concurring in part and dissenting in part.
I agree with the principal opinion in holding that Mr. Stevens’ trial counsel was *900constitutionally ineffective during the penalty phase of the trial, and, therefore, that Mr. Stevens is entitled to a new sentencing proceeding. I write separately because I believe that counsel’s shortcomings not only affected Mr. Stevens’ sentence, but also his conviction. Therefore, I would grant habeas relief with respect to his conviction as well as his sentence.
1.
As the principal opinion explains, counsel for Mr. Stevens were aware “that a comprehensive mental health evaluation would be a crucial prerequisite for determining trial strategy, yet after Dr. Lennon failed to provide them with such an evaluation they neglected to consult an alternative health expert.” Slip op. at 891. Counsel did not offer any strategic basis for not seeking out additional expert advice, and the State’s proffered explanation for the reasonableness of counsel’s choices does not suffice to justify their actions. See id. The principal opinion, therefore, concludes, correctly, that “[i]n light of Stevens’s lawyers’ admissions and the overall weakness of the defense they presented at trial, we are inclined to believe that their performance was ineffective.” Slip op. at 892.1 The principal opinion turns then to Strickland’s prejudice prong to determine whether Mr. Stevens’ conviction may. have been affected by counsel’s failure.
After reviewing the standards for mental illness under Indiana law and also the standard for demonstrating prejudice under Stnckland> the principal opinion con-eludes that “the Indiana Supreme Court’s conclusion that no jury could conclude that Stevens did not appreciate the wrongfulness of his conduct at the time of the murder was not implausible.” Id. at 893. In other words, the Supreme Court of Indiana’s conclusion — that Mr. Stevens had not suffered any prejudice as a result of his counsel’s missteps — was not unreasonable.
2.
In reaching this conclusion, I believe that the principal opinion reads the decision of the Supreme Court of Indiana too broadly. With respect to the question of whether trial counsel’s investigation of the expert was adequate, the Supreme Court of Indiana concluded: “The trial court determined that defense counsel adequately investigated issues of substance abuse and mental illness and reasonably chose to pursue a different strategy. The post conviction court did not err in denying relief on this claim.” Stevens v. State, 770 N.E.2d 739, 749 (Ind.2002) (footnote omitted). This language speaks in terms of attorney performance — the focus of Strickland’s, first prong — rather than to prejudice — the focus of Strickland’s second prong. Later in its opinion, the Supreme Court of Indiana also addresses counsel’s decision to present a voluntary manslaughter theory as opposed to presenting a mental disease defense. After reviewing counsel’s actions, the state supreme court again uses language that speaks to the performance inquiry of the first prong of Strickland: *901“The post-conviction court found that counsel’s decision to pursue the voluntary manslaughter strategy, while ultimately unsuccessful, did not amount to deficient performance.” Id. at 753. And, again, the Supreme Court of Indiana upheld that decision: “We conclude that the evidence as a whole does not lead unerringly and unmistakably to a decision opposite that reached by the post-conviction court, and we find that defense counsel’s choice of defense theory did not constitute ineffective assistance of counsel.” Id. In sum, the Supreme Court of Indiana rested its rejection of the merits of Mr. Stevens’ ineffective assistance of counsel claim on Strickland’s performance prong; it did not address directly the question of prejudice.
3.
AEDPA instructs that, in the usual case, we apply a deferential standard to a state-court determination that is challenged by way of a habeas petition. See Canaan v. McBride, 395 F.3d 376, 382 (7th Cir.2005) (“Ordinarily, § 2254(d) requires that we determine whether the state court’s decision was ‘contrary to, or involved an unreasonable application of, clearly established Federal law as determined by the Supreme Court,’ or “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.’ ”). “This standard only applies, however, to a claim that was adjudicated on the merits in State court proceedings.” Id. (internal quotation marks and citations omitted). “As a practical matter, a federal court cannot apply the deferential standard provided by § 2254(d) in the absence of any state court decision on the issue.” Id. This principle is equally applicable to the separate inquiries under each prong of Strickland. In the words of the Supreme Court, “our review is not circumscribed by a state court conclusion with respect to prejudice, as neither of the state courts below reached this prong of the Strickland analysis.” Wiggins v. Smith, 539 U.S. 510, 534, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003). Thus, because the Supreme Court of Indiana did not reach the question of prejudice, we need not apply AEDPA deference;2 we look to see *902only whether there is a reasonable probability that, but for counsel’s errors, the result of the trial would have been different.
4.
In this case, the evidence presented at Mr. Stevens’ post-conviction hearing established that Dr. Lennon’s ideas concerning mental illness were “completely abandoned” by the American Psychiatric Association in the 1980s and that “anyone who subscribes to the ‘myth of mental illness’ is not really in the mainstream of current thought among professionals.” Post Conviction Record (“PCR”) at 2015 (testimony of Dr. Kaplan). Mr. Stevens’ counsel “had concluded that Dr. Lennon was [a] ‘quack,’ ” but did not engage in further, critical investigation of Dr. Lennon’s testimony or his theories that would have led them to seek the assistance of another expert. Slip op. at 896.
Had Mr. Stevens’ counsel sufficiently investigated Dr. Lennon’s testimony, and, as a result, secured additional expert evaluation, the jury would have been presented with a psychiatric explanation for Mr. Stevens’ actions. At the state post-conviction hearing, Dr. Kaplan testified that, at the time Mr. Stevens committed the offense, “he was laboring under dissociative disorder not otherwise specified, borderline personality disorder, and intoxication from LSD and the aftereffects of marijuana,” PCR at 2004, and that Mr. Stevens’ capacity to make a personal choice at the time of the crime was “nil,” id. at 2029. Dr. Coons echoed Dr. Kaplan’s diagnoses, see id. at 1885-86, and also concluded that, at the time of the offense, Mr. Stevens was “laboring under extreme emotional disturbance” and was impaired in his ability both to conform his conduct to the requirements of the law and to appreciate the wrongfulness of his conduct, id. at 1891-92. In short, had counsel secured an expert who adhered to mainstream theories within the psychiatric and psychological communities, the expert would have presented evidence establishing the elements of a mental illness defense under Indiana law. See slip op. at 892 (quoting Ind.Code § 35-41-3-6).
Given this evidence, I believe that there is at least a reasonable probability that, had the jury been privy to this evidence, its result would have been different. I must respectfully register my disagreement with the principal opinion that, in these circumstances, there is a fundamental difference between mounting an insanity defense and establishing statutory mitigating factors for purposes of capital sentencing. See slip op. at 897. Here, the experts not only presented evidence that could establish that Mr. Stevens was operating “under the influence” of a mental disease at the time of the murder — a mitigating factor under Indiana’s capital sentencing law, see Ind.Code § 35-50-2-9(e)(2), but also that he was “unable to appreciate the wrongfulness of the conduct at the time of the offense” — a de*903fense to the crime, see Ind.Code § 35-41-3-6.
Additionally, I believe that the principal opinion overstates the downside to presenting a mental disease defense during the guilt phase. The principal opinion states that “[t]he strategic reasons that might, at a stretch, have justified this decision at the guilt phase, fall apart when we consider that at the sentencing phase Stevens had nothing left to lose.” Slip op. at 896. As a practical matter, Mr. Stevens had nothing to lose at the guilt phase. Absent expert testimony that, at the time of the killing, Mr. Stevens’ actions were caused by a mental disorder and he was unable to control his actions, the jury was left with no other alternative than to conclude that Mr. Stevens not only chose to engage in predatory pedophilia, but also that he willingly disposed of his victims as dictated by his own interests.
5.
Furthermore, even if the decision of the Supreme Court of Indiana, either by itself or in conjunction with the post-conviction trial court’s decision, see Stevens, 770 N.E.2d at 749 n. 5, could be construed as reaching Strickland’s prejudice prong, I could not join the principal opinion’s conclusion that the decision of the Supreme Court of Indiana was a reasonable one. The expert testimony during the post-conviction hearing shows that Mr. Stevens’ upbringing was “a recipe for developing someone with a dissociative disorder.” PCR at 2020. Both experts testified that this disorder, in conjunction with Mr. Stevens’ other infirmities, caused him to kill Zachary and prevented him from recognizing the wrongfulness of his actions. Additionally, Dr. Kaplan explained how the nature of Mr. Stevens’ disorders and the events leading to the killing of Zachary were inconsistent with premeditation. Because the testimony of competent experts would have explained Mr. Stevens’ actions in terms of his mental illness and also would have diminished the impact of any inculpating evidence of premeditation, I do not believe that a determination that Mr. Stevens was not prejudiced by his counsel’s lapse of judgment could be characterized as a reasonable one.
Conclusion
For the foregoing reasons, I not only would grant Mr. Stevens habeas relief with respect to his sentence, I also would grant the writ with respect to his conviction. On this ground, I respectfully dissent in part from the judgment of the court.
. As the principal opinion notes, counsel's sole reliance on a voluntary manslaughter defense is troubling for several reasons. See slip op. at 892. First, the facts as presented to the trial court did not justify a voluntary manslaughter instruction. Second, securing a voluntary manslaughter instruction, like pursuing a mental illness defense, would have opened the door to Eastin's testimony. See id. Finally, nothing precluded counsel from pursuing a voluntary manslaughter defense and at the same time presenting evidence of mental illness. Both focus on the defendant’s ability to form the requisite criminal intent to commit murder, and the record does not suggest any justification for only pursuing the former course.
. In addition to considering the decision of the Supreme Court of Indiana, Judge Manion suggests that, in assessing the question of prejudice, we should look at rationale relied upon by the state trial court as well, specifically the state trial court’s "finding that the additional psychological evidence altered the nature of the sexual encounters between Stevens and Zachary.” Concurring op. at 907. Judge Manion acknowledges that "[tjhe Supreme Court of Indiana did not specifically mention the different description of the sexual encounter, but never rejected the trial court’s findings and twice stated that additional evidence 'would have opened the door to the admission of substantial incriminating evidence not otherwise presented during the guilt phase.’ ” Id. at 907-08 (quoting Stevens v. State, 770 N.E.2d 739, 753 (Ind.2002)).
For purposes of our review under AEDPA, the operative state-court decision "is that of the last state court to address the claim on the merits.” Garth v. Davis, 470 F.3d 702, 710 (7th Cir.2006). In this case, the Supreme Court of Indiana, although specifically mentioning other evidence that could have been introduced had Mr. Stevens mounted a mental illness defense (namely Eastin's testimony), did not mention the alternative description of Mr. Stevens’ last sexual encounter with Zachary. This omission may have been inadvertent, or the evidence may have been encompassed by the reference to "substantial incriminating evidence.” However, it also may be the case that the Supreme Court of Indiana did not agree with the trial court's conclusion that the additional facts fundamentally altered the jury’s view of the “relationship” between Mr. Stevens and Zachary. One of the rationales for modern statutory rape laws is that sexual relations with a minor, especially one as young as Zachary, are "inherently nonconsensual.” See, e.g., Daryl J. Olszewski, Comment, Statutory Rape in Wisconsin: History, Rationale & the Need for *902Reform, 89 Marq. L.Rev. 693, 699 (2006). The Supreme Court of Indiana may have concluded that, even absent this evidence that the last encounter was forcible, not consensual, the jury already would have considered Mr. Stevens to be a sexual predator. In short, the jury's view would not have changed with the introduction of this additional evidence.
Thus, even assuming that the Supreme Court of Indiana's commentary on "open[ing] the door to the admission of substantial incriminating evidence,” Stevens, 770 N.E.2d at 753, is meant to address Strickland’s second prong, it is not at all clear that the Supreme Court of Indiana meant to encompass within this statement reference to the state trial court’s findings concerning the nature of the relationship between Zachary and Mr. Stevens. Consequently, these findings should not factor into this court’s AEDPA analysis.