CLAY, Circuit Judge.

NOT RECOMMENDED FOR FULL-TEXT PUBLICATION
File Name: 16a0372n.06

ON APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF MICHIGAN

BEFORE: COLE, Chief Judge; CLAY and GIBBONS, Circuit Judges.

CLAY, Circuit Judge. In 2005, a Michigan jury convicted Petitioner Terry Ceasor of one count of first-degree child abuse, in violation of Mich. Comp. Laws § 750.136b(2), based on allegations that he had caused his girlfriend's 16-month-old son, Brenden Genna, to suffer a subdural hematoma and retinal hemorrhaging. The prosecution's theory was that Ceasor inflicted these injuries by violently shaking or slamming Brenden while the baby's mother, Cheryl Genna, was out of the house. The linchpin of the prosecution's theory was the expert testimony of Dr. Holly Gilmer-Hill, who opined that Brenden's subdural hematoma and retinal hemorrhages were (1) symptoms commonly associated with shaken baby syndrome ("SBS"), (2) "caused by an intentional act," and (3) inconsistent with Ceasor's version of the facts—that Brenden's injuries resulted from an accidental fall from the couch. Following Ceasor's conviction, the trial court sentenced him to a term of two to 15 years in prison. On direct appeal, the Michigan Court of Appeals affirmed Ceasor's conviction and sentence, and the Michigan Supreme Court denied leave to appeal.

In 2008, Ceasor filed a petition for a writ of habeas corpus in the United States District Court for the Eastern District of Michigan. After staying Ceasor's federal habeas proceedings and permitting him to exhaust a claim for post-conviction relief in state court, the district court ultimately issued an order and judgment denying Ceasor's habeas petition. This timely appeal followed.

Ceasor's appeal raises two inextricable issues related to ineffective assistance of counsel. The first issue, which is more accurately characterized as a sub-issue of the main issue before this Court, is whether Ceasor has demonstrated the strength of his claim that his trial counsel rendered constitutionally ineffective assistance by failing to retain an expert witness to rebut Dr. Gilmer-Hill's testimony due to his ignorance (or misapprehension) of Michigan law governing public funding for indigent defendants. The second issue is whether Ceasor's appellate counsel rendered ineffective assistance by (1) relying solely on the trial record to support Ceasor's ineffective assistance of trial counsel claim and (2) failing to file a separate motion to remand, pursuant to Mich. Ct. R. 7.211(C)(1), for an evidentiary hearing under People v. Ginther, 212 N.W.2d 922 (Mich. 1973) (a "Ginther hearing"). For the following reasons, we REVERSE the district court's judgment denying habeas relief and REMAND for an evidentiary hearing on the merits of Ceasor's ineffective assistance of appellate counsel claim.

I. BACKGROUND

Ceasor's trial for first-degree child abuse lasted almost three days. Thereafter, the jury deliberated for nearly two days, and requested a deadlocked jury instruction, before returning a unanimous verdict of guilty. The following facts were averred to and established at trial.

A. Brenden's Injuries and Treatment

In June 2004, Ceasor began dating Cheryl Genna, Brenden Genna's mother. Both Genna and Ceasor described 16-month-old Brenden as an "active" child. Genna also had an older daughter, Derian, who was approximately seven-years-old at the time Brenden was injured.

At about 7:00 pm on October 2, 2004, Genna and her two children went to Ceasor's home in Port Huron, Michigan. They had visited the home before, and Genna's children had met and been alone with Ceasor "[s]everal times" in the past. At some point during the evening, Genna and Derian went to Blockbuster and China Lite, leaving Ceasor alone with Brenden. When they returned, Brenden was fine and showed no sign of injury.

On October 3, 2004, the day Brenden sustained his injuries, Genna woke up around 9:30 am and observed Ceasor coming back from giving Brenden his morning bottle. Both Genna and Ceasor—who testified in his own defense at trial—stated that it was normal for Ceasor to help care for Brenden by feeding him or changing his diaper. At around 10:30 am, Genna and Derian went to McDonald's to pick up breakfast while Ceasor remained at the house with Brenden. Once again, there was no indication that Brenden sustained injuries of any kind during the time he was alone with Ceasor.

Around 1:00 pm, Genna decided to drop Brenden off at his grandmother's house so she could take Derian swimming. However, when Ceasor saw that Brenden was still asleep, he told Genna "just let him sleep, go ahead and go," and Genna agreed to let Ceasor babysit while she and Derian went swimming.

Genna and Derian left Ceasor's house around 2:30 pm, and were away from the house for approximately an hour-and-a-half. The questions surrounding what occurred during this hour-and-a-half formed the basis of the prosecution's child abuse charge against Ceasor.

At trial, Ceasor testified to the following. Brenden was still sleeping for at least 20 to 30 minutes after Genna and Derian left the house, but Ceasor eventually brought Brenden out to the living room so the baby could sit on the couch with him while he watched football. Initially, Brenden was relatively inactive, but when Ceasor brought him a jar of Gerber bananas, fruit snacks, and a granola bar from the kitchen, Brenden became excited and began to eat. After cleaning up, and as he was returning from the kitchen, Ceasor saw that Brenden was standing on the couch and facing the TV. Ceasor crawled up to the back of the couch on his hands and knees and he and Brenden began playing a game called "gotcha," with Ceasor crawling behind the couch and Brenden running across the couch cushions. While they were playing, Ceasor noticed that Brenden's foot got stuck between the couch cushions a couple of times. Ceasor testified that during their game, Brenden was laughing and they were "having a good time."

When Brenden stopped playing in order to drink from his sippy cup, Ceasor "figured [he was] occupied enough" and stepped away to go to the bathroom. However, as he was urinating, Ceasor heard a thud that sounded like "two hits." Upon hearing the thud, Ceasor accidentally urinated on his hand, washed his hands without drying them, and ran out to the living room. There, he found Brenden wedged between the couch and the coffee table. At trial, Ceasor testified that there was "no way that [Brenden] went down in this position on his own," saying:

It wasn't like he was playing in this position. And, um, when I came out and saw him there, his head was, his head was flung back as far as the neck could go. And when I picked up [Brenden] he was like, it was like he was dead and he was like limp noodles . . . . I tried talking to him. I sprayed some water off my hands that were wet. Um, I touched his head. I, um, I tried everything I could do. I was calling his name.

Ceasor was on his way to call 911 when Genna and Derian came home from the pool. Genna initially laughed off Ceasor's statement that Brenden had fallen, thinking Ceasor was joking. However, when Ceasor said, "I'm serious," and Genna saw that Brenden was unconscious, she began screaming the baby's name. Rather than call an ambulance, Ceasor and Genna decided to use Genna's car to drive Brenden to the emergency room at Port Huron Hospital, which was located a couple of minutes away from Ceasor's house. Genna did not observe any injuries to Brenden at this time, but noticed that his hair was wet.

Both Ceasor and Genna testified that Brenden regained consciousness shortly after they arrived at Port Huron. Thereafter, Genna began calling family members, including her mother, sisters, and Brenden's father. Around the time Genna called Brenden's father, Ceasor informed the hospital staff that he was not Brenden's father, but Genna's boyfriend.

Dr. Christopher Hunt, Brenden's attending physician at Port Huron and one of the prosecution's witnesses at trial, testified that when he initially assessed Brenden, the baby's pupils were unequal in size and he was unresponsive to verbal commands and painful stimuli. Dr. Hunt did not observe or document any signs of trauma or injury, including retinal hemorrhaging, and noted that Brenden's breathing sounds were normal. The medical staff did not have to intubate Brenden, but took a CAT scan that Dr. Hunt later reviewed with a radiologist. The CAT scan revealed that Brenden had a subdural hematoma with a "slight mass effect," meaning that the blood beneath Brenden's dura had "start[ed] to push the brain to the opposite side" of his skull. Because a subdural hematoma is a serious injury, Dr. Hunt administered an anti-seizure medication to prevent continued swelling and ordered that Brenden be transferred to Children's Hospital in Detroit, where there was a pediatric neurosurgeon on staff.

When Dr. Hunt spoke to Genna and Ceasor about Brenden's history, he noted Ceasor as "the boyfriend." Dr. Hunt also documented a change in Ceasor's explanation about how Brenden sustained his injuries, observing that although Ceasor originally said that Brenden had fallen off the couch and hit his head on the coffee table, he later claimed that "he didn't know how [the injuries] occurred." (R. 7-7, PageID# 295-96). Concerned that Brenden's injuries were the result of "possible child abuse," Dr. Hunt testified that although subdural hematomas are frequently sustained by patients who have fallen and hit their heads, Brenden lacked the typical trauma associated with a fall-induced hematoma, including "soft tissue damage." He also conceded, however, that bruising does not always occur "right away."

Both Ceasor and Genna stayed with Brenden while he was in Port Huron's emergency room. Eventually, some of Genna's family members began to arrive, and Ceasor told Genna's sister that he "was there at the house with Brenden by [him]self" at the time the baby was injured. Later, however, Genna informed a police officer from the St. Clair County Sheriff's Department, Deputy Michael Garvin, that she was present when Brenden was injured and had picked him up after he fell. Genna gave this account while sitting approximately ten feet away from Ceasor in the Port Huron waiting room, and after Genna told this lie, Ceasor told Deputy Garvin and another police officer from the sheriff's department, Detective Terry Baker, that he had seen Genna "already kneeling next to" an unconscious Brenden just moments after the baby fell. At trial, Genna testified that she lied to Deputy Garvin about being present when Brenden was injured because she was "afraid," "scared," and "in shock." She also testified that at the time of Brenden's injuries, she was embroiled in a dispute with Brenden's father over child support. Both Genna and Ceasor testified that Ceasor never told Genna to lie to the police.

At some point while Brenden was hospitalized, Genna noticed an ovular "mark on the back of [Brenden's] head" that was two or two-and-a-half inches long and "had red dots on it." (R. 7-6, PageID# 232-33). Although Genna testified that she told the medical staff at Children's about the red mark on Brenden's head, she could not recall whether she told medical personnel at either hospital about a second injury—a bite mark—that she observed on his tongue.

Around 7:25 pm, an ambulance transported Brenden and Genna from Port Huron Hospital to Children's Hospital in Detroit. One of Brenden's attending physicians at Children's was Dr. Holly Gilmer-Hill, the expert witness at Ceasor's trial.

Although a resident treated Brenden the day he was transferred from Port Huron, Dr. Gilmer-Hill discussed the baby's injuries with his parents the following day, October 4, 2004. During this conversation, Genna told Dr. Gilmer-Hill that Brenden had fallen off the couch. Dr. Gilmer-Hill physically examined Brenden, but did not notice or document "any external bruising or swelling of the scalp." Her review of Brenden's CAT scan revealed that he had sustained a "subdural hemorrhage" with "shift," i.e., pressure caused by blood and swelling. Subsequently, Brenden was placed under observation and given Mannitol—a medication that draws fluid out of the brain—in order to minimize swelling.

When she saw Brenden again on October 6, Dr. Gilmer-Hill learned that the Children's ophthalmology staff had found retinal hemorrhages in both of Brenden's eyes. At trial, Dr. Gilmer-Hill testified that "[i]t takes a good deal of force to cause [a retinal hemorrhage], and the combination of subdural blood with retinal hemorrhage is child abuse. It is patently demonic. [It] [i]s diagnostic for child abuse." (R. 7-7, PageID# 319). When pressed on this point by the prosecution, Dr. Gilmer-Hill opined that "being [severely] shaken or slammed onto a surface, either hard or soft," would cause retinal hemorrhaging. (Id.).

Brenden remained at Children's Hospital between Sunday, October 3—the day he was injured and admitted—and the following Friday, October 8. He was not prescribed any medications at the time of or following discharge. After seeing Brenden on October 6, Dr. Gilmer-Hill did not treat him again before he was released, and she admitted on cross-examination that she only observed Brenden for approximately 25 minutes during the entirety of his five-day stay. Dr. Gilmer-Hill also conceded that: Brenden's skeletal x-ray did not reveal any broken or fractured bones; there were no marks on Brenden's body suggesting that someone had vigorously held or shaken him; Brenden did not need surgery to treat his injuries; and Brenden did not experience seizures while being treated at Children's.

According to Genna, by the time he was discharged on October 8, Brenden was "doing better" and was walking and talking. However, because Brenden was placed in his father's custody following discharge, Genna could not testify as to whether he experienced any injury-related difficulties in the three months following his hospital stay. However, at the time of Ceasor's trial, which was held a little more than a year later after Brenden was injured, Genna testified that Brenden was "doing great" and did not need to follow up or report for regular check-ups at either hospital.

B. Trial Testimony Regarding the Cause of Brenden's Injuries

At trial, the prosecution's case against Ceasor was based almost exclusively on the expert testimony of Dr. Gilmer-Hill. Relevant to this appeal, Dr. Gilmer-Hill opined that: SBS is an "accepted" syndrome in the medical community; subdural hematomas and retinal hemorrhages, taken together, are symptoms "diagnostic of child abuse"; one cause of retinal hemorrhaging is shaking a child or slamming him against a hard or soft surface; Brenden's injuries were not accidental because "a fall from a couch onto a carpeted floor . . . [could] not account for [those] injuries"; the degree of force that caused Brenden's injuries was more akin to "a fall out of a second story window" or "a high speed car accident" than a fall from a couch; and Ceasor's and Genna's inconsistent accounts of whether Genna was present when Brenden was injured suggested that Brenden had been abused because the abusers in SBS cases frequently offer conflicting or inconsistent accounts about how the victim's injuries occurred.

With regard to SBS generally, Dr. Gilmer-Hill testified as follows:

Typically[,] [SBS] involves shaking of [an] infant, usually a child less than two years old. Violent shaking. Not just shaking a child . . . a little bit to revive them or because they have fainted or something like that, but really violently shaking the child such that the head whips back and forth on the body, which is the axis.
The head is larger relative to the body in a child than it is in an adult, and so it causes . . . a big lever of force, and it causes severe forces within the head. The brain is not fixed within the skull and it can move. So, the brain slams back and forth inside the skull. The bridging veins between the brain and the skull can tear, which can cause a subdural hemorrhage. The[ victim] can [experience] bleeding in the back of the eye, which is [a] retinal hemorrhage[,] from the force of the shaking, and usually it involves an aspect of impact, too. Usually the child is struck as well, or slammed down on a, a sofa or a soft surface, even against a wall or thrown up against the ceiling. There are a lot of variations.

To support her opinion about the alleged link between subdural hematomas, retinal hemorrhages, and SBS, Dr. Gilmer-Hill referenced, among other things, an experimental study conducted by Dr. Ann-Christine Duhaime, an American neurosurgeon. On cross-examination, Dr. Gilmer-Hill stated that Duhaime's study simulated accidental injuries using cats and rats. When Ceasor's trial counsel asked Dr. Gilmer-Hill about a study conducted by forensic pathologist Dr. John Plunkett, which found that children could sustain serious trauma or even fatal head injuries from short falls, Dr. Gilmer-Hill rejected Plunkett's study on the ground that it was not "widely accepted" in her profession. Additionally, Dr. Gilmer-Hill disagreed with Dr. Plunkett's finding that children can sustain serious or fatal injuries "even on carpeted surfaces."

On re-cross, Dr. Gilmer-Hill stated that her expertise on SBS was derived from American neurosurgical literature (as opposed to literature published in other countries or by professionals in other fields). When asked whether she habitually discounted the findings of medical studies with which she disagreed, Dr. Gilmer-Hill stated that she had not "seen any well-done, rigorous studies that disagreed with [her] theory [about SBS]." (Id. at 329-30).

Finally, Dr. Gilmer-Hill testified that Brenden's injuries were non-accidental because the history offered by Ceasor and Genna—a fall from the couch—was "not consistent with the mechanism of the injur[ies]." (Id. at 315, 319-20). Although Dr. Gilmer-Hill conceded that she was not an expert in biomechanics, she nonetheless testified that a child falling from a height of five or six feet experiences "[m]uch less" gravitational force than a child being shaken. Further, when asked about nurse's notes generated at Children's that both documented and diagrammed that Brenden had "bruising to the forehead"—an injury potentially consistent with an accidental fall—Dr. Gilmer-Hill testified that she did not rely on this part of Brenden's history in formulating her opinion because the notes were the "only place" that documented bruising and she "[did not] see the bruise [her]self." (R. 7-7, PageID# 323). Dr. Gilmer-Hill also claimed not to have seen an intake form from Children's pediatric surgery service documenting that Brenden had oral redness.

After the prosecution rested its case, Ceasor testified on his own behalf as the defense's only witness. Among other things, Ceasor corroborated Genna's earlier testimony that Ceasor frequently helped care for Brenden, had never expressed impatience with Brenden being an "active" child, and had never physically disciplined Genna's children or his own thirteen-year-old son. Ceasor also specifically denied being upset with or feeling the need to discipline Brenden on the day he was injured. At the conclusion of Ceasor's testimony, the defense rested its case.

C. Jury Deliberations and Verdict

The jury deliberated for two days, watching a video recording of Dr. Gilmer-Hill's expert testimony twice and requesting "further instructions" from the trial court after failing to reach a unanimous verdict. At the end of the second day of deliberations, the jury returned a unanimous verdict of guilty, and Ceasor was sentenced to a prison term of two to 15 years.

D. Procedural Background

On direct appeal, Ceasor's appellate counsel argued, inter alia, that Ceasor had been deprived of his right to effective assistance of counsel under the Sixth Amendment of the United States Constitution because trial counsel "fail[ed] to contact, investigate, and urge Ceasor to hire an expert to refute the medical testimony presented by the prosecution." (R. 7-12, PageID# 572). More specifically, appellate counsel maintained that trial counsel's failure to retain an expert was objectively unreasonable because: (1) expert testimony could have "directly refuted" Dr. Gilmer-Hill's conclusion that Brenden's injuries were non-accidental; (2) SBS "is a hotly contested matter" in the scientific community; and (3) issues involving brain injury "are generally beyond easy understanding by legal professionals or lay jurors," "require[] precise attention to the particular facts" of the case, and "demand[] consult[ation] with qualified experts." (Id. at 574-75). Appellate counsel also argued that Ceasor was prejudiced by trial counsel's failure to retain an expert because there was a "high probability" that the jury would have rendered a verdict of not guilty if the defense had offered expert testimony "to refute the prosecutor's claims and support the defense['s] theory that [Brenden] was injured during an accidental fall." (Id. at 575-76). Finally, appellate counsel asserted that even if Ceasor could not afford to retain an expert, trial counsel should have requested public funding to pay for consultation with an expert. See Mich. Comp. Laws § 775.15.

The Michigan Court of Appeals affirmed Ceasor's conviction in an unpublished opinion dated July 12, 2007. People v. Ceasor, No. 268150, 2007 WL 2011747 (Mich. Ct. App. July 12, 2007). At the outset, the court of appeals noted that "[b]ecause [Ceasor] did not move for a Ginther hearing, [the court's] review [wa]s limited to errors apparent on the record." Id. at *3 (footnote omitted). Indeed, appellate counsel failed to move for a Ginther hearing—which allows a defendant to proffer facts or evidence in support of his ineffective assistance of counsel claim—by filing a separate motion to remand under Mich. Ct. R. 7.211(C)(1) ("Rule 7.211(C)(1)"). Instead, counsel represented that the "record alone show[ed] that [trial] counsel was ineffective" and conditioned Ceasor's request for a Ginther hearing on a finding that the "record [wa]s lacking." (R. 7-12, PageID# 575).

Examining only the trial record, the Michigan Court of Appeals found that even if there were scientific studies that conflicted with Dr. Gilmer-Hill's trial testimony, the existing record did not show "that an expert would have been willing to opine," based on the specific facts in Ceasor's case, that Brenden's injuries were not the result of being shaken or slammed or that his injuries were accidental and unrelated to abuse. Ceasor, 2007 WL 2011747, at *4 (emphasis added). The court also concluded that "the record [did] not support [Ceasor's] contention that his counsel failed to contact or try to procure an expert to support [the defense's] theory" because there was evidence that trial counsel sought and was granted additional time to consult an expert witness. Id. In light of the evidence that trial counsel at least attempted to locate an expert witness to testify on Ceasor's behalf, as well as the "presumption that [trial] counsel's decision to not call an expert witness was a matter of sound trial strategy," the court concluded that "[t]he fact that SBS may be a disputed diagnosis does not mean that an expert would have found after reviewing the evidence that [Brenden's] injuries resulted from an accident." Id. Following the Michigan Court of Appeals' affirmance of Ceasor's conviction, the Michigan Supreme Court denied leave to appeal on October 29, 2007. People v. Ceasor, 740 N.W.2d 257 (Mich. 2007)

On August 21, 2008, Ceasor filed a pro se petition for a writ of habeas corpus in the district court. In February 2010, he filed an amended habeas petition through counsel. Pursuant to the parties' stipulation, however, the district court ultimately stayed Ceasor's federal habeas proceedings so that he could exhaust a claim for post-conviction relief in state court.

Before the state trial court, Ceasor filed a motion for relief from judgment arguing that appellate counsel rendered ineffective assistance when he failed to file a separate motion to remand for a Ginther hearing, pursuant to Rule 7.211(C)(1), in raising Ceasor's ineffective assistance of trial counsel claim. More specifically, Ceasor observed that the Michigan Court of Appeals rejected his ineffective assistance of trial counsel claim on the grounds that there was insufficient evidence in the "existing [trial] record" to rebut the presumption that trial counsel's decision not to call an expert witness was a matter of "sound trial strategy." Ceasor, 2007 WL 2011747, at *3-4. The court of appeals only reached this conclusion, Ceasor argued, because appellate counsel neglected to file a proper motion to remand for a Ginther hearing under Rule 7.211(C)(1) and instead relied on a trial court record that was "devoid of any facts related to [trial] counsel's failure to hire an expert." Relatedly, Ceasor argued that if appellate counsel had properly moved for a Ginther hearing under Rule 7.211(C)(1), Ceasor would have established on remand that:

i. An expert would have informed the jury of the growing body of evidence undermining the [SBS] hypothesis.
ii. An expert would have countered Dr. Gilmer-Hill's many incorrect assertions and would have prepared trial counsel to effectively cross-examine Dr. Gilmer-Hill.
iii. An expert could have offered testimony that the injuries Brenden Genna suffered were consistent with a short fall, inconsistent with shaking, and unlikely to have been caused by intentionally applied force.
iv. Trial counsel failed to call an expert not because of trial strategy but solely because [Ceasor] could not afford one, and trial counsel never considered asking the Court for funds to hire an expert.

To support these contentions, Ceasor proffered the affidavits of four experts who reviewed his case pro bono: Dr. John Plunkett, a forensic pathologist; Dr. Peter Stephens, a medical doctor board certified in anatomical pathology, clinical pathology, and forensic pathology; Dr. Ronald Uscinksi, a clinical neurosurgeon with "special expertise in the literature surrounding pediatric head injuries"; and Dr. Christopher Van Ee, a Ph.D. in biomedical engineering. Together, these experts swore to the following. Since approximately 2004, the theories underlying SBS have been challenged and called into question due to their purported lack of a scientific basis. This is, at least in part, because the biomechanical and forensic literature demonstrates that shaking without impact is unlikely to cause subdural hematomas or retinal hemorrhages. Instead, such injuries may result from an accidental impact, including a short fall, or from a variety of natural causes. In fact, the gravitational force from the impact of a short fall far exceeds the force from shaking, and short falls can result in serious or even fatal head injuries. Further, injury biomechanics confirm that when a child is manually shaken, he or she will suffer a neck injury or gripping-style chest injuries well before sustaining a subdural hematoma or retinal hemorrhage.

With regard to Ceasor's case in particular, the experts opined that Brenden's injuries were consistent with a short fall from the couch onto the coffee table or the floor and inconsistent with abusive shaking. In evaluating Dr. Gilmer-Hill's expert testimony at trial, the experts asserted that Dr. Gilmer-Hill had misrepresented the findings in Dr. Duhaime's study and given the jury incorrect information regarding the biomechanics of infant head injury, short distance falls, and abusive shaking. They also noted that Dr. Gilmer-Hill's apparent misapprehension of the literature on SBS and pediatric head injury likely stemmed, at least in part, from her limited focus on American neurosurgical literature at the exclusion of international literature and literature from other disciplines, including pathology, pediatrics, and biomechanics. Each of the experts expounded on criticisms of SBS (or Dr. Gilmer-Hill's understanding thereof) that existed at the time of Ceasor's December 2005 trial or earlier, and Drs. Plunkett and Van Ee expressly represented that they would have offered the opinions included in their respective affidavits if they had been asked to testify at trial.

Finally, Ceasor attached affidavits from himself and his uncle—who had accompanied Ceasor to each of his meetings with trial counsel—swearing to the following. Trial counsel informed Ceasor that he would need an expert witness "in order to succeed at trial due to the complexity of the medical issues involved" in his case. (R. 24-2, PageID# 1368, 1371). At some point during summer 2005, trial counsel consulted with a potential expert witness, Dr. Faris Bandak. Thereafter, trial counsel asserted that Ceasor owed Dr. Bandak $1,500 for this initial consultation and would need to pay an additional fee of at least $10,000 in order to compensate Dr. Bandak for testifying at trial. When Ceasor informed trial counsel that he could not afford these fees because he had already exhausted his and his family's financial resources by retaining an attorney, trial counsel "refused to entertain other options for expert testimony" and, as a result, "never retained an expert for . . . trial." (Id. at 1369, 1372, 1374). Trial counsel also failed to pursue "other avenues to attain an expert for [Ceasor], such as petitioning the court for fees for an expert due to [Ceasor's] indigency." (Id. at 1249).

Overall, Ceasor's main argument before the trial court was that appellate counsel's failure to request a Ginther hearing "precluded [him] from presenting the evidence essential to support his [ineffective assistance of trial counsel] claim." (Id. at 1254). Such evidence included: (1) testimony from expert witnesses revealing the flaws affecting Dr. Gilmer-Hill's trial testimony and affirmatively demonstrating that Brenden's injuries were consistent with Ceasor's version of the facts; and (2) testimony from lay witnesses demonstrating that trial counsel's decision not to retain an expert witness was not a "reasonable trial strategy," but instead based on Ceasor's inability to pay the fee of a specific expert. The final page of Ceasor's motion noted that "[t]his claim for relief has not been raised previously and could not have been raised on direct appeal because [Ceasor] was represented on direct appeal by the appellate attorney who provided the ineffective assistance that is the basis of this motion." (Id. at 1255).

Citing Mich. Comp. Laws § 6.508(D) ("Rule 6.508(D)"), the trial court denied Ceasor's motion in a three-page order, reasoning that Ceasor's ineffective assistance of appellate counsel claim was merely his ineffective assistance of trial counsel claim "re-framed" in pursuit of a different result. Consequently, the trial court did not assess whether appellate counsel performed deficiently by failing to file a motion to remand for a Ginther hearing under Rule 7.211(C)(1), let alone mention the affidavits filed with Ceasor's motion. Relying on Rule 6.508(D), the Michigan Court of Appeals and Michigan Supreme Court denied leave to appeal.

On June 20, 2012, Ceasor filed a second amended habeas petition in the district court, and the district court re-opened his federal habeas proceedings. As exhibits to his habeas memorandum, Ceasor filed the same affidavits previously presented to (but apparently not considered by) the trial court. On January 13, 2015, the district court issued an order and judgment denying Ceasor's habeas petition.

In its order denying habeas relief, the district court observed that the trial court failed to address the merits of Ceasor's ineffective assistance of appellate counsel claim "on the erroneous ground that the claim had already been raised and decided against [Ceasor] by the Michigan Court of Appeals on direct review." Ceasor v. Ocwieja, No. 5:08-CV-13641, 2015 WL 164008, at *4 (E.D. Mich. Jan. 13, 2015). Since the state courts generated no results or reasoning to which the district court could defer on Ceasor's ineffective assistance of appellate counsel claim, the district court reviewed this claim de novo.

Ultimately, the district court found that appellate counsel's decision not to file a separate motion for an evidentiary hearing was a "reasonable recognition that the allegations of ineffective assistance could be determined from the trial transcript alone." Id. at *6 (citation and quotation marks omitted). The court also found that even "assuming that appellate counsel was ineffective for failing to move for a Ginther hearing," Ceasor could not show prejudice because he had not demonstrated that "a Ginther hearing would have been granted or that the Michigan Court of Appeals would have reversed his conviction had such a hearing been held." Id.

Ceasor timely appealed the district court's denial of his habeas petition and denial of a certificate of appealability, and this Court granted a certificate of appealability as to Ceasor's ineffective assistance of appellate counsel claim.

II. DISCUSSION

A. Jurisdiction

Although Respondent John Ocwieja (the "Warden"), does not raise a jurisdictional challenge to Ceasor's habeas appeal, because Ceasor is no longer in prison, we find it necessary to briefly discuss the issue of jurisdiction. See Steel Co. v. Citizens for a Better Env't, 523 U.S. 83, 95 (1998). As noted in our order granting a certificate of appealability, Ceasor was paroled in October 2009. "The federal habeas statute gives the United States district courts jurisdiction to entertain petitions for habeas relief only from persons who are 'in custody in violation of the Constitution or laws or treaties of the United States.'" Maleng v. Cook, 490 U.S. 488, 490 (1989) (citing 28 U.S.C. § 2241(c)(3); 28 U.S.C. § 2254(a)) (emphasis in original). The Supreme Court has "interpreted the statutory language as requiring that the habeas petitioner be 'in custody' under the conviction or sentence under attack at the time his [habeas] petition is filed." Id. at 490-91 (emphasis added). Because Ceasor filed the instant habeas petition in August 2008, and remained in custody until October 2009, the district court had jurisdiction to entertain Ceasor's habeas petition under 28 U.S.C. §§ 2241 and 2254, and this Court has jurisdiction to hear the appeal thereof pursuant to 28 U.S.C. § 1291.

Ceasor's appeal raises a second jurisdictional question: whether his petition for habeas relief has been mooted by his release from prison. See Demis v. Sniezek, 558 F.3d 508, 512 (6th Cir. 2009) ("Because the exercise of judicial power under Article III of the Constitution depends on the existence of a live case or controversy, mootness is a jurisdictional question."). "Article III, Section 2 of the United States Constitution authorizes the federal judiciary only to hear cases or controversies." Gentry v. Deuth, 456 F.3d 687, 693 (6th Cir. 2006). "[T]herefore[,] federal courts may not exercise jurisdiction when the controversy has been mooted, that is to say, when the 'issues presented are no longer live or the parties lack a legally cognizable interest in the outcome.'" Id. (quoting Los Angeles County v. Davis, 440 U.S. 625, 631 (1979)) (internal quotation marks omitted). Nonetheless, we have previously recognized the following with regard to habeas petitions in particular:

Although the Supreme Court had seemed to limit habeas relief to "the body of the petitioner" in Fay v. Noia, 372 U.S. 391, 430-31 (1963), . . . the Court subsequently expanded the writ's scope in Carafas v. LaVallee, 391 U.S. 234, 237 (1968), stating that the petitioner's challenge was not mooted by his release from incarceration prior to his hearing because, "in consequence of his conviction, he cannot engage in certain businesses; he cannot serve as an official of a labor union for a specified period of time; he cannot vote in any election held in New York State; he cannot serve as a juror." Ibid. Therefore, "[o]n account of these 'collateral consequences,' the case is not moot." Id. at 237-38.

Carafas and its progeny, including several opinions issued by this Court, have recognized that "the appropriate remedy for a writ of habeas corpus issued pursuant to an unlawful criminal conviction includes relief not only from the conviction's direct consequences (e.g. incarceration), but also from its collateral consequences." Id. (emphasis in original); see, e.g., Benton v. Maryland, 395 U.S. 784, 790-91 (1969); Sibron v. New York, 392 U.S. 40, 55 (1968); Abela v. Martin, 380 F.3d 915, 921 (6th Cir. 2004), abrogated on other grounds by Guilmette v. Howes, 624 F.3d 286 (6th Cir. 2010); Green v. Arn, 839 F.2d 300, 302 (6th Cir. 1988). Further, the Supreme Court has "allowed federal courts to presume the existence of collateral consequences" in cases where the petitioner is challenging an allegedly unconstitutional conviction, Gentry, 456 F.3d at 694-95, stating, "we have been willing to presume that a wrongful criminal conviction has continuing collateral consequences (or, what is effectively the same, to count collateral consequences that are remote and unlikely to occur)." Spencer v. Kemna, 523 U.S. 1, 8 (1998).

Accordingly, we will presume the existence of collateral consequences flowing from Ceasor's allegedly wrongful conviction. See id.; see also Leyva v. Williams, 504 F.3d 357, 363 (3d Cir. 2007). Because Ceasor's habeas petition has not been mooted by his release from prison, we proceed to the merits of his claim.

B. Standard of Review

In habeas proceedings, we review the "district court's legal conclusions de novo and its factual findings for clear error." Smith v. Mitchell, 567 F.3d 246, 255 (6th Cir. 2009) (citation and quotation marks omitted). Our review of the state court's decision, on the other hand, is generally "governed by the standards set forth in the Antiterrorism & Effective Death Penalty Act of 1996," also known as "AEDPA." Barnes v. Elo, 231 F.3d 1025, 1028 (6th Cir. 2000). Under AEDPA, "a federal court shall not grant a petition for a writ of habeas corpus unless the state court adjudication of the claim '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.'" Id. (quoting 28 U.S.C. § 2254(d)(1)). This highly deferential standard of review is commonly known as "AEDPA deference." See Fitzpatrick v. Robinson, 723 F.3d 624, 634, 637 (6th Cir. 2013).

AEDPA deference is warranted in those cases where the state court "put[s] forward a merits-based ground for denying post-conviction relief—even if the court also posits an alternative, procedural ground for denying relief. Moritz v. Lafler, 525 F. App'x 277, 284 (6th Cir. 2013) (citing Hoffner v. Bradshaw, 622 F.3d 487, 505 (6th Cir. 2010)); Brooks v. Bagley, 513 F.3d 618, 624-25 (6th Cir. 2008). However, this Court will review a federal habeas claim de novo where there "there are simply no results [on the merits], let alone reasoning, to which this [C]ourt can defer." McKenzie v. Smith, 326 F.3d 721, 727 (6th Cir. 2003); see also Thompson v. Bell, 580 F.3d 423, 439 (6th Cir. 2009) ("Because the Tennessee state courts did not adjudicate Thompson's chemical competency claim on the merits, there is no state court decision to which this Court can defer pursuant to 18 U.S.C. § 2254(d).").

In the instant case, the trial court relied on Rule 6.508(D)(2) to deny Ceasor's post-conviction motion on the grounds that the claim raised by the motion had already been raised before and decided by the Michigan Court of Appeals. As we have previously recognized, a petitioner's claims may be procedurally defaulted under Rule 6.508(D) where the petitioner did not present his claims "in accordance with [Michigan's] procedural rules." Simpson v. Jones, 238 F.3d 399, 405-06 (6th Cir. 2000). However, as acknowledged by the district court and each of the parties below, Ceasor's ineffective assistance of appellate counsel claim was never reviewed by the Michigan Court of Appeals, and therefore could not have been decided against him. Cf. Mich. Comp. Laws § 6.508(D)(2). Instead, the court of appeals only addressed his ineffective assistance of trial counsel claim, which itself was presented by his allegedly ineffective appellate attorney. Nonetheless, after erroneously finding that Ceasor's ineffective assistance of appellate counsel claim had already been adjudicated on direct appeal, the trial court neglected to assess whether appellate counsel performed deficiently by failing to file a motion to remand for a Ginther hearing or whether such a failure was prejudicial. Additionally, the court made no mention of Ceasor's argument that appellate counsel was required to file a separate motion to remand under Rule 7.211(C)(1), and failed to evaluate whether it was unreasonable for appellate counsel not to file a motion to remand in light of the alleged absence of facts in the trial record regarding trial counsel's failure to hire an expert. In other words, the trial court failed to review Ceasor's ineffective assistance of appellate counsel claim on the merits. Because the trial court did not adjudicate this claim on the merits, let alone provide any reasoning by which this Court may assess its decision, the standard of review is de novo. See Thompson, 580 F.3d at 439; McKenzie, 326 F.3d at 727.

C. Analysis

Under the Supreme Court's holding in Evitts v. Lucey, 469 U.S. 387, 396 (1985), a defendant has a "constitutional right to counsel on his first appeal [as of right]." Lutze v. Sherry, 392 F. App'x 455, 458 (6th Cir. 2010). This right encompasses "the right to the effective assistance of counsel." Evitts, 469 U.S. at 397 (emphasis added); see also Evans v. Hudson, 575 F.3d 560, 564 (6th Cir. 2009) ("On an appeal of right, a criminal defendant is entitled to effective assistance of appellate counsel.").

The two-prong test set out in Strickland v. Washington, 466 U.S. 668 (1984), governs claims of ineffective assistance of counsel, including claims of ineffective assistance of appellate counsel. See Evans, 575 F.3d at 564 (citing Mahdi v. Bagley, 522 F.3d 631, 636 (6th Cir. 2008)). Strickland's first prong requires Ceasor to demonstrate that "counsel's representation was deficient in that it 'fell below an objective standard of reasonableness.'" Towns v. Smith, 395 F.3d 251, 258 (6th Cir. 2005) (quoting Strickland, 466 U.S. at 688). In assessing whether counsel's performance was deficient, this Court "must indulge a strong presumption that counsel's conduct falls within the wide range of reasonable professional assistance," and Ceasor "must overcome the presumption that, under the circumstances, the challenged action might be considered sound trial strategy." Strickland, 466 U.S. at 689 (citation and quotation marks omitted). The second prong requires Ceasor to show "prejudice," i.e., "a reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different." Id. at 694. "A reasonable probability is a probability sufficient to undermine confidence in the outcome." Id.

In evaluating ineffective assistance of appellate counsel claims, "we assess the strength of the claim appellate counsel failed to raise." Wilson v. Parker, 515 F.3d 682, 707 (6th Cir. 2008). Accordingly, we will grant habeas relief only if "there is a reasonable probability that inclusion of the issue would have changed the result of the appeal." McFarland v. Yukins, 356 F.3d 688, 699 (6th Cir. 2004).

1. Ineffective Assistance of Appellate Counsel

a. Deficient Performance

In support of his ineffective assistance of appellate counsel claim, Ceasor asserts that "[a]ppellate counsel's performance was deficient under Strickland because it was objectively unreasonable for him to fail to file a separate motion for the evidentiary hearing that was both required under the Michigan Court Rules and absolutely necessary to establish the record for the claim of ineffective assistance of trial counsel." Appellant's Br. at 40; see also Reply Br. at 1-5. In response, the Warden argues that "appellate counsel's decision to seek reversal on the existing record, rather than to file a separate motion seeking a remand, did not fall below an objective standard of reasonableness . . . where appellate counsel specifically indicated in his brief that if the appellate court found that the record was lacking, he requested a remand for a Ginther hearing." Appellee's Br. at 17-18 (emphasis in original). For the reasons stated below, we reject the Warden's argument.

The Michigan Supreme Court has held that "[i]f a convicted defendant believes that his attorney's representation was below an objective standard of reasonableness, the appropriate procedure is to seek a Ginther hearing." People v. Smith, 581 N.W.2d 654, 660 (Mich. 1998). "The purpose of a Ginther hearing is to allow a defendant to establish facts or evidence to assist in making his claims." Parker, 2004 WL 1392292, at *6 (citing Ginther, 212 N.W.2d at 925). "A defendant is not entitled to a Ginther hearing as a matter of right," but instead "must demonstrate that there are factual issues regarding his or her counsel's performance that require further inquiry." People v. Randolph, No. 293999, 2010 WL 5383526, at *3 (Mich. Ct. App. Dec. 28, 2010).

Here, appellate counsel did not move for a Ginther hearing under Rule 7.211(C)(1). Instead, counsel asserted the following in his brief:

[SBS] is a hotly contested matter as is evidenced by [trial] counsel's cross-examination of the prosecution's medical experts. The complexity of brain injury cases is unquestioned. The issues are generally beyond easy understanding by legal professionals or lay jurors[, and] [h]andling these cases requires precise attention to the particular facts and demands consult with qualified experts. It is blatantly obvious from the record that a defense expert was needed. The record alone shows that counsel was ineffective. But, if this Court finds that the record is lacking, then Mr. Ceasor requests that this Court remand this case to the trial court for a Ginther hearing.

In order to evaluate the reasonableness of appellate counsel's conduct, we must examine the state of the law at the time of the appeal. See Jacobs v. Mohr, 265 F.3d 407, 418 (6th Cir. 2001) ("[T]he reasonableness of counsel's performance is to be evaluated from counsel's perspective at the time of the alleged error and in light of all the circumstances") (citation and quotation marks omitted); see also, e.g., Goff v. Bagley, 601 F.3d 445, 464-66 (6th Cir. 2010) (evaluating the reasonableness of appellate counsel's conduct "[u]nder Ohio law at the time of [the petitioner's] direct appeal"). In People v. Nantelle, a published 1996 case, the Michigan Court of Appeals held that where a defendant "[does] not move for a Ginther hearing" on his ineffective assistance of counsel claim, the court's review is "limited to mistakes apparent on the record." 544 N.W.2d at 673 (emphasis added). The court of appeals reiterated this rule in subsequent opinions issued prior to Ceasor's appeal, rejecting a defendant's request for an evidentiary hearing when it did not appear in a separate motion to remand under Rule 7.211(C)(1). See, e.g., People v. Fisher, No. 262961, 2007 WL 283799, at *2 n.2 (Mich. Ct. App. Feb. 1, 2007); People v. Carter, No. 232862, 2003 WL 887594, at *4 (Mich. Ct. App. Mar. 6, 2003). In the instant case, when the court of appeals limited its review of Ceasor's ineffective assistance of trial counsel claim to "errors apparent on the record" based on appellate counsel's failure to move for a Ginther hearing, the court cited Nantelle in support. Nantelle, as a published case, is precedentially binding on future panels of the Michigan Court of Appeals. See Mich. Ct. R. 7.215(C)(2), (J)(1).

The Warden cites two unpublished 2013 opinions for the proposition that although it may have been a "best practice" for appellate counsel to file a separate motion to remand under Rule 7.211(C), "it [wa]s not necessarily deficient performance to fail to do so." Appellee's Br. at 28 (citing People v. Henry, Nos. 306449, 308963, 2013 WL 6331731, at *5 n.1 (Mich. Ct. App. Dec. 5, 2013); People v. Moore, No. 303750, 2013 WL 1500886, at *1 n.2 (Mich. Ct. App. Apr. 11, 2013)). We note, however, that the authorities cited by the Warden are unpublished and therefore non-binding on the Michigan courts. See Mich. Ct. R. 7.215(C)(1). More importantly, they do not represent the state of Michigan law at the time Ceasor's direct appeal was decided in 2007.

Prior to and at the time of Ceasor's 2007 appeal, the Michigan courts had consistently held—in both published and unpublished opinions—that a defendant was required to move for a Ginther hearing in order to avoid having his ineffective assistance of counsel claim adjudicated solely on the trial court record. See Fisher, 2007 WL 283799, at *2 n.2; Carter, 2003 WL 887594, at *4; Nantelle, 544 N.W.2d at 673. Thus, the 2013 cases cited by the Warden merely demonstrate that the Michigan courts have not been perfectly consistent in determining whether to entertain requests for Ginther hearings included only in the appellate brief in the years after Ceasor's appeal. However, the Warden does not cite, and we have not found, any cases decided prior to Ceasor's 2007 appeal contradicting Nantelle's holding that the review of a defendant's ineffective assistance of trial counsel claim shall be limited to mistakes apparent on the record unless he moves for a Ginther hearing.

Importantly, the Michigan Court of Appeals has also rejected requests for Ginther hearings on the separate and distinct ground that the attorney failed to file "a supporting 'affidavit or offer of proof regarding the facts to be established at a hearing'" under Rule 7.211(C)(1)(a). People v. Babby, No. 256308, 2005 WL 2679687, at *4 (Mich. Ct. App. Oct. 20, 2005) (quoting Mich. Ct. R. 7.211(C)(1)(a)(ii)); see, e.g., J. Nicks, 2015 WL 9392729, at *2; Singleton, 2009 WL 2170681, at *2; People v. Williams, 737 N.W.2d 797, 801 (Mich. Ct. App. 2007); People v. Hamby, Nos. 252735, 252850, 2005 WL 1398361, at *5 (Mich. Ct. App. June 14, 2005). In this case, appellate counsel did not file an affidavit or offer proof of facts that he planned to establish if the case were remanded for a Ginther hearing. Thus, unlike the motion for relief from judgment filed in the trial court several years later, Ceasor's appellate brief did not include affidavits from Ceasor or any potential experts, let alone explain whether there were experts willing to testify in support of Ceasor's version of the facts. The brief also failed to explain why Dr. Bandak, the one expert trial counsel had contacted, was not retained as an expert witness for trial. The Michigan Court of Appeals recognized this deficiency when it concluded that "the record [did] not support [Ceasor's] contention that his counsel failed to contact or try to procure an expert to support [his] theory" because "the trial court granted [Ceasor] a stipulated adjournment to consult an expert witness, and [Ceasor] then received additional adjournments because his counsel had located an expert on SBS willing to review the evidence." Ceasor, 2007 WL 2011747, at *4. Since the record made clear that Ceasor was granted additional time to retain an expert but lacked any indication as to why Ceasor ultimately failed to present expert testimony at trial, the court of appeals presumed that "defense counsel declined to present an expert witness because any expert consulted was unwilling to support [Ceasor's] position that [Brenden's] injur[ies] w[ere] accidental." Id.

In his brief before the Michigan Court of Appeals, appellate counsel asserted that "[t]he record alone shows that [trial] counsel was ineffective." (R. 7-12, PageID# 575) (emphasis added). Approving this characterization, the district court found that:

[A]ppellate counsel's decision to raise the ineffective assistance of trial counsel claim without separately requesting an evidentiary hearing was a "reasonable recognition that the allegations of ineffective assistance could be determined from the trial transcript alone. No additional evidence was really necessary for the [appellate] court to make a fair determination of the [S]ixth [A]mendment issue."

The trial record included testimony from: Brenden's mother, Genna; three law enforcement officers employed by the local sheriff's department; an emergency room nurse at Port Huron; the emergency room physician at Port Huron, Dr. Hunt; one of Brenden's attending physicians at Children's Hospital and the prosecution's expert witness, Dr. Gilmer-Hill; and the defendant, Ceasor. Both Ceasor's and Genna's testimony—including their statements that Ceasor had never expressed impatience with Brenden's high energy level or physically disciplined Genna's children, as well as Ceasor's account about playing "gotcha" with Brenden on the couch immediately before the baby was injured—supported the conclusion that Brenden was injured by an accidental fall from the couch while Ceasor was in the bathroom and Genna was out swimming with her daughter. Testifying for the prosecution, Dr. Gilmer-Hill opined that Brenden's injuries were inconsistent with such a history and instead showed that Ceasor had intentionally shaken or slammed Brenden during the hour-and-a-half that Genna was away. Dr. Gilmer-Hill also testified, on direct and cross-examination, that SBS was an accepted syndrome in her field and that Dr. Duhaime's study had demonstrated that a short fall had less force than shaking and could not cause a severe intracranial injury. Thus, although the trial record showed that there were competing theories as to how Brenden sustained his injuries (with only the prosecution's theory supported by expert testimony), the record did not show whether any expert would have been willing to testify about (1) SBS' controversial status in the medical community, (2) the alleged misinformation included in Dr. Gilmer-Hill's testimony (including her mischaracterization of Dr. Duhaime's study), or (3) the plausibility of Ceasor's version of the facts and the implausibility of the prosecution's theory of the case. The record also lacked any indication as to why Ceasor requested an adjournment to confer with an expert, but ultimately did not present any expert testimony at trial. See Ceasor, 2007 WL 2011747, at *4. The Michigan Court of Appeals recognized these significant omissions and affirmed Ceasor's conviction based on the record at trial. Id. at *3-4, *6.

The United States Supreme Court has recognized that "the services of a lawyer will for virtually every layman be necessary to present an appeal in a form suitable for appellate consideration on the merits." Evitts, 469 U.S. at 393 (emphasis added). In Ginther, the Michigan Supreme Court held that "[t]o the extent his claim depends on facts not of record, it is incumbent on him to make a testimonial record at the trial court level . . . which evidentially supports his claim and which excludes reasonable hypotheses consistent with the view that his trial lawyer represented him adequately." 212 N.W.2d at 925 (citation omitted) (emphasis added). At the time of Ceasor's 2007 appeal, the Michigan Court of Appeals had made clear that in order demonstrate the need for a Ginther hearing, the appellant was required to (1) move to remand for such a hearing under Rule 7.211(C)(1), see Fisher, 2007 WL 283799, at *2 n.2; Carter, 2003 WL 887594, at *4; Nantelle, 544 N.W.2d at 673, and (2) pursuant to the Rule's requirements, support the motion with an "affidavit or offer of proof regarding the facts to be established at a hearing," Mich. Ct. R. 7.211(C)(1)(a); see Williams, 737 N.W.2d at 801; Babby, 2005 WL 2679687, at *4; Hamby, 2005 WL 1398361, at *5. Appellate counsel neither filed a separate motion to remand under Rule 7.211(C) nor supported his perfunctory request for a remand, included only in his appellate brief, with an affidavit or proof of the facts that he planned to establish at the Ginther hearing. Without the latter, the Michigan Court of Appeals accurately characterized appellate counsel's assertion that an expert could have successfully challenged Dr. Gilmer-Hill's testimony as "entirely speculative." Ceasor, 2007 WL 2011747, at *4. Finally, appellate counsel's argument that the trial record alone demonstrated that trial counsel was ineffective for failing to retain an expert witness was patently unreasonable. As Ceasor points out, there was nothing in the trial record showing "Ceasor's inability to pay for an expert, why trial counsel did not present an expert, [or] what an expert might have said" at trial. Appellee's Br. at 41. Indeed, the trial record merely demonstrated that there was some controversy surrounding SBS diagnoses; there was no indication that this controversy actually affected the outcome of Ceasor's case. See Ceasor, 2007 WL 2011747, at *3-4. Because Ceasor's claim of ineffective assistance of trial counsel depended on facts not in the trial record, we find that it was objectively unreasonable for counsel to argue that this claim was demonstrated by the trial record alone. See Ginther, 212 N.W.2d at 925; Williams, 737 N.W.2d at 801; Nantelle, 544 N.W.2d at 673. Accordingly, we reject the district court's conclusion to the contrary as clearly erroneous.

Overall, we find that appellate counsel performed deficiently by (1) deciding not to file a separate motion to remand under Rule 7.211(C)(1), (2) failing to present an affidavit or offer of proof under subsection (a) of the same Rule, and (3) representing that the record alone showed that trial counsel was ineffective. For the aforementioned reasons, this conduct "fell below an objective standard of reasonableness." See Strickland, 466 U.S. at 688.

b. Prejudice

Having established deficient performance, Ceasor must also show prejudice. Ceasor "need not show that counsel's deficient conduct more likely than not altered the outcome," but instead must demonstrate "a probability sufficient to undermine confidence in the outcome." Id. at 693-94. As indicated above, appellate counsel's "failure to raise an issue on appeal could only be ineffective assistance if there is a reasonable probability that inclusion of the issue would have changed the result of the appeal." McFarland, 356 F.3d at 699. Thus, the adjudication of Ceasor's ineffective assistance of appellate counsel claim requires an assessment of the strength of his ineffective assistance of trial counsel claim. See Wilson, 515 F.3d at 707.

2. Ineffective Assistance of Trial Counsel

a. Deficient Performance

As noted by the district court, trial counsel could and did cross-examine Dr. Gilmer-Hill on several issues, including: Dr. Plunkett's study indicating that children have died as a result of falls from a height as short as two or three feet; questions of whether Brenden's injuries could have been caused by an earlier trauma or a recent vaccination; and two articles published in the American Journal of Forensic Medicine and Pathology discussing the effect of short falls on children. Ceasor, 2015 WL 164008, at *3, *6. Relying on trial counsel's "extensive[]" cross-examination of Dr. Gilmer-Hill, the district court concluded that it was a reasonable trial strategy to rely on this cross-examination in lieu of an expert witness and held that the reasonableness of this strategy "defeat[ed] [Ceasor's] ineffective assistance of trial counsel claim." Id. at *6. We disagree with this assessment.

It is well-established that a trial attorney's arguments are not evidence. See Darden v. Wainwright, 477 U.S. 168, 182 (1986); People v. Ullah, 550 N.W.2d 568, 575 (Mich. 1996). Similarly, an attorney's questions or examinations of witnesses are not evidence. See United States v. Ross, 703 F.3d 856, 885 (6th Cir. 2012); United States v. Campbell, 317 F.3d 597, 607 (6th Cir. 2003); People v. Guyton, No. 317970, 2014 WL 6783764, at *1 n.1 (Mich. Ct. App. Dec. 2, 2014); People v. Bunn, No. 182595, 1996 WL 33324020, at *1 (Mich. Ct. App. July 19, 1996). At trial, Dr. Gilmer-Hill testified that Brenden's subdural hematoma and retinal hemorrhaging were consistent with, and diagnostic of, child abuse. Dr. Gilmer-Hill also characterized the defense's alternative causation theories as unsupported by the medical literature or not widely accepted. In other words, the only expert testimony presented in a child abuse case with no direct evidence of abuse supported the prosecution's version of the facts and contradicted the defendant's. Further, although trial counsel attempted to undermine Dr. Gilmer-Hill's credibility by highlighting some of the weaknesses affecting her opinion, he lacked the ability to proffer evidence contradicting her opinions, including evidence that Ceasor's version of the facts was consistent with Brenden's injuries. He also lacked the ability to refute Dr. Gilmer-Hill's allegedly erroneous assertions about causation, the biomechanics of short falls, and the etiology of Brenden's subdural hematoma and retinal hemorrhaging. Nonetheless, as indicated by the affidavits attached to Ceasor's habeas petition, four experts have asserted that Dr. Gilmer-Hill's opinions were contradicted by both the medical literature and the facts of this case, including the hospital records from Port Huron and Children's. Thus, Ceasor has put forward a strong argument that "expert testimony was not only integral to the prosecution's ability to supply a narrative of [Ceasor's] guilt, it was likewise integral to [Ceasor's] ability to counter that narrative and supply his own." See People v. Ackley, 870 N.W.2d 858, 867 (Mich. 2015).

In 2015, the Michigan Supreme Court decided a case involving several similar issues: People v. Ackley. The prosecution's case against Ackley arose from the "unexplained and unwitnessed death of a child": the three-year-old daughter of Ackley's live-in girlfriend. Id. at 860. Ackley "denied hurting the child, and said that she must have died as the result of an accidental fall" from the bed while she was napping alone in her room. Id. At trial, the prosecution argued that Ackley "killed the child, either by blunt force trauma or [by] shaking [her]." Id. At a Ginther hearing, trial counsel testified that he contacted only one expert in preparation for Ackley's case: Dr. Brian Hunter, a forensic pathologist. Dr. Hunter, however, informed counsel that he was "not the best person" to testify for the defense because there was a deep divide between medical experts "about diagnosing injuries that result from falling short distances, on the one hand, and [SBS] . . . , on the other hand," and Dr. Hunter "was on the wrong side of this debate to be able to assist [Ackley]." Id. at 860-61. Dr. Hunter nonetheless referred trial counsel to a well-known forensic pathologist, Dr. Mark Shuman, who "had conducted substantial research on short falls" and was a "man of science" who could assess and potentially support Ackley's version of the facts. Id. at 861. Despite this referral, trial counsel never contacted Dr. Shuman, reached out to any other expert on short falls, or familiarized himself with the medical literature about the diagnoses at issue in Ackley's case. At the Ginther following his trial and conviction for first-degree felony murder and first-degree child abuse, Ackley proffered the affidavit of Dr. Werner Spitz, an expert in forensic pathology, who opined that the "bruises on the child's body were consistent with the intubation and CPR she received on the day of her death" and "the child's head injuries could not be attributed to [SBS] but were caused by a likely accidental 'mild impact.'" Id. at 861-62.

Analyzing Strickland's first prong, the Ackley court concluded that "counsel performed deficiently by failing to investigate and attempt to secure an expert witness who could both testify in support of the defendant's theory that the child's injuries were caused by an accidental fall and prepare counsel to counter the prosecution's expert medical testimony." Id. at 863. The court reasoned that trial counsel knew that: "the prosecution's theory of the case was that [Ackley] intentionally caused the child's unwitnessed injuries"; the prosecution "intended to prove [this premise] with expert testimony"; and "[t]his testimony would require a response." Id. Despite this knowledge, trial counsel consulted only with Dr. Hunter, an expert who identified himself as being on the wrong side of the SBS debate to assist Ackley. Counsel also failed to contact Dr. Shuman, despite Dr. Hunter's referral, a choice he "did not have sufficient information to legitimate" because he failed to investigate the law and facts of his case, including any medical treatises or articles that could explain the child's death. Id. The court noted that counsel's failure to adequately investigate his options for obtaining favorable expert testimony was particularly unreasonable "in light of the prominent controversy within the medical community regarding the reliability of [SBS] diagnoses." Id. at 864.

Ackley is instructive. Here, as in Ackley, because Brenden's injuries were unwitnessed (and Ceasor had no known history of child abuse), the prosecution needed an expert to support its theory that the child's injuries were caused by intentional shaking or slamming and could not have resulted from an accidental fall. See id. at 860, 863, 866-67. Because the prosecution's causation theory (and rebuttal of Ceasor's explanation) was based almost exclusively on the expert testimony of Dr. Gilmer-Hill, Ceasor has proffered a strong argument that trial counsel should have known that Dr. Gilmer-Hill's testimony "would require a response." See id. at 863. Further, we have every indication that trial counsel did in fact understand that Dr. Gilmer-Hill's testimony would need to be met with countervailing expert testimony. At the preliminary examination of Dr. Gilmer-Hill, which took place almost a year before Ceasor's December 2005 trial, defense counsel represented to the trial judge that Ceasor's trial was "going to be expert against expert." More damningly, two of Ceasor's affidavits assert that trial counsel acknowledged the need for expert testimony at attorney-client meetings, telling Ceasor that he would need an expert witness "in order to succeed at trial due to the complexity of the medical issues" in his case. (R. 24-2, PageID# 1368, 1371). Nonetheless, trial counsel failed to retain an expert witness for trial. According to Ceasor, this decision was based on Ceasor's inability to pay Dr. Bandak's fees of more than $10,000. Ceasor also asserts that after he informed trial counsel that he could not afford Dr. Bandak's fees, counsel "refused to entertain other options for expert testimony" and "never retained an expert for . . . trial." (Id. at 1369, 1372).

Under Michigan law, a court may "provide public funds for indigent defendants to retain expert witnesses" under Mich. Comp. Laws § 775.15. People v. Agar, No. 321243, 2016 WL 399933, at *2 (Mich. Ct. App. Feb. 2, 2016); see also People v. Tanner, 671 N.W.2d 728, 729-30 (Mich. 2003). Although "a trial court is not compelled to provide funds for the appointment of an expert on demand," Tanner, 671 N.W.2d at 730 (citing People v. Jacobsen, 532 N.W.2d 838, 839 (1995)), in this case, trial counsel never moved for funds to retain an expert in the first place.

We are mindful of the Supreme Court's admonition that "[t]he selection of an expert witness is a paradigmatic example of the type of strategic choice that, when made after thorough investigation of the law and facts, is virtually unchallengeable." Hinton v. Alabama, 134 S. Ct. 1081, 1089 (2014) (quotation marks and brackets omitted) (emphasis added). However, "strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation. In other words, counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary." Strickland, 466 U.S. at 690-91. In this case, trial counsel's alleged error was failing to investigate Michigan law governing public funds for indigent defendants that could have been used to either (a) pay Dr. Bandak's $11,500 fee or (b) compensate a different expert at a similar or lower rate. Recently, the Supreme Court found that an analogous failure to investigate a "state statute providing for defense funding for indigent defendants" constituted deficient performance where there was evidence that the "attorney knew that he needed more funding to present an effective defense." Hinton, 134 S. Ct. at 1088-89. In Hinton, trial counsel put on an expert witness that he himself deemed "inadequate" because he mistakenly believed he could not obtain more than $1000 under an Alabama statute providing for state reimbursement of expenses incurred by an indigent criminal defendant. Id. The Supreme Court found that counsel's choice to hire a sub-par expert based on his misapprehension of the funding statute not only demonstrated that counsel was ignorant of the law, but also that he failed to "perform basic research" regarding the availability of public funds. Id. at 1089. This conduct, the Court held, constituted deficient performance. Id.

Here, the affidavits sworn to by Ceasor and his uncle indicate that trial counsel's decision to try Ceasor's case without an expert was made without investigating Michigan law. This is because although trial counsel allegedly told Ceasor that the medical complexity of his case necessitated hiring a defense expert, there is no record that counsel ever moved for funding under Mich. Comp. Laws § 775.15 or pursued any other avenues for retaining an expert. Ceasor argues that because trial counsel failed to investigate available options for paying an expert despite Ceasor's indigency, the jury's determination of his guilt or innocence hung on the testimony of the prosecution's lone expert witness. Since Ceasor has never been granted an evidentiary hearing on this issue, in state or federal court, we do not presume that this assertion will be borne out by the evidence. However, we are persuaded that Ceasor has demonstrated the strength of the first prong of his ineffective assistance of trial counsel claim—deficient performance. See Wilson, 515 F.3d at 707.

b. Prejudice

With regard to the second prong, prejudice, Ceasor highlights several considerations suggesting that there is "a reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different." See Strickland, 466 U.S. at 694. First, "[t]here was no explanation for [Brenden's] injuries beyond the theor[y] presented by the expert[], and the prosecution produced no witnesses who testified that [Ceasor] was ever abusive." See Ackley, 870 N.W.2d at 865. Instead, witnesses who observed Ceasor at Port Huron testified that he seemed "shaken up" and upset by Brenden's injuries. Further, both Genna and Ceasor testified that Ceasor regularly contributed to Brenden's care, had been left alone with both of Genna's children "several times" in the months preceding Brenden's injuries, and never used violence to discipline Genna's children or his own child. Ceasor also testified that there was nothing Brenden did on the day that he sustained his injuries that made Ceasor angry or tried his patience. Second, Ceasor's conviction "turned on the jury's assessment of the prosecution's [causation] theory," which itself depended on the expert testimony of Dr. Gilmer-Hill. See id. Nonetheless, while the affidavits sworn to by Drs. Plunkett, Stephens, Uscinksi, and Van Ee signal that an expert could have challenged Dr. Gilmer-Hill's testimony based on the specific facts averred to in Ceasor's 2005 trial, the affidavits from Drs. Plunkett and Van Ee indicate that an expert would have testified in support of Ceasor's version of the facts if trial counsel had sought their testimony.

The crux of the prosecution's proof that Ceasor knowingly or intentionally caused Brenden serious physical harm—an element of first-degree child abuse that the prosecution was required to prove beyond a reasonable doubt, see Mich. Comp. Laws § 750.136b(2); People v. Nowack, 614 N.W.2d 78, 82 (Mich. 2000)—was Dr. Gilmer-Hill's expert testimony. At closing argument, the prosecution went out of its way to point out that this testimony was uncontroverted. Brenden's injuries—a subdural hematoma and retinal hemorrhaging—were medically complex and beyond the easy comprehension of the jury. Further, no amount of cross-examination or lay witness testimony could have rebutted Dr. Gilmer-Hill's medical opinions that these injuries were medically consistent with abuse and inconsistent with an accidental fall. Thus, we acknowledge, as the Ackley court did, that in many SBS cases "where there is 'no victim who can provide an account, no eyewitness, no corroborative physical evidence and no apparent motive to [harm],' the expert 'is the case.'" 870 N.W.2d at 867 (quoting Deborah Tuerkheimer, The Next Innocence Project: Shaken Baby Syndrome and the Criminal Courts, 87 Wash U. L. Rev. 1, 27 (2009)) (emphasis in original).

D. The District Court's Errors

Finally, we are unpersuaded by the district court's reasons for concluding that even "assuming that appellate counsel was ineffective for failing to move for a Ginther hearing," Ceasor cannot show prejudice because his ineffective assistance of trial counsel claim lacks merit. Ceasor, 2015 WL 164008, at *6. As indicated above, to be afforded habeas relief, Ceasor need only show that "there is a reasonable probability that inclusion of the issue would have changed the result of the appeal." McFarland, 356 F.3d at 699.

First, we reject as irrelevant the district court's reliance on the fact that Ceasor did not submit an affidavit showing that Dr. Bandak would have testified favorably at trial. Ceasor has provided affidavits from four other experts—two of whom explicitly represented that they would have testified favorably at his 2005 trial—as well as affidavits swearing that Dr. Bandak forewent testifying at Ceasor's trial because Ceasor could not afford to pay his fee. There is no indication that Dr. Bandak declined to testify because he questioned Ceasor's version of the facts, and if the district court doubted the veracity of the statements included in Ceasor's affidavits, or sought additional factual development on this issue, the proper course would have been to hold an evidentiary hearing. See McAdoo v. Elo, 365 F.3d 487, 500 (6th Cir. 2004) ("[W]hen a defendant diligently seeks an evidentiary hearing in the state courts in the manner prescribed, but the state courts deny him that opportunity, he can avoid [28 U.S.C.] § 2254(e)(2)'s barriers to obtaining a hearing in federal court."); Stanford v. Parker, 266 F.3d 442, 459 (6th Cir. 2001) ("Generally, a habeas petitioner is entitled to an evidentiary hearing in federal court if the petition 'alleges sufficient grounds for release, relevant facts are in dispute, and the state courts did not hold a full and fair evidentiary hearing.'").

Additionally, the district court's finding that it was reasonable for trial counsel to "extensively" cross-examine Dr. Gilmer-Hill in lieu of calling an expert witness is belied by trial counsel's alleged statements, in his meetings with Ceasor, that Ceasor would need an expert witness "in order to succeed at trial due to the complexity of the medical issues" in his case. Further, the trial record gives us reason to doubt the efficacy of trial counsel's cross-examination, which failed to highlight errors in Dr. Gilmer-Hill's opinions that a well-prepared attorney would have recognized. More importantly, even if trial counsel could effectively attack Dr. Gilmer-Hill's credibility through cross-examination, Ceasor has put forth a cogent argument that because trial counsel knew that the prosecution would rely on Dr. Gilmer-Hill's testimony to show that the medical evidence contradicted and effectively disproved Ceasor's version of the facts, it was objectively unreasonable to fail to take steps to retain an expert. See Hinton, 134 S. Ct. at 1088; Ackley, 870 N.W.2d at 865.

Finally, we reject as clearly erroneous the district court's conclusion that there was sufficient evidence against Ceasor, apart from Dr. Gilmer-Hill's testimony, to preclude a finding of prejudice, i.e., a reasonable probability that the outcome of Ceasor's trial would have been different if trial counsel had presented an expert to challenge Dr. Gilmer-Hill's testimony. See Byrd, 689 F.3d at 639-40. The district court relied on two types of evidence in support of this conclusion. First, the district court observed that Ceasor and Genna told "inconsistent stories to the police, which called into question their credibility." Ceasor, 2015 WL 164008, at *7. Although we agree with the general proposition that inconsistencies affecting a criminal defendant's version of events may undermine his or her credibility, see, e.g., United States v. Edmond, 815 F.3d 1032, 1047 (6th Cir. 2016), the only testimonial inconsistency in this case is that Ceasor and Genna changed their story about whether Genna was present when Brenden was injured. However, both Ceasor and Genna admitted to police that Genna was away from the house when Brenden was injured, the truth came out within days of Brenden's injuries and more than a year before Ceasor was tried, and the trial testimony showed that Ceasor never told Genna to lie to the police. The trial record also demonstrates that Ceasor consistently told the police and medical personnel that Brenden fell off the couch shortly after they played a game of "gotcha," and that Ceasor reiterated this account of how Brenden was injured even after he was confronted with the inconsistency regarding Genna's presence at the time of Brenden's injuries. And, perhaps most importantly, this inconsistency was plainly immaterial to Brenden's treatment because it did not alter the alleged cause of his injuries and no witness testified that Brenden's treatment would have changed had Ceasor told a different story about whether Genna was present when the child was hurt.

The district court also relied on the purported lack of external trauma noted by Drs. Hunt and Gilmer-Hill to support its conclusion that Ceasor could not show prejudice. However, Dr. Hunt, who saw Brenden at Port Huron within hours of when he was injured, acknowledged that bruising does not always occur "right away." Dr. Gilmer-Hill, on the other hand, conceded that bruising and oral redness were documented in the nurse's notes at Children's, but said she did not rely on this part of Brenden's history because the nurse's notes were the "only place" that noted bruising and she "didn't see the bruise [her]self." (R. 7-7, PageID# 323-24). In our view, this isolated and contested evidence regarding the lack of external trauma to Brenden's body, standing alone and weighed against the evidence that Ceasor often served as a second caretaker for Brenden, never physically disciplined Genna's children or his own, and repeatedly offered a consistent account to physicians and police about how he believed Brenden's (unwitnessed) injuries occurred, falls far short of being dispositive of the issue of Ceasor's guilt. See Ackley, 870 N.W.2d at 865-66.

Based on the strength of Ceasor's ineffective assistance of trial counsel claim, see Wilson, 515 F.3d at 707, we find that habeas relief is warranted because there is a reasonable probability that if appellate counsel had (1) properly moved to remand for a Ginther hearing under Rule 7.211(C)(1) and (2) submitted an affidavit or other offer of proof in support of this claim, such performance "would have changed the result of [Ceasor's direct] appeal." See McFarland, 356 F.3d at 699. Having found that appellate counsel performed deficiently by (1) arguing that the trial record alone supported Ceasor's claim and (2) failing to comply with the requirements of Michigan law in 2007, including Rule 7.211(C)(1)(a)'s "affidavit or offer of proof" requirement, see supra Part II.C.1.a, we remand this case to the district court for an evidentiary hearing on whether Ceasor was prejudiced by appellate counsel's deficient performance.

We note that because the state courts never considered Ceasor's ineffective assistance of appellate counsel claim on the merits, the evidentiary limitation articulated in Cullen v. Pinholster, 563 U.S. 170, 180-82 (2011), and derived from 28 U.S.C. § 2254(d), does not bar the district court from considering proof in support of Ceasor's ineffective assistance of appellate counsel claim at an evidentiary hearing, see Bies v. Sheldon, 775 F.3d 386, 394 n.5 (6th Cir. 2014); McClellan v. Rapelje, 703 F.3d 344, 351 (6th Cir. 2013). We also find that Ceasor diligently sought an evidentiary hearing before both the trial court and the district court, and therefore has complied with the requirements of § 2254(e)(2). See Williams v. Taylor, 529 U.S. 420, 432 (2000) ("Under the opening clause of § 2254(e)(2), a failure to develop the factual basis of a claim is not established unless there is lack of diligence, or some greater fault, attributable to the [petitioner] or the [petitioner's] counsel.").

In examining the issue of prejudice, the district court shall consider the affidavits submitted with Ceasor's habeas petition, the sworn statements of counsel, if available, and any other evidence the district court finds relevant to the question of prejudice. At this juncture, we offer no comment as to whether Ceasor will be able to demonstrate prejudice at the evidentiary hearing. Should the district court find prejudice, it may enter a conditional writ of habeas to allow the state courts to consider a new appeal or a renewed request for a Ginther hearing on Ceasor's ineffective assistance of trial counsel claim. See, e.g., Goff, 601 F.3d at 472-73, 482 (granting a conditional writ of habeas corpus unless the state courts reopened the petitioner's direct appeal within 120 days to permit petitioner to raise his allocution claim); Johnson v. Mitchell, 585 F.3d 923, 946 (6th Cir. 2009) (granting a conditional writ of habeas vacating the petitioner's death sentence unless the state courts conducted a new sentencing hearing within 180 days). If Ceasor cannot show prejudice, further habeas relief should be denied. See Strickland, 466 U.S. at 691-92.

IV. Conclusion

For the reasons stated in this opinion, we REVERSE the district court's judgment denying habeas relief and REMAND for an evidentiary hearing on the prejudice prong of Ceasor's ineffective assistance of appellate counsel claim.