 2025) (Schenck, P.J., concurring in part and dissenting in part).

5.   E.g., Liljeberg v. Health Servs. Acquisition Corp., 486 U.S. 847, 863 (1988) (post judgment relief under federal Rule 60(b)); see generally Thomas v. 462 Thomas Family Properties, LP, 559 S.W.3d 634 (Tex. App.—Dallas 2018, pet. denied); cf. Roberts v. Wal-Mart Louisiana, L.L.C., 54 F.4th 852, 854 (5th Cir. 2022) (judgment not void or open to collateral attack though federal district judge rendering it was disqualified by ownership of stock in party where decision turned on purely legal issue, parties had right to de novo review and no other indication of illicit interest in the case).

6.   In other even more extreme cases, such as a judge habitually taking bribes to fix criminal cases, the U.S. Supreme Court has found the presumption of regularity pierced to the point that discovery is proper. Bracy v. Gramley, 520 U.S. 899, 900 (1997). Likewise, our Texas Supreme Court required the same relative to the judge (since convicted of the same) and a juror in a civil case. Ford Motor Co. v. Castillo, 279 S.W.3d 656, 663 (Tex. 2009). Roberson has made no like showing or adequately explained why the information he now describes as harmful was unavailable for presentation in his previous writ applications.

7.   “To establish an enforceable and workable framework, the Court's precedents apply an objective standard that, in the usual case, avoids having to determine whether actual bias is present. The Court asks not whether a judge harbors an actual, subjective bias, but instead whether, as an objective matter, ‘the average judge in his position is ‘likely’ to be neutral, or whether there is an unconstitutional ‘potential for bias.’ ” Williams v. Pennsylvania, 579 U.S. 1, 8 (2016) (citing Caperton v. A.T. Massey Coal Co., Inc., 556 U.S. 868, 881 (2009)); see In re Murchison, 349 U.S. 143, 136–37 (1955).

8.   Applicant also alleges that Judge Bentley expressed bias during and after trial. In particular, Applicant cites Judge Bentley's statement at trial regarding sexual assault allegations and the prosecution's burden of proof. He also cites a statement attributed to Judge Bentley published in a magazine suggesting Judge Bentley enjoyed being a part of Roberson's trial and particularly appreciates presiding over death penalty cases. Putting aside the obvious timing concerns with this argument appearing at this stage in the case despite its operative facts being known for decades, neither statement reflects either the deep-seated animus or the extrajudicial source typically (if not invariably) necessary to state a viable recusal, much less due process, issue. See Lietky v. United States, 510 U.S. 540, 550-51 (1994); United States v. Grinnell Corp., 384 U.S. 563, 583 (1966).

9.   Our decision in Ex parte Elizondo recognized a constitutional claim for actual innocence applicable to capital and non-capital cases alike using this burden of proof for Applicant. 947 S.W.2d 202, 209 (Tex. Crim. App. 1996). As I discussed in Wood, I see no legislative effort (or authority) to bar access to this Court to establish entitlement to relief under it. Instead, I believe the claim is more properly recognized under the Texas Constitution and in capital cases involving a death sentence, proof of innocence by a preponderance would be more appropriate.

10.   I do not read our opinion in Roark as resolving the question of whether it is ever safe to shake a baby, but only whether expert evidence of a syndrome by which a series of symptoms are present without further evidence of physical injury is in fact evidence of guilt of abuse and reliable to the point of sustaining a criminal conviction under Article 11.073. We are not the first court to find evidence of this type has a tendency to corrupt the truth-seeking process. Cf. State v. Saldana, 342 N.W.2d 227 (Minn. 1982) (expert opinion on abuse accommodation syndrome suggesting that conduct of complainant appearing to undermine credibility of the accusation was in fact proof of guilt); Commonwealth v. Gallagher, 519 Pa. 291, 547 A.2d 355 (1988) (same).

11.   Roberson's -05 application includes the testimony of Dr. Green, who concluded Nikki's death was caused by a “severe, undiagnosed viral pneumonia” and complicated by a secondary necrotizing bacterial bronchopneumonia several days to a week before death, which may have been hastened by the respiratory-suppressing medications prescribed by physicians unaware of her pneumonia. He further opined the subdural blood observed in the autopsy is explained by the oxygen deprivation Nikki experienced due to her pneumonia, and found no basis for suggesting the subdural bleeding and brain swelling was caused by trauma. Jurors would be free to weigh this against any evidence to the contrary.

12.   As noted, because I believe the Constitution provides the same standard and does so regardless of procedural barriers in a death case, I will treat the statute as a redundancy.

13.   For purposes of securing a conviction of murder—as opposed to capital murder—proof of intent to cause serious bodily injury (as opposed to death) would suffice. Tex. Penal Code §§ 19.02(b)(2), 1.06(b) (defining “serious bodily injury” as “injury that creates a substantial risk of death or that causes death, serious permanent disfigurement, or protracted loss or impairment of the function of any bodily member or organ”).

14.   Further expert opinion now suggests that the alleged fall from the bed could have produced the injuries and fatality, though it does not suggest that death would likely result. Given the need for at least a preponderance of evidence of innocence, I do not see this opinion as adding materially to a prospective juror's challenge in determining mental state.

15.   Even as to these first two critical questions, there is considerable strain on the judgment as we attempt to determine what jurors would do without some of the expert opinions. While I see nothing a trial court could add to that inquiry, we are effectively keeping multiple mental plates spinning while we count the number of angels on the head of a pin before we ask whether a jury would find Roberson acted with intent to kill.

16.   As noted, murder is committed “knowingly” when the actor engages in conduct while aware that death is reasonably certain to result from the conduct. Medina v. State, 7 S.W.3d 633, 640 (Tex. Crim. App. 1999).

17.   Repub. Tex. Const. of 1836, art. IV, § 21, reprinted in 1 H.P.N. Gammel's The Laws of Texas 1822-1897, at 1069, 1074 (Austin, Gammel Book Co. 1898).

18.   Not that my view on the wisdom of legislation matters, but science by definition is constantly open to testing and rejection. What defense lawyers peddle as “new and better” science may, despite our declarations of agreement, prove to be tomorrow's junk. Phrenology may yet return some day. Still, for purposes of statutory relief, we do the homework that's assigned. So long as experts are more familiar with the subject matter than lay jurors and trial judges are within their discretion in finding it admissible, we will confront these issues.

19.   Penry v. Lynaugh, 492 U.S. 302 (1989).

20.   As I noted in Wood, the Court may have acted cautiously in Herrera and all like cases. Therein exists, in prodding the state courts to avoid the result so many justices found warranted lending their public voice to, the notion that the Justices were willing to “assume” that executing an innocent person would violate the federal Constitution. Herrera v. Collins, 506 U.S. 390, 417 (1993); id. at 419 (O'Connor, J., concurring) (listing as grounds for such a holding that providing relief from execution to a habeas applicant with a viable claim of innocence is a “principle of justice so rooted in the traditions and conscience of our people as to be ranked as fundamental.”).

21.   This standard, and this one alone, in my view would amount to a functional equivalent of acquittal precluding a new trial. See Smith v. State, 499 S.W.3d 1, 6 (Tex. Crim. App. 2016) (describing instructed verdict as a vestige of common law). I would find Mrs. Smith to meet it.

1.   To illustrate my point, consider the following arguments of counsel for both the State and the Defense made just before the jury retired to deliberate Applicant's guilt or innocence:Arguments by the State in the opening part of its closing argument:“I anticipated ․ that we were going to have a battle with defense counsel as to what exactly happened, as to whether or not, as Mr. Roberson said in his confession, [the child victim] fell off the bed.” RR Vol. 46, p. 15.“Well, the defense counsel, they conceded that point in their opening statement. You heard defense counsel, you heard [Defense Counsel] say that this is a shaken baby case.” RR Vol. 46, p. 15.“Beating that child, for you to find him not guilty of capital murder, you would have to believe that in his beating that 28-pound child, at his size he wasn't aware that that might cause her death. That's what you'd have to find. He wasn't aware that that would cause her death, the beating of [the child victim].” RR Vol. 46, p. 18.“And he came up with this story that we've heard. And there are different variations of it, but it's essentially the same. ‘Everything was fine. We were watching movies. She fell asleep gently in my arms. I was awakened at five a.m., and, low and behold, the baby is crying on the floor, blood on her face, from this 22-inch fall.’ * * * Not only was she blue, but her face was battered. You heard the evidence that those marks began to appear even more throughout the treatment. She was bloodied, facial bruising, bruising of the shoulders. You heard the testimony that the back of her head was mush. That was the description. Mush.” RR Vol. 46, p. 20.“You heard ․ testimony that injuries ․ injuries to the severity that [the child victim] received are normally seen in car wrecks, when small children are brought in, car wrecks. That's the type of force that Mr. Roberson put [the child victim], his 2-year-old daughter, through.” RR Vol. 46, p. 21.“You heard from Dr. Squires ․ And you heard her testimony that these were inflicted injuries consistent with not just shaken baby syndrome, this is not a child that was just shaken out of frustration, but shaken impact syndrome. This is a child that was not only shaken, but was beaten about the head. Child abuse, she ruled. You heard from Dr. Urban, the Medical Examiner. Not just shaken, but blunt force injuries to [the child victim], received multiple blows to the head. Multiple blows to the head. Not just, ‘I lost it’, you know, ‘Please be quiet.’ Sits her gently on the bed. But we're talking shaking and beating is what [the child victim] sustained.” RR Vol. 46, p. 25.Arguments by the Defense in their closing argument:“Yes, this is a shaken baby case, but no, this is not a murder case.” RR Vol. 46, p. 34.“․ this case is being presented to you with an acceptance of responsibility, not dodging it, and an acceptance of appropriate responsibility.” RR Vol. 46, p. 41.“Dr. Squires was asked by the State and by us whether this case was a classic shaken baby case. And it is.” RR Vol. 46, p. 42.“Your act is bad. Yes; we do not deny that. But you have not formed a specific intent to murder.” RR Vol. 46, p. 43.“The State has alluded that there's been an ongoing and escalating series of abuse administered to this child. The evidence isn't there to support that, not credible.” RR Vol. 46, p. 45.“Shaken baby syndrome is what applies. It's got a new name of sudden impact, but it's still the same thing as noted by both the expert witnesses. And the one thing I will agree with what [the Prosecutor] said, ‘Listen to the expert witnesses. Evaluate their testimony.’ And if you do, it's clear that this was not an act of murder.” RR Vol. 46, p. 51.Arguments by the State in the final part of its closing argument:“Knowingly is essentially he's beating his child that weighs 28 pounds and he wants you to believe that this man didn't reasonably know that that could cause that child's death. He did know. * * * ‘I'm reasonably certain this will kill my 2-year-old daughter, slam, slam.’ ” RR Vol. 46, p. 54.“But how violent would a struggle be between a 230 pound man and a 28 pound baby? What would you expect? A lot of fight back from [the child victim]? He wants you to believe it's a pillow fight. Talked about all the things that could have caused these injuries. Well, could have been impact on a mattress. Could have been a pillow. This wasn't a pillow fight. Are we going to believe that Mr. Roberson in his fit of rage, ‘I'm really going to shake you like this, but I'm going to make sure that your head hits the pillow or the mattress. I don't want to cause too much damage. I want to fit into that category of criminally negligent homicide.’ ” RR Vol. 46, p. 57.“But are we talking about an either or? Are we talking about he just shook her? You know, an out of control parent is frustrated because their child is sick and they pick them up and they shake them, shake them. Is that an out of control parent? No. He did this and then he starts punching them he slams her down and he throws her on the ground. And that's not intentionally or knowingly? Talked about her battered face. Did he just shake her? No. Frenulum is busted, ripped, torn, blood coming out. Multiple impacts to the head. You saw bruising to the shoulder. Take a look at her face on the photographs. This is not just, ‘I shook her and then I realized, well, maybe I shouldn't be doing that because they may file some sort of criminally negligent homicide, so I stopped.’ * * * Dr. Urban; multiple impacts to the head.” RR Vol. 46, p. 61.“You've seen the autopsy report. The experts agree. There's seven doctors that wrote off on it as homicide. Intentionally inflicted injuries is what they characterize it as. Shaken impact syndrome. Multiple blows to the head. Not as ‘I'm out of control.’ It's an intentionally and knowingly produced injury.” RR Vol. 46, p. 63.“Fact of the matter is, simply put, he just beat her to death. * * * He beat this child to death just like he said he was going to.” RR Vol. 46, p. 64.“And the more she cries the madder he gets. And he picks her up and if you look at the bruising on her face, I think he just absolutely punched her in the face, slapped down on her face with his hand, busted her frenulum. And you look where the bruises are on her face. And he grabbed her face and he said ‘Shut the hell up.’ And pounded her head on whatever. And she kept crying. And did he stop then? No, he sure didn't. He picked her up then and he violently shook her like has been described. On and on and on and nobody's there to break it up. That's what he did. And then he threw her down on the floor where he says he found her. * * * And the last thing that she saw before he killed her was the hate in her dad's eye when he was shaking her to death is what she saw.” RR Vol. 46, p. 66.“․ but we can sure hold people like Mr. Roberson accountable for what he did to his own little daughter in beating her to death.” RR Vol. 46, p. 68.

1.   Roberson v. Dir., TDCJ-CID, No. 2:09CV327, 2014 WL 5343198, at *62 (E.D. Tex. Sept. 30, 2014). The Court of Appeals for the Fifth Circuit granted a certificate of appealability as to Applicant's third ground, that “the trial court's exclusion of his expert witness's testimony violated his due process rights and his right to present witnesses in his own defense.” Roberson v. Stephens, 614 F. App'x 124, 136 (5th Cir. 2015). After briefing, the Fifth Circuit affirmed. Roberson v. Stephens, 619 F. App'x 353, 359 (5th Cir. 2015). Applicant petitioned the Supreme Court to grant a writ of certiorari, but the Supreme Court denied Applicant's petition. Roberson v. Stephens, 577 U.S. 1033, 136 S. Ct. 607 (2015).

2.   Judge Parker sees it differently. Post at 1 (Parker, J., concurring and dissenting) (“Applicant's case is not a ‘shaken baby’ case ․ This case is a ‘blunt-force trauma’ case.”) (original emphasis). But the repeated references at trial about shaking belie this assertion.

3.   Judge Parker highlights this “previous history of abusive conduct” to show why this was not a “shaken baby” case. See post at 4–5 (Parker, J., concurring and dissenting). But the State had to use this evidence to prove its shaken baby theory. As Dr. Squires testified on re-direct examination:Q: How does that pattern of violence— Is there a typical of way it goes? Does it escalate, the violence?A: It's like-- I'm not sure how to say it. From my literature and my assessment and my experience doing this for 10 years, most of the time this is not an isolated event. It's not the first and only time something like this has happened. Typically a violent environment where people consistently do it and then one day it goes too far. So, it is my experience and in all the literature it is rare that it's a single event that this wonderful, caring, loving person suddenly one day does something like this.Based on Dr. Squires’ testimony alone, it would have been extremely difficult for the State to prove its theory of the case without introducing additional instances where Applicant was seen shaking Nikki.And Applicant's admission to a fellow county jail inmate, Ryan Lodygowski, that he hurt Nikki, is equally unpersuasive. As Detective Wharton stated, Lodygowski was “known to be unreliable” and “was obviously problematic.” Statement of Brian Wharton, Oct. 23, 2024.

4.   Judge Parker asserts that “ ‘shaken baby’ syndrome was Applicant's defense.” Post at 4 (Parker, J., concurring and dissenting) (original emphasis). She argues that “[c]asting doubt on ‘shaken baby’ syndrome actually hurts [Applicant's] case because it would make a jury even less likely to believe that the child's death was unintentional.” Id. (original emphasis, bracketed text inserted). But this argument misses the point for two reasons. First, the State repeatedly argued—and relied upon testimony—indicating that this was a “shaken baby” or “shaken impact” case. Second, Dr. Urban herself testified that it would be impossible to differentiate between evidence of shaking and evidence of actual impacts. Judge Parker would have the Court—and the jury—separate out evidence of shaking, contrary to what Dr. Urban testified was possible. Contra post at 14 (Keel, J., dissenting) (“But Applicant's conviction rested on both shaking and multiple impacts—‘shaking’ and ‘beating.’ ”). As discussed later, this is precisely why Applicant's Article 11.073 claims should be remanded for evidentiary development.

5.   This Court relied on Dr. Plunkett's expert testimony in Ex parte Henderson, 384 S.W.3d 833, 840–42 (Tex. Crim. App. 2012).

6.   Dr. Urban, relying on the old SBS scientific understanding, interpreted the absence of neck injuries as supporting the shaken baby theory.

7.   See Appendix A, State's Exhibit 29.

8.   Judge Parker asserts that the evidence at trial showed that “Nikki suffered from five different types of bleeding,” one of which was a “subdural” bleed “[i]nside the skull.” Post at 2 (Parker, J., concurring and dissenting). Dr. Green's affidavit provides a non-abuse explanation for why this bleed was observed during Nikki's autopsy.

9.   The records relied upon by Dr. Mack were part of records that were discovered in the courthouse basement in August 2018, some twelve years after Applicant's trial. Coincidentally, the discovery of the records was on the same day as the evidentiary hearing on Applicant's -03 hearing was supposed to begin.

10.   Judge Parker asserts that “Nikki had extensive injuries to the face and the back of the head, which were described as ‘massive’ and ‘very extensive,’ ” including bruising on her chin, the left side of her jaw and cheek, the right side of the forehead, the back of her head, and just inside each ear. Post at 2 (Parker, J., concurring and dissenting). However, as highlighted by Dr. Bora, sepsis could cause bruising similar to what Judge Parker asserts is evidence of guilt.

11.   Judge Parker highlights this “extensive bleeding all throughout the head” as evidence of physical abuse. Post at 3 (Parker, J., concurring and dissenting); see also id. at 2 (describing “five different types of bleeds”) (original emphasis). But as Dr. Auer explains, “subdural bleeding is reflexively associated with trauma, but here is due to a blood clotting disorder and reperfusion, both factors acting in concert.” Judge Parker argues that “the multiple types of internal bleeding ․ lend further support to the notion that a great deal of force was exerted against the child's head.” Id. at 3. But Dr. Auer—and several of Applicant's other experts—refute this assertion. Dr. Auer concluded that trauma could not be “the only cranial consideration for the oozing hemorrhages” because of “the absence of bruising or fracture.”

12.   The first nurse to see Nikki, Kelly Gurganus, testified that she only saw a minimal bruise that she described as looking like a handprint, and saw no evidence of black eye, blood or fracture. Dr. Squires—who saw Nikki upon her arrival to Palestine— testified that Nikki had “minimal bruising” and a “little chin abrasion” but “no scars, no unusual bruising or anything.”

13.   Dr. Bonnell concluded similarly, noting that “[t]he injury to the frenulum ․ could have occurred during intubation attempts and this complication is well-described in the literature.”

14.   Judge Parker asserts that evidence of Nikki having “extensive injuries to the face,” including “a bruise on her chin, bruising along the left side of her jaw and cheek, a bruise on the right side of the forehead, a scrape next to the left eye, [and] a scrape on the top of the head,” is evidence of physical abuse. Post at 2 (Parker, J., concurring and dissenting). But as Dr. Laposata concluded, as well as Dr. Auer, supra at 31–32, these minor injuries are attributable to the treatment Nikki received at the hospital, not physical abuse. For example, what Judge Parker described as injuries to the top of Nikki's head is attributable to the site where the pressure monitor was surgically inserted and later removed before Nikki was transported for autopsy. See also supra at 31, n.14.

15.   Judge Parker argues that Dr. Urban's trial testimony concluded that the “extensive bleeding all throughout the head” was a result of “multiple impacts over the entirety of the head.” Post at 3 (Parker, J., concurring and dissenting); see also post at 12, 14 (Keel, J., dissenting) (evidence of “multiple impacts”). But as Dr. Laposata concludes, the “contrast” between the amount of blood during the initial CT scan and the autopsy “is not consistent with a theory that Nikki's intracranial bleeding was caused by multiple violent impacts to the head.” Consequently, Dr. Urban's testimony was “misleading” because “the additional intracranial bleeding seen at autopsy is consistent with DIC and is not consistent with trauma.”

16.   Judge Parker asserts that “[t]he presence and severity of external injuries—that is, injuries outside the skull—debunk any notion that the homicide finding in the present case depends on the vitality of the scientific theory of shaken baby syndrome. Producing these external injuries required forceful contact with a hard surface.” Post at 3 (Parker, J., concurring and dissenting). Setting aside the fact that the evidence from Applicant's experts uniformly reject the notion that there were “external injuries,” Dr. Laposata concluded that even “minor trauma”—such as a fall from the bed that resulted in Nikki hitting her head on either the floor or the cinder blocks holding the mattress, as Applicant admitted—would be sufficient to lead to “major bleeding” in a chronic DIC patient like Nikki. Judge Parker argues that the external injuries “lend further support to the notion that a great deal of force was exerted against the child's head.” But expert testimony based on post-2002 research completely undermines this assertion.

17.   Also included in Applicant's -06 writ is a joint statement by ten pathologists that state that the signatories “are deeply troubled by the failure to revisit an autopsy performed over 20 years ago.” The pathologists reach the conclusion that the autopsy “is not reliable” and “does not comport with the current standard of care in our field.” They argue that “the conditions Nikki seems to have had at the time of her death, including an undiagnosed viral and bacterial pneumonia that had likely caused DIC and sepsis, have been confused with SBS/AHT.”

18.   This confession also was not before the jury when it decided whether Applicant was guilty—Goodness testified to this during punishment. Furthermore, as she herself conceded, this confession did not happen until after she pressured him: “he told me he didn't remember. And after I convinced him that was not going to fly with me, then he told that he lost it.” Even if this is construed as a “confession,” both courts and the research community have rejected confessions in the context of both parents in shaken baby cases and emotionally-challenged individuals. See Aleman v. Vill. of Hanover Park, 662 F.3d 897, 907 (7th Cir. 2011); Richard A Leo & Deborah Davis, From False Confession to Wrongful Conviction: Seven Psychological Processes, Psychiatry & L. 9, 38–40 (2010).

19.   I pause briefly to address a lingering concern I have over a tangentially relevant piece of evidence that was seen in both Roark and in this case: the evidence of uncorroborated sexual assault. Contra post at 5 (Parker. J., concurring and dissenting) (“Both children had some injuries that might be indicative of sexual assault but ultimately could not medically be ruled as such.”).The State's alternate theory of capital murder in this case was murder in the course of committing or attempting to commit the offense of aggravated sexual assault. See Tex. Penal Code § 19.03(a)(2). From opening statements, the State primed the jury with evidence that Applicant had sexually assaulted Nikki.The only evidence of sexual assault came from Andrea Sims. She testified to the jury that she worked as an “emergency room staff nurse and a Sexual Assault Nurse Examiner.” Yet on cross-examination, Sims admitted that she was “not actually certified as a SANE nurse.” Sims told the jury that she did “this sexual assault exam on the request of the Palestine Police Department, and they asked me to do one.” Yet Detective Wharton, the lead detective, later stated “[Sims] was not asked to [conduct a sexual assault exam] by me or any member of my team that I am aware of.”Sims testified to seeing “fresh tears” to Nikki's “anal area,” but admitted that “the pictures really don't show the tears that I was able to see.” Dr. Squires testified that based on the available evidence, she was not prepared to make a medical conclusion about sexual assault.Unsurprisingly, at the close of its case in chief, the State abandoned this allegation. Yet, even after the trial court denied Applicant's motions for mistrial and direct verdict, the State persisted to argue in closing:You heard from Andrea Sims that there was a probable sexual assault. Not only was she the Sexual Assault Examination Nurse, but she was also a registered nurse that was working in the emergency room that day. She saw evidence of three anal tears. You heard her conclusion, probably sexual assault.Detective Wharton put it best: “One cannot put the bullet back in the gun and expect a jury to be objective.” And the trial court recognized this very issue when it stated: “You know when you get off into sexual assault herein lies the problem. If you produce evidence, but it doesn't rise to the level to carry your burden on that, you know, we've built error into the case already.”

20.   Judge Keel says that Roark’s reach must be “limited to” its facts. Post at 19 (Keel, J., dissenting). But, this Court in Roark simply applied evolving science to Roark’s facts, not the other way around. We recognized a “marked shift” in the scientific community and “multiple similar cases” in SBS, id. at 185, 187, and did not simply limit Roark to the very facts presented in that habeas application.

21.   Judge Parker says that “Applicant has been given numerous bites of the apple for approximately fifteen years on his journey to seek relief from his conviction. He has, in fact, had an entire apple pie.” Post at 7 (Parker, J., concurring and dissenting). The reality, however, is that this Court's repeated dismissals of Applicant's claims without considering the merits have barred him from a seat at the table upon which the proverbial apple pie rests.

1.   See 707 S.W.3d 157, 159 (Tex. Crim. App. 2024) (age of the child).

2.   This observation and others about the child's injuries have been taken from the testimony of various experts, including Kelly Gurganus and Andrea Sims (emergency-room nurses who observed Nikki), Dr. Thomas Konjoyan (the emergency-room physician), Dr. Jill Urban (the physician who conducted the autopsy), and Dr. Janet Squires (a pediatric expert who testified for the State at trial).

3.   See id. at 160-62 for the statements in this paragraph as to Roark's case. In Applicant's case, Dr. Squires said that sometimes a special scope was needed to see retinal hemorrhages but that, for Nikki, the retinal hemorrhages were “very obvious” and “[e]verybody could see them.” Dr. Urban testified that it was not unusual for a child with a serious head injury to not have a skull fracture because children “are very malleable.”Nurse Sims, testified that there were fresh tears to Nikki's anal area that were an indicator of sexual assault. But Dr. Squires called it a tiny laceration that was “nonspecific” and that she “did not feel there were any findings” of sexual assault. Dr. Urban testified that she did not see any injuries to the anus. Sims also testified that the anal opening dilated quickly, which was also an indication of sexual assault, but Dr. Squires discounted that fact because laxity in the anal opening meant very little in a comatose child. Sims further testified the child had a torn frenulum, which could also be a sign of sexual assault by forcing something in the mouth, but Dr. Urban testified it could have been caused by a blow to the mouth. Dr. Urban also testified that the semen test of the sexual assault kit was negative and that she and her colleague made no affirmative finding of sexual assault.

4.   Id. at 164.

5.   Id. at 161.

6.   Id. at 161, 180.

7.   Id. at 159.

8.   See Tex. Code Crim. Proc. art. 11.071, § 5.

9.   Many of the observations I make were also made in Judge Yeary's concurring opinion to the denial of Applicant's previous requests for reconsideration of his third and fourth habeas applications. See Ex parte Roberson, Nos. WR-63,081-03, WR-63,081-04, slip op. (Tex. Crim. App. Oct. 10, 2024) (not designated for publication). I agree with his opinion.

10.   943 S.W.2d 469, 474 (Tex. Crim. App 1997).

1.   The remainder of this opinion will shorthand “ascertainable through the exercise of reasonable diligence on or before a specific date” as “available” or a form thereof and “field of scientific knowledge, a testifying expert's scientific knowledge, or a scientific method on which the relevant scientific evidence is based” as “science” or “scientific evidence.”

2.   He stated that he disproved that: (1) “A low-velocity fall cannot cause serious injury or death;” (2) “The location of a subdural hematoma (SDH) indicates whether an injury was accidental versus inflicted;” (3) “A child with an ultimately fatal head injury does not have a period of time during which he/she appears to be ‘normal’, i.e., have a lucid interval;” and (4) “Retinal hemorrhage (RH) with specific characteristics indicates an ultimate cause, i.e., accidental versus ‘shaken baby syndrome.’ ”

3.   Those articles included:• Ommaya AK, Goldsmith W, Thibault LE. Biomechanics and neuropathology of adult and paediatric head injury. Br J Neurosurg 2002; 16:220-42.• Uscinski R. Shaken baby syndrome: fundamental questions. Br J Neurosurg 2002; 16:217-19.• Geddes JF, Hackshaw AK, Vowles GH, Nickols CD, Whitwell HL. Neuropathology of inflicted head injury in children. I. Patterns of brain damage. Brain 2001; 124: 1290-8.• Geddes JF, Vowles GH, Hackshaw AK, Nickols CD, Scott IS, Whitwell HL. Neuropathology of inflicted head injury in children. II. Microscopic brain i1tjury in infants. Brain 2001; 124:1299-306.• Chadwick DL, Chin S, et al. Deaths from falls in children: how far is fatal? J Trauma 1991; 31:1353-5.• Barnes PD. Ethical issues in imaging nonaccidental injury: child abuse. Topics Magnet Res Imag 2002; 13: 85-94.• Fung EL-W, Sung RYT, Nelson EAS, Poon WS. Unexplained subdural hematoma in young children: Is it always child abuse? Pediat Internal 2002; 44:37-42.• Piatt JH. A pitfall in the diagnosis of child abuse: external hydrocephalus, subdural hematoma, and retinal hemorrhages. Neurosurg Focus 1999; 7.• Nassogne M-C, Sharrard M, Lucie Heitz-Pannier L, Armengaud D, et al. Massive subdural haematomas in Menkes disease mimicking shaken baby syndrome. Childs Nerv Syst 2002; 18:729-731.

4.   Ommaya AK, Goldsmith W, Thibault LE. Biomechanics and neuropathology of adult and paediatric head injury. Br J Neurosurg 2002; 16:220-42.

5.   Mertz, H.J. Anthropomorphic Test Devices. In: Nahum, A. M.; Melvin, J. W., eds. Accidental Injury: Biomechanics and Prevention. New York: Springer, 2002, Table 4.7; (2) Melvin, J. Injury assessment reference values for the CRABI 6-month infant dummy in a rearfacing infant restraint with airbag. SAE International Congress and Exposition. Detroit, Michigan: SAE International, Warrendale, Pennsylvania; 1995; and (3) Klinich, K., et al. Estimating infant head injury criteria and impact response using crash reconstruction and finite element modeling. Stapp. Car. Crash. J. 46:165-194, 2002.

6.   Monson, K. L., Goldsmith, W., Barbaro, N. M., Manley, G., 2000. Static and dynamic mechanical and failure properties of human cerebral vessels. In: Crashworthiness, Occupant Protection, and Biomechanics in Transportation Systems 2000, ASME. Ed. by H. F. Mahmood, S. D. Barbat, and M. R. Baccouche. AMO V. 246 / BED V. 49:255-265; and (2) Zhang, L., Bae, J., Hardy, W. N., Monson, K. L., Manley, G. T., Goldsmith, W., Yang, K. H., King, A. I., 2002. Computational study of the contribution of the vasculature on the dynamic response of the brain. Stapp Car Crash Journal 46:145-163.

7.   Strictly speaking, the assertion that Nikki had no neck injury is untrue. Urban found small hemorrhages in Nikki's neck and back muscles.

8.   Applicant's -04 and Roark were pending in this Court at the same time, yet we unanimously dismissed Applicant's -04 shortly before handing down the Roark opinion. Tellingly, the author of the Roark opinion, the Honorable Barbara Hervey, joined the majority in dismissing the -05 that relied on her opinion, and she would have been most acutely aware of her opinion's meaning and reach.