RILEY, Chief Judge.

Earl Forrest was convicted of three counts of capital murder in Missouri state court and was sentenced to death. The Missouri Supreme Court affirmed the conviction and sentence on direct appeal, see State v. Forrest (Forrest I), 183 S.W.3d 218, 232 (Mo.2006) (en banc), and later affirmed the denial of Forrest's motion for post-conviction relief, see Forrest v. State (Forrest II), 290 S.W.3d 704, 718 (Mo. 2009) (en banc). The district court1 denied Forrest's application for writ of habeas corpus under 28 U.S.C. § 2254, but granted a certificate of appealability on the question of ineffective assistance of counsel during the penalty phase of Forrest's trial. Forrest appeals the denial on this ground. With appellate jurisdiction under 28 U.S.C. § 1291, we affirm.

I. BACKGROUND

On December 9, 2002, Forrest, who had been drinking alcohol, and Angelia Gamblin, his girlfriend, drove to the home of Harriet Smith to resolve "a dishonored agreement" between Forrest and Smith. Forrest I, 183 S.W.3d at 223.2 The agreement was that Forrest would introduce Smith to a source of methamphetamine in return for a lawn mower and a mobile home. See id. While Gamblin waited in the car, Forrest went into Smith's house and "demanded that Smith fulfill her part of the bargain." Id. A melee ensued during which Forrest killed Michael Wells, a visitor at Smith's residence, with a "close contact"3 shot to the face. See id. Smith attempted to escape with Gamblin's car, but hit a tree limb while backing out of the driveway. With the car stuck, Forrest coerced Smith back into the house where Forrest killed her, shooting her a total of six times, including two close contact shots to her face. See id.

Forrest and Gamblin fled, taking with them a lockbox containing about $25,000 of methamphetamine. See id. After Forrest and Gamblin returned home, Sheriff Bob Wofford and Deputy Sharon Joann Barnes, who were investigating the shooting, knocked on the door and asked for Forrest. See Forrest II, 290 S.W.3d at 707. Forrest approached the door and opened fire, wounding Sheriff Wofford and killing Deputy Barnes. See id.; Forrest I, 183 S.W.3d at 223. After Forrest and Gamblin were shot, Forrest surrendered. See Forrest I, 183 S.W.3d at 223.

A. Trial and Direct Appeal

Represented by Missouri public defenders Sharon Turlington and David Kenyon (collectively, defense counsel), Forrest was charged in Missouri state court with three counts of first degree murder for the deaths of Smith, Wells, and Deputy Barnes, for which the state sought the death penalty. See id. During the guilt phase of Forrest's trial, the jury found Forrest guilty on all three charges. See id.

During the penalty phase, defense counsel called fourteen lay witnesses to provide mitigating testimony. Forrest's brother William testified that he and Forrest regularly used illegal drugs and that William, Forrest, and their father drank alcohol heavily. William said their father always punished Forrest more strictly than William and recalled an occasion when their father brought Forrest to their front yard and "slapp[ed] him around" publicly. William explained their father pressured him and Forrest to be "tough guys" and to fight other boys. William testified that when they were adults, he and Forrest once had a fight during which Forrest hit William so hard William saw stars and William stabbed Forrest in the stomach with a buck knife. Defense counsel also presented the testimony of four of Forrest's stepchildren4 to demonstrate he was a good father-figure. The children's mother, Nancy Young, testified to the same effect. Several friends and acquaintances, including Gamblin, testified Forrest abused alcohol and methamphetamine but that he still was a good person, parent, and friend.

Defense counsel also called three mental health experts. Clinical psychologist Dr. Robert Smith diagnosed Forrest with long-term depression, brain damage, and substance dependence. Dr. Smith explained that these conditions combined to cause mood swings as well as difficulty concentrating and problem-solving. Neuropsychologist Dr. Michael Gelbort concluded, according to a battery of neuropsychological tests, Forrest suffered from frontal lobe dysfunction, which accentuated Forrest's impulsive behavior. Dr. Gelbort also explained brain scans had not been performed on Forrest, but that the type of impairment from which Forrest was believed to suffer would "rarely" appear on PET,5 SPECT, MRI, or CT scans. Psychiatric pharmacist Dr. Roswell Lee Evans testified that long-term drug and alcohol abuse affects the ability to exercise judgment and recall memories. Dr. Evans also testified that considering the volume of alcohol consumed, Forrest was most likely in an "alcoholic blackout" at the time of the murders.

The jury unanimously recommended a death sentence for each conviction, finding multiple statutory aggravators applicable: Forrest killed Smith during the murder of Wells; he killed Smith and Wells to obtain something of pecuniary value (methamphetamine); and he killed Deputy Barnes while she performed her official duty as a peace officer. See Forrest I, 183 S.W.3d at 223. On appeal, the Missouri Supreme Court upheld Forrest's conviction and sentence. See id. at 232.

B. Post-Conviction Relief and Habeas Proceedings

After his unsuccessful direct appeal, Forrest moved for post-conviction relief, see Mo. Sup.Ct. R. 29.15, raising dozens of claims, including the four claims at issue here that defense counsel was ineffective during the penalty phase. Forrest's ineffective assistance claims cited defense counsel's failure to (1) conduct a PET scan; (2) present certain medical records; (3) call a particular psychologist to testify about Forrest's probability of future dangerousness; and (4) investigate three specific acquaintances of Forrest. After an evidentiary hearing as to those issues, the Missouri circuit court denied the motion.

On appeal, Forrest challenged, among other things, the state circuit court's conclusions on each of his four currently relevant ineffective assistance claims as to the penalty phase. The Missouri Supreme Court affirmed, analyzing each of Forrest's four ineffective assistance claims considered at the hearing and concluding that in each instance defense counsel's performance was constitutionally adequate and non-prejudicial. See Forrest II, 290 S.W.3d at 708-12, 714-15, 718.

Forrest then applied for a writ of habeas corpus in federal district court for the Western District of Missouri, presenting, among other claims, the four alleged instances of penalty phase incompetence as parts of a single ineffective assistance claim and arguing the Missouri Supreme Court improperly viewed these points individually, rather than cumulatively. The district court denied the application, rejecting Forrest's cumulative analysis assertion under Middleton v. Roper, 455 F.3d 838, 851 (8th Cir.2006), and concluding federal relief was barred under 28 U.S.C. § 2254(d). The district court granted Forrest a certificate of appealability as to his claims of ineffective assistance during the penalty phase. Forrest appeals on this basis.

II. DISCUSSION

"When considering the district court's denial of a habeas petition, `we review the district court's findings of fact for clear error and its conclusions of law de novo.'" Middleton, 455 F.3d at 845 (quoting Lyons v. Luebbers, 403 F.3d 585, 592 (8th Cir.2005)). The Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), 28 U.S.C. § 2254(d), prohibits us from granting habeas relief unless the state adjudication

28 U.S.C. § 2254(d). "AEDPA modified a federal habeas court's role in reviewing state prisoner applications in order to prevent federal habeas retrials and to ensure state-court convictions are given effect to the extent possible under law." Abernathy v. Hobbs, 748 F.3d 813, 816 (8th Cir. 2014). Forrest "carries the burden of pro[ving]" this hurdle is met. Cullen v. Pinholster, 563 U.S. ___, ___, 131 S.Ct. 1388, 1398, 179 L.Ed.2d 557 (2011). Forrest's underlying ineffective assistance claim is governed by Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), which requires Forrest to "show both that counsel's performance was deficient and that the deficiency prejudiced the defendant." Williams v. Roper, 695 F.3d 825, 830 (8th Cir.2012). We agree with the district court that the Missouri Supreme Court was reasonable in concluding defense counsel satisfied Strickland's performance prong.

A. Strickland's Performance Prong

"To prove that counsel's performance was deficient, the defendant must show `that counsel made errors so serious that counsel was not functioning as the "counsel" guaranteed the defendant by the Sixth Amendment.'" Strong v. Roper, 737 F.3d 506, 517 (8th Cir.2013) (quoting Strickland, 466 U.S. at 687, 104 S.Ct. 2052). Under this standard, counsel must "`make reasonable investigations or ... make a reasonable decision that makes particular investigations unnecessary.'" Id. (quoting Strickland, 466 U.S. at 691, 104 S.Ct. 2052). "`Judicial scrutiny of counsel's performance is highly deferential, indulging a strong presumption that counsel's conduct falls within the wide range of reasonable professional judgment.'" Id. (quoting Bucklew v. Luebbers, 436 F.3d 1010, 1016 (8th Cir.2006)). Here, our review is governed by both § 2254(d) and Strickland, meaning we are to be "`twice deferential: we apply a highly deferential review to the state court decision; the state court, in turn, is highly deferential to the judgments of trial counsel.'" Strong, 737 F.3d at 517 (quoting Nooner v. Norris, 402 F.3d 801, 808 (8th Cir.2005)); see also Harrington v. Richter, 562 U.S. 86, ___, 131 S.Ct. 770, 788, 178 L.Ed.2d 624 (2011) (stressing the difficulty of overcoming both § 2254(d) and Strickland "when the two apply in tandem").

With two exceptions discussed below, see infra sections II.B and C, Forrest does not claim the Missouri Supreme Court's decision was "contrary to" clear Supreme Court precedent. Forrest bases his arguments on the "unreasonable application" language of § 2254(d)(1) and the "unreasonable determination of facts" language of § 2254(d)(2). In reviewing for an unreasonable application of clearly established law, we look not "`to the quality of the reasoning process articulated by the state court'" or "the statement of reasons explaining the state court's decision," which "would `place the federal court in just the kind of tutelary relation to the state courts that [AEDPA was] designed to end.'" Williams, 695 F.3d at 831, 834-35 (quoting Hennon v. Cooper, 109 F.3d 330, 334-35 (7th Cir.1997)). "[W]e examine the ultimate legal conclusion reached by the [state] court" and ask "whether there is `any reasonable argument' that the state court's judgment is consistent with Strickland." Id. at 831-32 (quoting Richter, 562 U.S. at ___, 131 S.Ct. at 788). In judging the state court's conclusions we must also remember that "[t]he Strickland standard is a general one, so the range of reasonable applications is substantial." Richter, 562 U.S. at ___, 131 S.Ct. at 788.

Challenges to the state court's factual findings place a similarly weighty burden on the petitioner. "Unless [Forrest] rebuts by clear and convincing evidence the presumption that [a state court's] finding was correct, 28 U.S.C. § 2254(e)(1), and shows that the finding was unreasonable, id. § 2254(d)(2), we may not grant him relief." Taylor v. Roper, 577 F.3d 848, 859 (8th Cir.2009).

1. PET Scan

Though three medical experts testified during Forrest's penalty phase, Forrest argues defense counsel was unreasonable in failing to have Forrest undergo a PET scan. The record shows defense counsel considered the possibility of brain scanning as early as May 2003. In June of 2004, Dr. Evans suggested obtaining a PET scan, and in September that year Dr. Gelbort gave a similar recommendation. In September of 2004, Turlington, through another attorney, contacted Dr. Ken Smith to arrange for a PET scan, but Dr. Ken Smith refused to perform the scan without first performing an MRI. Following Dr. Ken Smith's insistence on an MRI, defense counsel chose to proceed to trial without a PET scan.

At the post-conviction relief hearing, Turlington testified an MRI likely would not show the type of damage Forrest was believed to have. At trial, Dr. Gelbort agreed. Turlington testified also that she has used PET scans in other trials, but with PET scans "you're not guaranteed that there's going to be anything on a scan... and then you created evidence that undercuts your own evidence." The problem with this, she testified, is that Missouri state law prevents a motion for transportation to the scan location from occurring ex parte and under seal, making it easy for the state to learn of negative results. Turlington explained, "I can say one hundred percent for sure if we could have done [a PET scan] ex parte and had the order filed under seal, we would have done the scan, one hundred percent for sure."

Dr. David Preston, a physician specializing in nuclear medicine, conducted a PET scan after Forrest's conviction. The scan, according to Dr. Preston, showed brain damage and reduced metabolism in Forrest's frontal lobes. At the post-conviction relief hearing, Dr. Preston testified the scan cannot demonstrate diminished capacity, predict future behavior, or determine state of mind-for example, the scan cannot predict moods or criminal behavior. Drs. Smith, Evans, and Gelbort each testified Dr. Preston's results would have helped illustrate and reinforce their conclusions, but would not have changed their testimonies.

The state circuit court found defense counsel made a "fully-informed" and "strategic" decision to forego the scan in light of the possible negative results and inability to move ex parte for transportation to the scan location. The state circuit court therefore concluded defense counsel's decision was reasonable and entitled to deference. See Strickland, 466 U.S. at 689-91, 104 S.Ct. 2052. The Missouri Supreme Court agreed defense counsel's decision was reasonable for this reason. See Forrest II, 290 S.W.3d at 709.

Forrest first challenges the finding that foregoing a PET scan was a strategic decision, arguing the failure was actually because defense counsel neglected the issue until the last minute and then "simply ran out of time." The record does not compel such a conclusion. Forrest notes (1) Drs. Evans and Gelbort suggested a PET scan could substantiate their tests, and (2) September 2004 was the first time defense counsel contacted a doctor to arrange for a PET scan before the trial began on October 4, 2004. But this naked time line of events does not compel an assumption of negligence or disprove the Missouri state courts' finding that strategic concerns motivated the decision. Nor does the record evidence establish defense counsel's ultimate decision was forced by time pressure rather than produced by strategic deliberation. To the contrary, Turlington testified that but for Missouri law preventing defense counsel from obtaining PET scans ex parte and under seal, defense counsel would have acquired a PET scan. Turlington also testified negative or inconclusive results were a real possibility and defense counsel considered the state's potential access to these results throughout pretrial preparation. The record supports a finding that defense counsel's decision was considered and strategic, and Forrest did not satisfy his heavy burden of proving the contrary. See 28 U.S.C. § 2254(d)(2).

Forrest also argues "[a] reasonably competent trial attorney would have realized that the testimony of the trial experts was vulnerable to attack" on the basis that they lacked corroboration through medical testing, and a reasonable trial attorney "would have conducted a PET scan and presented available medical records to foreclose the prosecutor's arguments that completely undermined" Forrest's experts. To the extent the scan results would have been admissible during the penalty phase, the record indicates even favorable PET scan results showing diminished energy usage would merely have corroborated Forrest's existing experts and could not definitively prove the existence of a mental disease or defect.

More importantly, Turlington testified that in her experience, "you're not guaranteed that there's going to be anything on a[PET] scan"6 and that Missouri law prevented an ex parte order to transport Forrest to the scan location. At the time, the Missouri Supreme Court had recently held that "[n]ormally, a defendant is not entitled to an ex parte hearing on" a motion to transport a defendant for a mental evaluation. State v. Anderson, 79 S.W.3d 420, 434 (Mo.2002) (en banc). The Missouri Supreme Court in the present case agreed with defense counsel that "[a]n ex parte order will not be granted to transport a defendant for a mental examination." Forrest II, 290 S.W.3d at 709 n. 3 (citing Anderson, 79 S.W.3d at 434). With knowledge of the scan, the state might have acquired any unfavorable results, meaning the consequences of negative results were potentially severe. See Mo. Sup.Ct. R. 25.06(A) (mandating that upon the state's written motion containing a reasonable request, a trial court must "order the defendant to disclose to the state that material and information requested which is found by the court to be relevant and material to the state's case"); Dees v. Caspiri, 904 F.2d 452, 455, n. 3 (8th Cir. 1990) (per curiam) (explaining that "under Missouri Rule 25.06 the defendant ... may be compelled to reveal the reports of experts he does not intend to call at trial" and noting that "counsel was appropriately cautious of this rule's potential application" to any adverse expert opinions which further investigation might reveal); State v. Carter, 641 S.W.2d 54, 57-58 (Mo.1982) (en banc).

Without the benefit of hindsight, see Abernathy, 748 F.3d at 816, two options would have appeared plausible: rely on Forrest's three existing experts and forego the scan, or attempt to corroborate the experts' testimony by running the risk of materially undermining it. The Missouri Supreme Court was reasonable in concluding defense counsel's choice of the former option fell within Strickland's "wide range of reasonable professional judgment," Bucklew, 436 F.3d at 1016. See 28 U.S.C. § 2254(d)(1).

2. Medical Records

Forrest next claims defense counsel was ineffective in deciding not to introduce medical records from the early 1990s. These records show one instance of an "alleged assault" in which Forrest "was struck in the head with a baseball bat," resulting in an "8 cm scalp laceration." The records also contain reports of depression, suicide attempts, and drug and alcohol abuse.

In a deposition admitted during the post-conviction relief hearing, Kenyon testified he had believed the records were not "mitigation friendly" because the baseball bat incident was the result of a dispute between Forrest and another man over drug debts each owed the other. Kenyon testified he wanted to keep out suggestions that Forrest dealt drugs and particularly that Forrest was involved in drug-related violence. Turlington testified she had believed the baseball bat incident was not "crucial" to the case in light of Dr. Gelbort's expected testimony regarding brain damage.

The state circuit court concluded defense counsel's decision was not unreasonable because (1) most of the medical records' contents were presented through other evidence, having been supplied by counsel to Drs. Gelbort and Evans, thereby making the records cumulative, and (2) the facts surrounding the baseball bat incident were potentially harmful. The Missouri Supreme Court agreed, deferring to defense counsel's strategic decision and concluding that "failing to offer cumulative evidence" does not constitute constitutionally deficient performance. Forrest II, 290 S.W.3d at 710.

First, Forrest agrees the medical records are to some degree substantively cumulative of penalty phase testimony. Forrest argues this finding is nevertheless unreasonable because, like "airtight evidence of DNA test results" relative to "antiquated blood-typing," the credibility of medical records is irrefragable compared to the more controvertible credibility of expert witness testimony on the same subject. Forrest's comparison to DNA sequencing and blood-typing overstates the relative credibility here. Assuming evidence should be seen as non-cumulative where it is significantly more credible than existing evidence, the difference between viewing Forrest's medical records and learning their contents from witnesses — both lay witnesses with direct knowledge and experts who perused the records and interviewed Forrest — is not so great as to make the Missouri Supreme Court's finding unreasonable. Having found the records cumulative, it was reasonable to conclude defense counsel's decision met the requirements of Strickland. See, e.g., Elam v. Denney, 662 F.3d 1059, 1066 (8th Cir.2011) (deciding "the state courts did not unreasonably apply Strickland in concluding that counsel's decision not to call [two experts] to provide cumulative testimony... was not constitutionally deficient performance"); Sherron v. Norris, 69 F.3d 285, 291 (8th Cir.1995) (concluding that given the testimony in hand, "counsel's decision to forgo what would at best have been cumulative testimony was not professionally unreasonable").

Second, the Missouri Supreme Court's conclusion was also reasonable in light of defense counsel's strategic concerns. Forrest argues "the Missouri Supreme Court's strategy finding rested upon a glaring factual error that the medical records contained evidence that [Forrest] was dealing drugs." But the strategic risk was not that the records directly referenced the drug deal — the fear was that the reference to the baseball bat incident could lead to questions about the incident itself, a topic defense counsel understandably wished to avoid.

Forrest also claims defense counsel's decision was unreasonable because the records were necessary to fortify the testimony of Forrest's experts against the state's cross-examination and because Dr. Gelbort opened the door to the baseball bat incident anyway. Though the medical records might have been useful in bolstering the testimony of Forrest's experts, the marginal return of introducing the records to supplement existing testimony reasonably did not justify the risk of giving the state a lead-in to harmful information. Further, the simple fact that Dr. Gelbort mentioned Forrest's baseball bat injury while testifying does not mean there was never a danger in producing the hospital records that note the "alleged assault" of Forrest with a baseball bat during Forrest's involvement in an "altercation."

Given both the diminished utility of largely cumulative records and the strategic reasons to avoid discussing the baseball bat incident, the Missouri Supreme Court reasonably concluded defense counsel's performance was constitutionally sufficient. See 28 U.S.C. § 2254(d)(1).

3. Dr. Cunningham

Forrest next argues defense counsel was ineffective in deciding not to hire Dr. Mark Cunningham, a clinical and forensic psychologist, to testify during the penalty phase. At the post-conviction relief hearing, Dr. Cunningham testified on two points: (1) "protective and risk factors" from Forrest's childhood, and (2) predictions on the likelihood of Forrest's future violence in prison. Dr. Cunningham testified that Forrest's age, adjustment in prison, and high school diploma made him "unlikely to be involved in violence in prison." Dr. Cunningham also testified about Forrest's past, and in particular, his family life, substance abuse, and brain dysfunction.

During his post-conviction relief deposition, Kenyon testified he had interviewed Dr. Cunningham as a possible expert witness on Forrest's risk of future violence. Kenyon testified he and Turlington ultimately decided against hiring Dr. Cunningham out of fear his testimony would prompt the state to ask questions about Forrest's involvement in a California homicide. Kenyon also testified he was concerned Dr. Cunningham's rough predictions as to what his statistical analysis would produce portended testimony that could be more harmful than helpful. Turlington confirmed the decision not to call Dr. Cunningham was based on the belief that his testimony would not help.

The state circuit court ultimately concluded the decision not to call Dr. Cunningham reflected a reasonable strategy, given the fear Dr. Cunningham's testimony could be unhelpful, possibly leading to questions about Forrest's connection to another homicide. The state circuit court also found Dr. Cunningham's testimony questionable and largely cumulative of other evidence. The Missouri Supreme Court agreed with the state circuit court's conclusions. See Forrest II, 290 S.W.3d at 715.

Forrest challenges the conclusion that defense counsel's strategy was reasonable, emphasizing that before trial, defense counsel moved to strike the state's endorsement of a witness whom the state wished to testify on the California homicide. Forrest suggests it was unreasonable to fear questions about this homicide because, as Forrest describes it, the state had "agreed to not bring up" evidence of the homicide or call the witness, and because the state court had "entered an interlocutory order sustaining the defense's motion to exclude" the witness. In reality, the state and defense counsel agreed during the hearing that the state would not address the California homicide so long as defense counsel did not open the door to Forrest's general propensity for violence — for example, by eliciting testimony Forrest "was a wonderful person" and "never harmed a fly." After this apparent agreement, the state court, "[f]or lack of a better term," "enter[ed] an interlocutory or temporary ruling granting the motion to strike the endorsement." The court then noted it would "make rulings as appropriate" "dependent on the evidence as it is presented at trial" and the court's ruling did "not foreclose counsel, either for the State or the defense, from addressing this issue again at a time they deem appropriate." Following this ruling, admission of evidence on the California homicide would have appeared to remain a danger. With Dr. Cunningham testifying on Forrest's historical behavior and risk of future violence, it was reasonable to fear Dr. Cunningham's testimony would tread too closely to the line and expose Forrest to damaging testimony with little to gain.

Forrest also claims questions about the California homicide would have been inadmissible during Forrest's penalty phase under State v. Debler, 856 S.W.2d 641 (Mo.1993) (en banc). However, "`[d]uring the penalty phase, both the state and the defense may introduce any evidence pertaining to the defendant's character,'" including "[e]vidence of a defendant's prior unadjudicated criminal conduct." State v. Christeson, 50 S.W.3d 251, 269 (Mo.2001) (en banc) (quoting State v. Ervin, 979 S.W.2d 149, 158 (Mo.1998) (en banc)). Debler simply requires that the defendant be given timely notice of the state's intent to use such evidence. See State v. Chambers, 891 S.W.2d 93, 106-07 (Mo.1994) (en banc); see also Ervin, 979 S.W.2d at 158 ("Since the decision in Chambers, this Court has consistently held that the error in Debler was lack of notice.").

"`Decisions relating to witness selection are normally left to counsel's judgment, and this judgment will not be second-guessed by hindsight.'" Hanes v. Dormire, 240 F.3d 694, 698 (8th Cir.2001) (quoting Williams v. Armontrout, 912 F.2d 924, 934 (8th Cir.1990) (en banc)). Here, the record shows, and the state circuit court found, defense counsel's decision was a strategic one, based on justifiable concerns about the helpfulness (and potential harmfulness) of Dr. Cunningham's testimony. Defense counsel's decision based on this risk calculation does not fall outside the bounds of Strickland.

4. Lay Witnesses

Lastly, Forrest asserts ineffective assistance in defense counsel's failure to interview and ultimately call three lay witnesses: Anthony Jacobs, Curtis Fuller, and Dennis Smock. Jacobs, Forrest's longtime friend, would have testified about Forrest's drug and alcohol abuse and that Forrest was good with children. One of defense counsel's investigators contacted Jacobs looking for another witness and during the call Jacobs told the investigator he had known Forrest. Fuller, Forrest's childhood neighbor, would have testified Forrest's father drank heavily and yelled at and hit Forrest when Forrest was a child. Turlington explained defense counsel did not investigate Forrest's childhood acquaintances. Smock, Forrest's former employer, would have testified Forrest was a hard worker, but required more supervision to complete even basic tasks. Smock's name appeared on a list of acquaintances which Forrest provided defense counsel. Turlington could not remember any strategic reason that defense counsel did not further investigate these witnesses.

The state circuit court concluded defense counsel was "not ineffective for failing to adduce cumulative evidence" from Jacobs, Fuller, and Smock. The Missouri Supreme Court agreed this did not amount to deficient performance. See Forrest II, 290 S.W.3d at 711. Forrest now challenges this conclusion, arguing Turlington's failure to articulate a strategic justification made defense counsel's inaction necessarily unreasonable. But this argument ignores the "strong presumption" of reasonableness permeating defense counsel's every decision, see Strickland, 466 U.S. at 689, 104 S.Ct. 2052, by apparently assuming that defense counsel must recall and provide (years after the facts) a conscious strategic foundation for every action — and for every action not taken. Strickland's presumption is not toppled so easily, and it is Forrest's burden to show that the course defense counsel took is not among the "countless ways to provide effective assistance" in this case. Id.

In any case, "the duty to investigate does not force defense lawyers to scour the globe on the off chance something will turn up; reasonably diligent counsel may draw a line when they have good reason to think further investigation would be a waste." Rompilla v. Beard, 545 U.S. 374, 383, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005). Indeed, Strickland itself presented a situation where "[c]ounsel's strategy ... decision not to seek more character or psychological evidence than was already in hand was ... reasonable." Strickland, 466 U.S. at 699, 104 S.Ct. 2052. And months after the Missouri Supreme Court rendered its decision, the Supreme Court again confirmed that "there comes a point at which evidence from more distant relatives can reasonably be expected to be only cumulative, and the search for it distractive from more important duties." Bobby v. Van Hook, 558 U.S. 4, 11, 130 S.Ct. 13, 175 L.Ed.2d 255 (2009) (per curiam); see also Hanes, 240 F.3d at 698 ("While some of the potential witnesses' testimony could have been helpful in rebutting or clarifying some collateral evidence, we do not believe any of the proffered testimony was so important as to put counsel's failure to consult with or call these witnesses outside the wide bounds of strategic choices that counsel is afforded.").

In this case, defense counsel found three experts to provide mental health testimony as well as fourteen lay witnesses willing to testify on Forrest's substance abuse, difficult childhood, quality as a person, and ability as a parent. We find nothing to suggest defense counsel should have recognized the investigation was materially incomplete, requiring additional fleshing out, or that either Jacobs, Fuller, or Smock was necessary for a reasonably complete investigation.

Forrest attempts to distinguish Smock as "the only former employer of [Forrest's] to testify" and Fuller as "the only nonfamily member who could corroborate [William's] trial testimony of [Forrest's] turbulent childhood." With numerous witnesses already testifying on the same subjects, Forrest has not satisfied his substantial burden of showing the failure to pursue these tenuous leads for cumulative testimony overcomes Strickland's presumption of reasonableness or makes defense counsel's otherwise thorough investigation unconstitutionally deficient. See Strickland, 466 U.S. at 691, 104 S.Ct. 2052 ("[A] particular decision not to investigate must be directly assessed for reasonableness in all the circumstances, applying a heavy measure of deference to counsel's judgments."). Relief cannot be granted for Forrest on this basis. See 28 U.S.C. § 2254(d)(1).

B. Claim-by-Claim Analysis

Forrest presented both the state circuit court and the Missouri Supreme Court with four discrete failures of defense counsel and a separate claim of error for each, having made no attempt to unite these failures under a common theory of ineffectiveness. Yet Forrest now claims the Missouri Supreme Court's decision "was contrary to" and "involved an unreasonable application of, clearly established Federal law," 28 U.S.C. § 2254(d)(1), insofar as the Missouri Supreme Court assessed claim by claim the reasonableness of defense counsel's performance. Forrest contends clear Supreme Court precedent obligated the Missouri Supreme Court to bundle the individual claims of attorney error and determine whether the body of these alleged faults, en masse, overcome Strickland's presumption of reasonableness. See Strickland, 466 U.S. at 689, 104 S.Ct. 2052. We do not understand the Strickland standard to demand this sort of cumulative performance inquiry, see Wainwright v. Lockhart, 80 F.3d 1226, 1233 (8th Cir. 1996) ("Errors that are not unconstitutional individually cannot be added together to create a constitutional violation. Neither cumulative effect of trial errors nor cumulative effect of attorney errors are grounds for habeas relief." (internal citation omitted)), nor does Forrest direct us to a Supreme Court decision supporting his assertion.

Forrest "carries the burden of pro[ving]" the requirements of § 2254(d) have been met-a "`difficult'" and "`highly deferential standard.'" Cullen, 563 U.S. at ___, 131 S.Ct. at 1398 (quoting Richter, 562 U.S. at ___, 131 S.Ct. at 786, and Woodford v. Visciotti, 537 U.S. 19, 24-25, 123 S.Ct. 357, 154 L.Ed.2d 279 (2002) (per curiam)). Forrest cites language in Strickland that "the performance inquiry must be whether counsel's assistance was reasonable considering all the circumstances." 466 U.S. at 688, 104 S.Ct. 2052 (emphasis added). Forrest also looks to the Supreme Court's statement that "[i]t will generally be appropriate for a reviewing court to assess counsel's overall performance throughout the case in order to determine whether the `identified acts or omissions' overcome the presumption that a counsel rendered reasonable professional assistance." Kimmelman v. Morrison, 477 U.S. 365, 386, 106 S.Ct. 2574, 91 L.Ed.2d 305 (1986) (quoting Strickland, 466 U.S. at 690, 104 S.Ct. 2052).

While this language plausibly may be read to support Forrest's cumulative view of performance, Forrest's burden under 28 U.S.C. § 2254(d)(1) is to show "the holdings, as opposed to the dicta, of [the Supreme] Court's decisions," Williams v. Taylor, 529 U.S. 362, 412, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000), provide a "clear answer to the question presented" in Forrest's favor, Wright v. Van Patten, 552 U.S. 120, 126, 128 S.Ct. 743, 169 L.Ed.2d 583 (2008). Though Kimmelman says it is "appropriate" to look to overall performance and "inadvisable" to do otherwise, 477 U.S. at 386, 106 S.Ct. 2574, such precatory language does not evince a holding establishing the clear rule Forrest proposes. By requiring that counsel's performance be "reasonable considering all the circumstances," 466 U.S. at 688, 104 S.Ct. 2052, Strickland does not purport to aggregate each discrete and potentially unrelated claim of ineffectiveness into a single performance inquiry. The statement contemplates that "[a] fair assessment of attorney performance requires that every effort be made ... to reconstruct the circumstances of counsel's challenged conduct, and to evaluate the conduct from counsel's perspective at the time." Id. at 689, 104 S.Ct. 2052. This the Missouri Supreme Court has done, basing its conclusions on evidence of defense counsel's knowledge and beliefs at the time of each decision. See Forrest II, 290 S.W.3d at 709-11, 715, 718.

The Missouri Supreme Court reasonably applied Strickland by analyzing individually each of the four discrete claims before the court and considering those circumstances appearing relevant to each claim.

C. Trial Strategy Standard

In describing the ineffective assistance standard, the Missouri Supreme Court stated "[t]rial strategy is not a basis for ineffective assistance of counsel." Forrest II, 290 S.W.3d at 708. Forrest contends "[t]his formulation of the Strickland performance test is contrary to Wiggins[v. Smith, 539 U.S. 510, 520-21, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003) ] because it failed to consider whether counsel's strategy was objectively reasonable under prevailing professional norms in light of the information available to counsel at the time of trial."

The Missouri Supreme Court made this statement shortly after recognizing that "[d]eficient performance exists when the representation is `below an objective standard of reasonableness.'" Forrest II, 290 S.W.3d at 708 (quoting Goodwin v. State, 191 S.W.3d 20, 25 (Mo.2006) (en banc)). And the court followed its statement by explaining that "decisions made after considering the law and facts and pondering alternative strategies generally are not disturbed by a court on review." Id. This largely paraphrases the intended operative standard that "strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable." Strickland, 466 U.S. at 690, 104 S.Ct. 2052; see also Abernathy, 748 F.3d at 816 (recognizing our obligation under Strickland to "`refrain[ ] from ... second-guessing of trial counsel's strategic decisions'" (quoting Nave v. Delo, 62 F.3d 1024, 1035 (8th Cir.1995))). While the "`shorthand reference' to the Strickland standard ... `may perhaps be imprecise,'" "[w]e are satisfied that the state courts understood the familiar Strickland standard and issued a decision that was not contrary to established federal law." Williams, 695 F.3d at 832 (quoting Woodford, 537 U.S. at 23-24, 123 S.Ct. 357).

III. CONCLUSION

With our review cast in the light of twice-magnified deference, see Richter, 562 U.S. at ___, 131 S.Ct. at 788; Strong, 737 F.3d at 517, Forrest cannot meet his burden to demonstrate the state courts erred and his defense "counsel made errors so serious that counsel was not functioning as the `counsel' guaranteed [him] by the Sixth Amendment." Strickland, 466 U.S. at 687, 104 S.Ct. 2052. Because Forrest has not shown the Missouri Supreme Court's decision was either "contrary to, or involved an unreasonable application of, clearly established Federal law" or "based on an unreasonable determination of the facts," 28 U.S.C. § 2254(d), we affirm the denial of Forrest's application for § 2254 relief.

BYE, Circuit Judge, dissenting.

I would reverse the district court and grant Earl Forrest a new sentencing. I believe Forrest's trial counsel were constitutionally ineffective in failing to present a PET scan at the penalty phase of Forrest's trial and the Missouri Supreme Court unreasonably determined trial counsels' actions did not constitute constitutionally ineffective assistance of counsel. See Forrest v. State, 290 S.W.3d 704 (Mo.2009). Therefore, I respectfully dissent.

The majority believes trial counsels' decision to forego a brain scan was "considered and strategic." I disagree. I believe the Supreme Court of Missouri made an unreasonable determination of the facts because the court failed to take into account that the testimony of defense experts was undermined by a lack of medical evidence.

The center of the defense strategy at trial was to show Forrest should not be sentenced to death because he suffered from brain damage. The defense called several experts to testify about Forrest's impaired functioning due to substance abuse, brain damage to Forrest's right frontal lobe, Forrest's impaired judgment and impulse control, and other cognitive disorders. During its cross-examination of defense experts and in closing arguments, the state effectively attacked the experts' findings of Forrest's brain damage by emphasizing there were no medical records or medical tests presented to corroborate the subjective opinions of these experts. The state repeatedly argued to the jury Forrest had failed to prove brain damage because the defense did not retain a physician to scan Forrest's brain to demonstrate the brain damage existed with objective scientific evidence.

During the penalty phase, the state challenged the findings and credibility of the defense experts by repeatedly characterizing their testimony as untrustworthy. As only one of myriad examples, during the penalty phase cross-examination, the state asked: "Is there a single objective feature which you can show us, an MRI, an x-ray, anything where you can point to something and say here is this brain damage, or here is this problem with his brain?" The defense expert was required to acknowledge no brain scans had been undertaken to corroborate the findings of brain damage. The state again reemphasized the fact no objective testing, including a PET scan, was conducted to prove Forrest was brain damaged or suffered from brain dysfunction. The state's strategy was effective-the jury foreman later told a newspaper the jury did not believe Forrest had brain damage based on the evidence presented. Appellant's Add. 30.

A PET scan would have neutralized the state's attacks about a lack of objective evidence of brain damage and would have provided the jury an appropriate view of such brain damage. Trial counsel should have realized having the PET scan available would help their case; in fact, trial counsel discussed a PET scan more than a year in advance yet failed to procure one, even after defense experts recommended a PET scan be procured. When trial counsel finally decided to procure a PET scan at the eleventh hour, counsel backed out when an MRI was requested. There is no strategy to this; counsel were simply ineffective. Being presented with a biological mechanism to account for violent behavior sways sentencing courts. Benedict Carey, Study of Judges Finds Evidence from Brain Scans Led to Lighter Sentences, N.Y. Times, August 16, 2012. The PET scan would have enabled the expert witnesses to propose a viable biological mechanism to show the jury Forrest's brain damage and should have been presented by defense counsel.

The majority also believes "the record indicates even favorable PET scan results... would merely have corroborated Forrest's existing experts." I disagree with this conclusion and believe the Missouri Supreme Court was unreasonable in determining a PET scan would have been cumulative. See Forrest, 290 S.W.3d at 709. Evidence regarding the main issue of a case-here, Forrest's brain damage — the decision of which turns on the weight of the evidence, cannot be cumulative because the weight and depth of the evidence is what was to be weighed by the jury. Black v. State, 151 S.W.3d 49, 56 (Mo. 2004). Additionally, a PET scan is a different device than an examination by a treating physician. The defense experts were all cross-examined on the issue of a lack of any test or objective evidence of brain damage. While a PET scan may have been cumulative in showing the same brain damage testified to by the experts, it would not have been cumulative as to its effect on the jury — the PET scan would have played a separate, unique roll in the penalty phase by providing objective evidence on which the jury could have relied to find Forrest had incurred brain damage. Moreover, the most compelling scientific research should have been presented if it were available, which it could have been, particularly when there was a substantial chance experts would be effectively cross-examined on such technology as being missing from this case.

Finally, the majority reasons defense counsel made a strategic decision in foregoing any brain scans for fear a scan would come back negative. The Missouri Supreme Court also found counsel acted strategically in avoiding a brain scan. Forrest, 290 S.W.3d at 709. I disagree and believe the Missouri Supreme Court was unreasonable in reaching such a conclusion. We do not, and should not, require all defense counsel presenting evidence of brain damage to obtain a PET scan of their client. However, this case presents an abnormal circumstance. Counsel knew at least a year in advance they planned to mount a defense strategy almost solely reliant on Forrest's brain damage, yet failed to obtain any objective evidence about Forrest's brain. Additionally, their own expert witnesses explicitly recommended trial counsel obtain PET scans of Forrest. Under these circumstances, trial counsel were ineffective in failing to pursue a PET scan or any other type of objective evidence of brain damage or brain abnormalities.

Before this court is "[t]he issue [of] whether the failure to discover and present evidence of [Forrest]'s mental condition undermines our confidence in the outcome of the ... sentencing phase of the trial." Antwine v. Delo, 54 F.3d 1357, 1365 (8th Cir.1995). The failure to present PET scan evidence showing brain damage strongly undermines my confidence in the outcome of the penalty phase of Forrest's trial. For these reasons, I would reverse the district court and would order a new penalty phase hearing where Forrest could present a PET scan as mitigating evidence.