Bauer, Circuit Judge.
Petitioner-appellant Mack Sims seeks a writ of habeas corpus, arguing his due process rights were violated because the state withheld evidence favorable to his case. In November of 1993, security guard Shane Carey was shot in Elkhart, Indiana. Approximately fifteen to twenty minutes after the shooting, the Elkhart police found Mack Sims near a walking path around twenty feet from where the shooting occurred. After Carey identified him at trial as the shooter, Sims was convicted of attempted murder and sentenced to a term of imprisonment of 35 years. In 2012, during a post-conviction evidentiary hearing, Sims learned the prosecution withheld evidence that Carey, the only witness who could identify the shooter, was hypnotized before trial to enhance his recollection of the shooting. After the Indiana courts denied habeas relief, Sims filed a petition for a writ of habeas corpus in federal court. The district court held that the Indiana court did not unreasonably apply established federal law and denied the petition. Because we disagree, we reverse.
I. BACKGROUND
A. The Night of the Shooting
In September of 1993, Shane Carey began working as a security guard at Sister Virginia's Adult Basic Education ("Sister Virginia's") in Elkhart, Indiana. On November 2, 1993, the school was in session when Carey arrived for his shift at 6:00 p.m. Shortly after arriving he traversed the premises and returned to his car to read a book after he was satisfied the school was safe. Sister Virginia's parking lot, where his car sat, had some lighting but was not well lit.
*1081Around 7:00 p.m., Carey noticed three black men walk behind a nearby building and emerge in the parking lot a few minutes later. The men walked toward Carey's vehicle; two walked to the passenger side and one to the driver's side. When about two feet from the car the individual walking towards the driver's side door grabbed a gun from his coat and fired it through the window. Carey did not see the gun, but soon realized he had been shot in the face. He made his way into Sister Virginia's and emergency help was summoned. Carey testified at trial that the shooting took place "shortly after 7:00 [p.m.]."
Officers Tom Lerner and William Wargo arrived on the scene at 7:27 p.m. Lerner instructed Wargo to secure Carey's vehicle and the area around it while he entered Sister Virginia's to speak with Carey. Carey was having difficulty speaking but was able to provide Lerner with a description of the assailant as a black male with short hair or a shaved head and a large build, possibly in his late twenties, wearing a three-quarter-length coat with dark pants and dark combat boots.
At approximately 7:30 p.m. while standing near Carey's car, Officer Wargo heard a noise near the train tracks about twenty-five feet southeast of Carey's vehicle. Wargo then observed a black male crouching behind a dumpster wearing a black three-quarter length jacket and hat looking onto the crime scene. This individual was near a walking path that ran next to the train tracks and was often used by locals. The officers ordered the subject out of the bushes and he came forward without protest. They identified this individual as Mack Sims and patted him down for weapons; none were found.
Although there were civilian witnesses in the area, none were able to give an identifying description of the shooter. Police officers searched the surrounding area extensively but never found a gun, nor did they recover a shell casing from the area or any other physical evidence.
B. The Trial of Mack Sims
On November 4, 1993, Sims was charged with attempted murder. His trial began on August 23, 1994. The prosecution called ten witnesses: six were law enforcement officers that described their role in the investigation and three were individuals present at Sister Virginia's the night of the shooting that were unable to identify the shooter. Thus, the state relied almost exclusively on the only witness who could possibly identify the shooter, Carey, to establish their case against Sims.
Carey testified that he got a good look at the assailant. He described in court what he saw the night of the shooting in more detail than was in the incident report:
Q: Now, as he approached the car, did you get a look at him?
A: Yeah. I was looking him square in the eyes.
Q: Describe what you saw.
A: What I saw was a man with-it was a somewhat full face, well-I'd say a well-rounded face. What I mainly noticed was the eyes. And I noticed underneath the left eye the skin tone, I'd say, or shades were slightly different than the other. One side just underneath the eye was a little bit lighter and the other side was very dark. And I noticed-I did notice the eyes. It was a very cold stare.
Q: Did you see him reach for anything?
A: I saw him make a hand movement, and that's about it. I saw a flash but nothing more.
Q: Okay. Do you recognize in this courtroom today the person you saw and who shot you on November 2nd, 1993?
A: Yes, I do.
*1082Q: Where is he?
A: He's sitting right there in front of me (indicating).
Mr. Wicks [ the prosecutor]:
I would ask the record reflect the witness has identified the defendant, Mack Sims, Your Honor.
The Court:
Let the record reflect the witness has identified the defendant.
Carey continued to describe the assailant,
I noticed the shoes looked somewhat like boots, I would say, and dark color pants. What I mainly noticed was the coat. The coat was slightly long and dark in color, either a dark black or maybe bluish or-like a Navy or midnight blue or something to that extent. One of the things I did notice on the coat was a patch on the arm ... It was a small patch. I didn't really get a good look at the shape, but it couldn't have been more than a couple inches in diameter.
Carey then observed the coat Sims was wearing when he was arrested and identified a patch on it as being the one he witnessed when the assailant approached his vehicle. Carey also testified that detective John Faigh came to speak with him at the hospital the day after his surgery. He recalled being presented with six photos. Carey chose a photograph from this lineup noting that the picture "looked like" the assailant. Carey then identified in court the photograph of Sims that he picked out of that lineup.
During cross-examination defense counsel pointed out that Carey's identification of the assailant in the photographic lineup was not unequivocal. Carey indicated that he had unequivocally identified the assailant in a photographic lineup that appeared nowhere in the record:
A: The identification that I was provided with in the emergency room was [unequivocal]. They did show me something in the emergency room. It was difficult to see, but what I did see, that was him.
Q: Oh, there was another picture that was shown to you?
A: There was another picture in the emergency room.
Q: There is nothing in the reports about that: is that correct?
A: Not that I know of.
Carey was then asked about a situation shortly before trial in which he was unable to identify the assailant in a photographic lineup:
Q: And you also recall, maybe even last week, looking at a photo lineup when Mr. Wicks was around?
A: Uh-huh.
Q: And you again said that it looks like but you couldn't be sure because of some facial hair?
A: Yeah, the facial hair did throw me off, but I will not forget the eyes. The eyes are the one thing that I do remember.
Q: Now you indicated-yeah, you have talked about the eyes, but you didn't say anything in your statement about the eyes.
A: That's the one thing that I do remember.
Q: But you didn't tell the police about it at the time?
A: No. But like I said, I was also very groggy.
Sims later testified and his mug shot showed that he had facial hair, a mustache and goatee, at the time he was arrested.
Additionally, Carey testified that the photograph of Sims in the emergency room that was not part of the record was shown to him by itself and not as part of a photographic lineup:
Q: And you indicated now that there was another picture that you were shown, a single picture you were shown, in the emergency room?
*1083A: I was shown that in the emergency room. If I recall, my-my parents were in the emergency room, but I'm not sure if they were in there when they showed me this picture. They did show me a picture in the emergency room.
Carey stated the lighting in the parking lot was "somewhat-subdued would be the word I'd use, somewhat faint." Carey also testified that he was not wearing his glasses the night of the shooting, although his vision was not so poor that he was required to wear them to drive.
Defense counsel impeached Carey regarding inconsistencies in his description of the assailant. The defense pointed out that the description of the assailant included that he had short hair or was bald. However, Sims later testified that his hair was not short or shaved but was curly and longer at the time he was arrested. The defense pressed Carey on his description of the shoes worn by the assailant. Carey had described the assailant as wearing black combat boots; Sims later testified that he was wearing black and white Nike sneakers. The defense also noted that Carey had indicated the assailant was wearing black pants, but Sims testified that he was wearing blue jeans. The defense asked Carey if there was a hood on the coat, to which Carey responded that his memory had improved over time on the matter: "I recall-I did not recall it at the time. Later on I did recall a hood, and it was part of the way up." During direct examination Sims testified that he was wearing an Orlando Magic baseball cap when he was arrested. Furthermore, the defense pointed out that the distinct patch detailed in direct examination and used to identify Sims's coat was never mentioned in his description of the assailant.
After Carey stepped down from the witness stand the defense moved for a mistrial based upon the testimony elicited from Carey:
[Carey] testified that while he was in the emergency room immediately after being shot that he was shown a single photograph of the defendant. The defendant would indicate to the Court that the type of identification process, being a single photograph, is prejudicial and suggestive and clearly taints any subsequent identification that may have been made in this matter ... Based on that single photographic identification, I believe that the in-court identification was tainted by the suggestive nature of the initial identification.
The court denied the motion stating there was no evidence from the state or defendant that indicated a single photograph was ever shown to the victim. The court stated the only way it could be assumed that a single photograph lineup occurred was by accepting everything Carey said was true, which was not required. The court also found the in-court identification sufficient to overcome any undue suggestiveness of a single photograph lineup.1
*1084Closing argument took place on August 24, 1994. The prosecution leaned heavily, almost exclusively, on Carey's testimony: "He has identified as the shooter this defendant, Mack Sims, and he has never hesitated a bit in that identification. And that, of course, is what this case is about is the validity of that identification." The jury found Sims guilty of attempted murder.
C. The Sentencing and Appeal Process
Sims was sentenced to 35 years' imprisonment on December 1, 1994. The defense filed a motion for a new trial on December 29, 1994, arguing "the Court erred when it allowed the identification evidence of Shane Carey in that it was tainted by an impermissibly suggestive pre-trial viewing of a picture of the Defendant by Mr. Carey." On July 15, 1995, the trial court dismissed the motion. Its response in full was: "The court has examined this alleged error and having further examined its notes, now determines that it believes that no error was committed or that no impermissible suggestive identification took place and that there is no merit to specification no. 1."
Sims did not fare much better on appeal. In an unpublished opinion, the Court of Appeals of Indiana opined that the "[e]xtra-judicial exhibition of a single photograph to a victim is an unduly suggestive identification procedure." However, the court noted the strength of Carey's in-court testimony, in particular that he looked Sims directly in the eye, noticed the light-colored patch of skin under one eye, and Sims was only two or three feet away from Carey when he was shot. These factors considered with the identification of Sims in photographic lineups convinced the court that the totality of the circumstanced provided a sufficient basis, independent of the improper photograph display, to support the admissibility of the in-court identification of Sims.
D. Post-Conviction Proceedings
On February 8, 2012, in what was supposed to be an evidentiary hearing regarding a post-conviction relief petition filed by Sims, the information that formed the basis of this case was revealed. Graham Polando, a deputy prosecuting attorney for the state of Indiana stated in open court the following: "I consulted with Judge [Charles] Wicks who was the trial deputy [for Sim's attempted murder trial in 1994], he asked me not to disclose what he told me, but he indicated that the victim in this case identified the ... defendant, Mr. Sims, only after hypnotism." This fact was never disclosed to defense counsel.
On June 8, 2012, the Elkhart County Superior Court held an evidentiary hearing to address the hypnotism issue. Carey testified that when viewing the lineup administered by detective Faigh the day after the shooting, he merely stated the individual "looked like" the assailant because "at the time [he] was not extremely sure." Carey also testified that the prosecuting attorney in the case, Wicks, brought up the idea of hypnosis saying "they could put [him] under hypnosis... but [indicated] there might be a problem in court in the future." Carey responded that he did not have a problem with any future legal issues regarding the hypnosis so long as he was able to recall the person who shot *1085him. Carey also testified that Wicks set up the appointment and the state paid for it.
Carey testified that he attended one session of hypnosis in which he "fell asleep ... [and] literally entered a dream state in which I ... recall[ed] the shooting itself. And during that time I had another opportunity to uh, see the person that shot me." Sims's attorney then asked Carey:
Q: Okay. So it was really only after this hypnosis that you were sure of the person was that shot you?
A: Yeah.
Carey indicated the session took place months before trial "when [Wicks] and I first started talking about who the perpetrator was." After the session, Wicks created another photographic lineup and this time Carey was able to identify Sims due to the skin coloration on the face of the assailant that Carey noticed while reliving the night of the shooting under hypnosis. In fact, Carey stated that only "after the hypnotism the birthmark really stood out."
The court reconvened the hearing on June 12 and Charles Wicks, the prosecutor in Sims's case and now a Judge for the Elkhart County Superior Court, testified. He stated that he could not recall whether Carey disclosed he had been hypnotized. A few breaths later Wicks defended not disclosing the hypnotism asserting it was not exculpatory in nature because Carey never wavered in his identification of Sims as the assailant. Wicks then testified that he gave Carey the information of the individual that performed the hypnotism because even though Carey had given a "fairly complete summary of what happened to him that evening, he said he would like to be able to recollect the evening better." The hypnotist was George Atkins, a licensed physician's assistant who Wicks knew from Kiwanis Club and had used in the past to help personal injury clients recall traumatic events.
Later in the hearing the court questioned deputy prosecuting attorney Polando regarding the statement Wicks made to him, asking whether the following quotation was correct: " 'I never told Jim Stevens' (meaning Deputy Public Defender Stevens), 'that the victim was ... only able to identify Mr. Sims after he was hypnotized." Polando confirmed this was correct, and added that Wicks also told him not to tell anyone. When Wicks retook the stand he stated he did not recall saying this to Polando.
On September 6, 2012, the Elkhart Superior Court entered an order denying Sims's Second Amended Petition for Post-Conviction Relief. In its order, the court identified the main issue as "whether Carey was able to sufficiently identify Petitioner before hypnosis." The court then discounted Wicks' statement to Polando that Carey was only able to identify Sims after hypnosis because Wicks "had not had time to consider the case before speaking with DPA Polando." The court also gave weight to the fact that other evidence contradicted this statement. The court noted that Officer Lerner testified that Carey could identify the assailant from the beginning and that his description of the assailant matched Sims. The court then cited with approval the photographic lineup presented to Carey the day after the shooting in which Carey noted the picture of Sims "looked like" the assailant. The court concluded "[f]rom the record of the case, Carey was able to identify the Petitioner well before hypnosis."
After outlining the test established by Brady v. Maryland , 373 U.S. 83, 87, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), the court noted the evidence was favorable to the defense, at a minimum as impeaching evidence, that would have been beneficial to the defense to discredit the state's only eye-witness. The court then immediately *1086concluded that the evidence was not material because there was not a reasonable probability that disclosure would have changed the result of the proceeding. It was persuaded because the defense vigorously cross-examined Carey at trial and the Indiana Rules of Evidence only rendered inadmissible testimony of a witness as to matters recalled only through hypnosis. Thus, the court reasoned, Carey's testimony would not have been found inadmissible because Carey could sufficiently identify his assailant prior to hypnosis. The court also noted that Carey never identified anyone other than Sims from a photographic lineup and that his description of the assailant did not substantially change after hypnosis.
On July 15, 2013, the Court of Appeals of Indiana affirmed this decision. The court noted that under Indiana law, evidence derived from a hypnotically entranced witness should be excluded because the evidence is affected by confabulation, a process that causes the subject to fill in memory gaps with fantasy. The court also noted that because the subject is confident that the recalled memories are based in fact and accurate, the witness will likely be impervious to cross-examination. However, the court held that the state had carried its burden of demonstrating by clear and convincing evidence the witness's in-court identification had a factual basis independent of the hypnosis and therefore would have been admissible under the Indiana Rules of Evidence. The court noted that Carey testified at trial that he was able to look directly into the face and eyes of his assailant and that Carey's description of Sims the night of the shooting matched Sims. The court also noted Carey identified Sims three times in photographic lineups before he underwent hypnosis. The court found that Carey's identification had a sufficient pre-hypnosis foundation to have been admissible because Carey "was subjected to vigorous cross-examination regarding his numerous identifications of Sims." The court also gave weight to Carey's testimony at the evidentiary hearing that the hypnosis did not help him identify Sims, but rather made him "extremely sure" of his identification.
After the Indiana Supreme Court denied relief, Sims filed a petition for a writ of habeas corpus in federal court. The district court found that the Indiana Court of Appeals' decision was neither contrary to, nor involved an unreasonable application of, clearly established federal law. The court also found that the state court's decision was not based on an unreasonable determination of the facts. Accordingly, the district court denied habeas relief, but certified its appealability pursuant to 28 U.S.C. § 2253(c), finding that reasonable jurists could differ on whether the post-conviction relief court erred in finding relief was not warranted.
II. ANALYSIS
We review de novo a district court's denial of a petition for writ of habeas corpus. Carter v. Thompson , 690 F.3d 837, 843 (7th Cir. 2012). This review is governed by the Anti-terrorism and Effective Death Penalty Act of 1996 ("AEDPA"). The statute significantly limits the scope of our review; "habeas relief cannot be granted for persons in custody pursuant to a judgment of a state court unless the adjudication of the claim: '(1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established federal law, as determined by the Supreme Court of the United States; or (2) resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.' "
*1087Czech v. Melvin , 904 F.3d 570, 573 (7th Cir. 2018) (quoting 28 U.S.C. § 2254(d) ). A federal court may issue a writ of habeas corpus under the "contrary to" clauses of AEDPA if the state court applied a rule different from law set forth in Supreme Court precedent, or if it decides a materially indistinguishable case differently than the Supreme Court. Bell v. Cone , 535 U.S. 685, 694, 122 S.Ct. 1843, 152 L.Ed.2d 914 (2002) (citing Williams v. Taylor , 529 U.S. 362, 405-406, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000) (O'Connor, J., concurring) ). A state court decision is an "unreasonable application" if it correctly identifies the governing legal rule, but applies it unreasonably. Williams v. Taylor , 529 U.S. 362, 408-409, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000).
On appeal, Sims's principal argument is that the state withholding the evidence that Carey was hypnotized prior to trial violated Sims's constitutional rights as established in Brady . Under Brady , a defendant's due process rights are violated if the state withholds favorable evidence from the defense that is material to the defendant's guilt or punishment. Brady , 373 U.S. at 87, 83 S.Ct. 1194. The state court found, and the parties do not dispute, the hypnosis evidence would have been favorable to the defendant and it was not disclosed by the state. We agree. Therefore, the only issue before us is whether the Indiana court's decision, that suppression of evidence that the state's star witness was hypnotized was not material under Brady , is contrary to or an unreasonable application of clearly established federal law.2 Because the Indiana court's decision was both, we reverse the district court and grant the writ.
A. "Clearly Established" Federal Law
The Supreme Court has clearly established that strong and non-cumulative impeachment evidence related to an important trial witness is material under Brady. A new trial is not "automatically require[d] ... whenever a combing of the prosecutors' files after the trial has disclosed evidence possibly useful to the defense but not likely to have changed the verdict. A finding of materiality of the evidence is required under Brady ." Giglio v. United States , 405 U.S. 150, 154, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972) (internal citations and quotations omitted). Evidence is material under Brady if "there is a reasonable probability that, had the evidence been disclosed, the result of the proceeding would have been different." Kyles v. Whitley , 514 U.S. 419, 433, 115 S.Ct. 1555, 131 L.Ed.2d 490. A "reasonable probability" exists if the suppression of the favorable evidence "undermines confidence in the outcome of the trial." Id. (quoting United States v. Bagley , 473 U.S. 667, 678, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985) ).
Nearly half a century ago, the Supreme Court held "[w]hen the reliability of a given witness may well be determinative of guilt or innocence, nondisclosure of the evidence affecting credibility" justifies a new trial under Brady . Giglio , 405 U.S. at 154, 92 S.Ct. 763. In Giglio , the defendant was convicted of passing forged money orders. Id. at 151, 92 S.Ct. 763. The prosecution's case centered around the testimony of the defendant's co-conspirator, Robert *1088Taliento, who was the only witness able to link the defendant to the crime. Id. Taliento confessed and described the scheme to a grand jury. Id. At trial, Taliento testified and identified the defendant as the instigator of the scheme. Id. Defense counsel vigorously cross-examined Taliento seeking to impeach him regarding a possible arrangement for prosecutorial leniency if he agreed to testify. Id. Taliento testified that no agreement had been reached, but the defendant later discovered he agreed to testify before the grand jury if the state agreed not to prosecute him. Id. at 151-52, 92 S.Ct. 763. Even though the impeachment evidence only went to his credibility, the Court found the information was material under Brady. Id. at 154, 83 S.Ct. 1194.
Conversely, cases in which the Supreme Court has found the suppression of impeachment evidence was not material under Brady are easily distinguishable. For example, in Turner , the Court found suppressed impeachment evidence was not material because it was "largely cumulative of impeachment evidence petitioners already had and used at trial," and because the impeachment evidence only involved minor witnesses. Turner v. United States , --- U.S. ----, 137 S.Ct. 1885, 1894, 198 L.Ed.2d 443 (2017).
Thus, the Supreme Court has long recognized that suppression of strong and non-cumulative evidence related to the credibility of an important witness is material under Brady , at least when the witness's testimony is critical to the prosecution's case. See also Kyles , 514 U.S. at 441-42, 115 S.Ct. 1555 (holding that the state's case relied heavily on the testimony of eyewitnesses who identified the defendant as the murderer, therefore failure to disclose impeachment evidence related to those witnesses was material under Brady ) and Wearry v. Cain , --- U.S. ----, 136 S.Ct. 1002, 1007, 194 L.Ed.2d 78 (2016) (withholding impeachment evidence of state's star witness violated Brady ).
B. "Unreasonable Application of" and "Contrary to" Controlling Precedent
A close reading of the Indiana Court of Appeals' decision shows where that court went astray from the law established by the Supreme Court of the United States. The state court actually acknowledged that "[e]vidence derived from a hypnotically entranced witness is inherently unreliable as not having probative value and is therefore inadmissible." Sims v. State , 990 N.E.2d 523, 2013 WL 3526759 at *4 (Ind. Ct. App. 2013), citing Rowley v. State , 483 N.E.2d 1078, 1081 (Ind. 1985). Rowley , in turn, followed Strong v. State , 435 N.E.2d 969, 970 (Ind. 1982), which reviewed case law from around the country in rejecting hypnotically enhanced testimony. Rowley and Strong had gone on to hold that a witness who has undergone hypnosis may testify to identify a wrongdoer in a criminal trial, nevertheless, if the prosecution can show by clear and convincing evidence that the in-court identification has a sufficient independent factual basis.
In Sims's case, the Indiana Court of Appeals veered away from the Brady materiality standard. Instead of deciding whether the concealed evidence was important enough to undermine confidence in the result of the trial without it, the state court analyzed whether Carey's in-court identification of Sims had a sufficient independent factual basis so as to have been admissible . We assume that the finding of admissibility was correct under state law. But the state court then made the leap that was contrary to, and an unreasonable application of, Brady and its progeny: it concluded that because Carey's testimony would still have been admissible, "it is not reasonably probable that the outcome of *1089Sims's trial would have been different had Carey's hypnosis been disclosed."
That was a clear error. Brady 's materiality standard is not an admissibility test. It requires the court to gauge the potential effects on the outcome of the trial if the concealed information had been available to the defendant. See Smith , 56 U.S. at 75-76 ; Kyles , 514 U.S. at 453-54, 115 S.Ct. 1555 ; Strickler , 527 U.S. at 289-90, 119 S.Ct. 1936. The Indiana court did identify Brady 's overarching rule, but failed to correctly apply, or even recognize, the materiality standard outlined by the Supreme Court. Courts must consider the overall strength of the prosecution case, the importance of the particular witness's credibility to the prosecution case, the strength of the concealed impeachment material, and how the concealed material compares to other attacks the defense was able to make on the witness's credibility. See Kyles , 514 U.S. at 441, 445, 451, 454, 115 S.Ct. 1555 ; Giglio , 405 U.S. at 154-55, 92 S.Ct. 763 ; Smith , 565 U.S. at 76, 132 S.Ct. 627 ; Wearry , 136 S.Ct. at 1006-07.
Giglio , Kyles , Wearry , and Smith all involved concealment of strong and non-cumulative impeachment evidence for the witnesses whose credibility was critical to the prosecution cases. They provide a body of clearly established law showing when impeachment evidence is material under Brady . Turner and Strickler , by contrast, show that concealed impeachment evidence may not be material when the prosecution case is strong apart from the witness in question and when the concealed impeachment evidence would have added little weight to the defendant's attacks on the witness's credibility.
Concealing the hypnosis of Carey in this case falls on the material side of the line mapped by these Supreme Court cases. Without Carey's identification of Sims as the shooter, the prosecution had no case. No physical evidence tied Sims to the shooting. His presence by the dumpster shortly after the shooting was suspicious, of course, but far short of what would have been needed to convince a jury to convict.
The fact that Carey had been hypnotized would have undermined his credibility and changed his cross-examination quite dramatically. As the Supreme Court explained in Rock v. Arkansas , 483 U.S. 44, 59-60, 107 S.Ct. 2704, 97 L.Ed.2d 37 (1987), there are several serious problems that undermine the accuracy and credibility of hypnotically enhanced testimony:
Responses of individuals to hypnosis vary greatly. The popular belief that hypnosis guarantees the accuracy of recall is as yet without established foundation and, in fact, hypnosis often has no effect at all on memory. The most common response to hypnosis, however, appears to be an increase in both correct and incorrect recollections. Three general characteristics of hypnosis may lead to the introduction of inaccurate memories: the subject becomes "suggestible" and may try to please the hypnotist with answers the subject thinks will be met with approval; the subject is likely to "confabulate," that is, to fill in details from the imagination in order to make an answer more coherent and complete; and, the subject experiences "memory hardening," which gives him great confidence in both true and false memories, making effective cross-examination more difficult.
Essentially, "[n]ot only do hypnotized witnesses find it difficult to distinguish their original memories from those brought out under hypnosis, but they also tend to become more confident about their recall despite the fact that it might contain false recollections." Edie Greene, Kirk Heilbrun, William H. Fortune, & Michael T. Nietzel, Wrightsman's Psychology and the Legal System 140 (6th ed. 2007); see also *1090Steven Jay Lynn, Elza Boycheva, Amanda Deming, Scott O. Lilienfeld, & Michael N. Hallquist, Forensic Hypnosis: The State of the Science, in Psychological Science in the Courtroom: Consensus and Controversy, 85 (Jennifer L. Skeem, Kevin S. Douglas, & Scott O. Lilienfeld 2009) ("23 studies have shown that hypnosis either increases confidence relative to a nonhypnotic group, or participants confidently report inaccurate memories of events they earlier denied occurred when they were not hypnotized").
The concealed hypnosis thus explains Carey's puzzling statement at trial that his memory of the incident actually improved over time.3 But more fundamentally, it calls into question everything Carey said at trial. Based on the Supreme Court's view of hypnosis, Carey would not know what he was able to recall independent of the hypnosis, nor what he was able to recall because of the hypnosis, or whether any of his testimony was true or based on fantasy. This is made more troubling by the fact that Carey provided significantly more information at trial than he did any time before trial. Carey discussed in detail looking the assailant square in the eyes, that he remembered witnessing a small patch on the assailant's jacket, that he remembered a small birthmark on the assailant's face. As the defense pointed out at trial, none of these details were in Carey's description to Officer Lerner nor did they appear in the police report. It is reasonable to infer the jury found these details persuasive without knowing that Carey's recollection of them might have been due entirely to the hypnosis session.
The "memory hardening" effect of hypnosis can explain Carey's admission during post-conviction proceedings that he was sure of who the person was who shot him only after being hypnotized. The effect Carey's increased confidence likely had on the jury helps to undermine our confidence in the verdict. Decades of research confirms that the confidence with which eyewitnesses identify criminal defendants can be a powerful predictor of verdicts regardless of the accuracy of the identification . The more confident the eyewitness is in his identification, the more likely the jury is to believe that the identification is accurate and to convict the defendant. See Brian L. Cutler, Steven D. Penrod, & Thomas E. Stuve, Juror Decision Making in Eyewitness Identification Cases, 12 Law and Human Behavior 41 (1988); Steven G. Fox & H.A. Walters, The Impact of General versus Specific Expert Testimony and Eyewitness Confidence upon Mock Juror Judgment , 10 Law and Human Behavior 215 (1986); Michael R. Leippe, Andrew P. Manion, & Ann Romanczyk, Eyewitness Persuasion: How and How Well Do Fact Finders Judge the Accuracy of Adults' and Children's Memory Reports? , 63 Journal of Personality & Social Psychology 181 (1992); Lynn et al. at 85; Gary L. Wells, What Do We Know About Eyewitness Identification? , 48 American Psychologist 553, 564 (1993); Gary L. Wells, R.C.L. Lindsay, & Tamara J. Ferguson, Accuracy, Confidence, and Juror Perceptions in Eyewitness Identification , 64 Journal of Applied Psychology 440 (1979).
It is not difficult to imagine what Sims's lawyer could have done at trial with the knowledge that Carey had been hypnotized. The known effects of hypnosis could explain Carey's confidence, his claim that his memory of the shooting had improved over time, and the otherwise benign changes in his descriptions of the shooter. Reasonable judges cannot be confident that, if the jury had known that Carey had been hypnotized before he identified Sims *1091at trial, they would have found his identification beyond reasonable doubt.
Given the well-known problems that hypnosis poses for witnesses' memories, we can be confident that Carey's identification testimony would have been subjected to withering cross-examination. As noted, the prosecution's case against Sims depended completely on Carey's credibility, which the suppressed hypnosis evidence would have severely undermined. The evidence would have cast doubt for the jury not only on Carey's in-court identification, but also on Carey's credibility as a witness more generally, including the accuracy of his prior identifications of Sims. The jury saw Carey identify Sims only once, in court. The prior identifications were heard only secondhand. If the jurors had known that Carey needed to be hypnotized to make the in-court identification, they would have been less likely to believe Carey was confident that Sims was his assailant, and therefore that his identification was accurate. From there and without the ability to observe Carey make his prior, untainted identifications, the jury could easily have questioned Carey's overall credibility as an eyewitness.
The Indiana appellate court noted that Carey was able to describe the clothing and physical attributes of the assailant who matched Sims's clothing and physical attributes when he was discovered on the scene. However, the trial court transcript illustrates the defense cross-examined Carey vigorously and pointed out several instances in which his description of the assailant did not match Sims's clothing or physical attributes the night of the incident. The undisputed details merely provide the shooter was a black male with a large build wearing a three quarter length coat. This is not the kind of identification that instills confidence especially in a case that the prosecution described as being all about "the validity of Carey's identification."
The state court noted Carey identified Sims three times in photographic lineups before hypnosis. Carey had glass in his eyes and did not recall the first lineup and was unable to identify the shooter in a lineup two days later. Furthermore, significant doubt was cast on these lineups by the fact that Carey testified that he was initially only shown a single picture. The defense moved for a mistrial, which was denied, and the issue was affirmed on appeal. However, the Indiana Court of Appeals affirmed this decision because they were convinced the totality of the circumstances constituted a sufficient independent basis to cure any unduly suggestive procedures. But the suppressed evidence of hypnosis now brings this ruling into question.
The dissent assails our opinion by asserting that Carey never wavered in his identification of Sims. This does not explain why Wicks felt it necessary to take the risk of setting up a hypnosis session for Carey without disclosing it. Nor does it appear to take into account the instances in which Carey equivocated. Furthermore, the only indication as to when the hypnosis session took place is Carey's testimony at the post-conviction evidentiary hearing that it was months before trial when he and Wicks "first started talking about who the perpetrator was."
Finally, these problems with hypnosis undercut the Indiana court's final reason for refusing post-conviction relief: Carey's testimony indicated he was able to identify the assailant, but hypnosis was able to make him "extremely sure." No one knows what effect the hypnosis had on Carey and it also belies the record for reasons discussed above.
Considering the overall weakness of the prosecution case without Carey, the importance *1092of his testimony, the explosive strength of the concealed hypnosis evidence, and the relatively mild impeachment of Carey that the defense managed at trial, habeas relief is required. The post- Brady cases involving strong concealed impeachment material for key prosecution witnesses- Smith , Giglio , Wearry , and Kyles -show beyond reasonable dispute that the prosecutor's deliberate concealment of the hypnosis evidence undermined confidence in the verdict that has kept Sims in prison for more than twenty years.
III. CONCLUSION
Given the suppression of the evidence was clearly a violation under Brady , the writ of habeas corpus should have been granted. Therefore, we reverse and remand the case to the district court with instructions to grant the writ of habeas corpus.