Marvin Craig White, convicted in Louisiana of armed robbery, appeals the dismissal of his 28 U.S.C. Sec. 2254 habeas corpus petition. We affirm.

While exiting the parking lot of his Sunset, Louisiana business around 6:00 p.m. on July 30, 1982, Gayle White, president and general manager of Louisiana Wholesale Drug, was hailed by a man he later identified as Marvin Craig White. The man, in apparent distress, asked to use the telephone. When Gayle White replied that the office was locked the man drew a handgun. Two accomplices appeared and the trio forced Gayle to open the company warehouse from which they stole more than $4000 of controlled drugs and other items. For 30 minutes or more Marvin Craig White savagely beat and threatened to kill Gayle White. The robbers left him in a pool of his own blood with his arms and legs bound together.

Gayle White finally worked himself free and alerted the St. Landry Parish sheriff's office and Sunset, Louisiana city police. He told them of his ordeal and described his assailants as three black men, the first being about six feet tall, of medium build, wearing a dark blue, uniform-style shirt.

During the next several weeks Gayle White examined over 100 mug shots; none were identified as one of the assailants. On September 21, 1982, Sgt. Charles Duplechain and Assistant Chief Roy Mallet of the Opelousas, Louisiana police department used hypnosis in an effort to review Gayle White's memory of the traumatic robbery. The Opelousas police were not involved in the investigation and Duplechain and Mallet were not familiar with any of the details. During the course of the interview Gayle White gave a narrative of the robbery including a more detailed description of the first robber.1

Four months later the deputy sheriff in charge of the investigation received a photograph of Marvin Craig White who had been arrested for a similar crime in another jurisdiction. This photograph was placed with two score others and the array was shown to Gayle White who immediately identified Marvin Craig White as the first assailant. Over a year later, without doubt or hesitancy, Gayle White picked Marvin Craig White from a live lineup whose composition had been selected carefully by Marvin Craig White's attorney.

In May 1985 Marvin Craig White was tried and convicted of armed robbery and was sentenced to prison for 99 years. The evidence linking him to the crime was Gayle White's testimony identifying him as the first of the three robbers. After exhausting state remedies the instant petition was filed. The district court adopted the recommendation of the magistrate judge and denied relief. We granted a certificate of probable cause.

The primary contention presented to the state courts, the federal trial court, and now this court, is that Gayle White's hypnotically refreshed identification was admitted erroneously because it violated confrontation and due process rights. We reject the proposition that either due process or confrontation guarantees require a per se bar to the admission of post-hypnosis testimony.2 As we previously have noted, the two challenges really pose a single inquiry: was the memory of Gayle White likely distorted by hypnosis to the point that the admission of his testimony resulted in a fundamentally unfair trial of Marvin Craig White.3 In considering similar challenges, we have balanced the inherent dangers of hypnotically "refreshed" testimony4 against a compendium of indicia of reliability.5

This case-by-case balancing approach6 requires consideration of a number of relevant factors. In our leading opinion, Wicker v. McCotter, we upheld the constitutionality of hypnotically refreshed testimony, favorably noting several factors: the post-hypnosis testimony corresponded substantially with the pre-hypnosis statements; the identification of the defendant at trial was probed fully on cross-examination; and the record reflected independent evidence corroborating the identification. In United States v. Harrelson,7 a case involving admissibility under the Federal Rules of Evidence rather than a constitutional challenge, we commented favorably on the absence of unduly suggestive hypnosis procedures and the fact that neither witness had failed an opportunity to identify the defendant before being hypnotized.

In United States v. Valdez,8 another Federal Rules of Evidence case, we were stricter in our assessment of hypnosis, holding that hypnotically refreshed testimony could not be used to identify "a person known by the witness to be under suspicion, whom the witness had nevertheless been unable to identify before being hypnotized."9 In that case a Texas Ranger had repeatedly interviewed a person as the prime suspect in an investigation of an extortion attempt, part of which the Ranger purportedly witnessed. The Ranger was thereafter unable to identify that person in a lineup until after undergoing hypnosis. In addition, the hypnosis session in Valdez reflected an unreliability because the questioners were familiar with the case and used suggestive cues to steer the answers of the Ranger.

The Supreme Court's decision in Rock v. Arkansas10 is instructive in the selection of the balancing factors, encouraging procedural safeguards like those described by Dr. Martin T. Orne,11 the FBI,12 and, more recently, the Seventh Circuit13 to minimize the danger of suggestion or confabulation. Other traditional guarantors of testimonial reliability are considered important by the Supreme Court.14 Verification by corroborating evidence, cross-examination to reveal inconsistencies, expert testimony on the dangers of hypnosis, and pre-hypnosis statements for purposes of comparison engender greater confidence in a jury's ability to weigh accurately the reliability of post-hypnotic testimony.15

These and other authorities16 reflect an emerging litany of factors courts consider in determining whether the admission of post-hypnotic testimony violates a defendant's rights to due process and confrontation. We are persuaded that on a case-by-case basis the court should determine whether the defendant has shown,17 from the totality of the circumstances,18 that the post-hypnosis testimony is unreliable.19 We agree with our Seventh Circuit colleagues that it is preferable, although not constitutionally mandated, that the hypnosis be performed by a professional trained in hypnosis; that the hypnotist be independent of either party; that the hypnotist's pre-session knowledge of the case be kept to a minimum and recorded as a check on suggestion; that the hypnotist obtain a description of the case from the subject before inducing hypnosis as a check on confabulation; that the session be recorded, preferably videotaped; and that only participants in the hypnotic session be present.20

We find in the instant case that the use of Gayle White's post-hypnosis testimony was constitutionally permissible. Although neither Duplechain nor Mallet was a professional psychiatrist or psychologist, each had taken both a basic and an advanced course in hypnosis from qualified experts. Neither knew any of the details of the case, limiting the possibility of suggestion. Neither had any law enforcement responsibility in the investigation of the crime in question. The session was audiotaped and a recording was made available and played for judge, jury, and defense counsel, ensuring an opportunity for fair challenge to the techniques and procedures used. Gayle White gave a narrative description of the crime and its perpetrators without any suggestions or pressure from Duplechain or Mallet. Gayle White and the two hypnotists were cross-examined vigorously on the circumstances and procedures surrounding the session.21 Although the deputy in charge of the investigation was present during the session, there is no indication that he spoke or in any way became involved.22 The deputy and the other law officers were not aware of the existence of Marvin Craig White until months after the hypnosis session, negating any suggestion that Gayle White's accurate description could have been created, in whole or in part, by the sort of impermissible particularized suggestion found in Valdez or Williams.23 Gayle White's post-hypnosis testimony, though more detailed than his brief description at the crime scene, was wholly consistent therewith. Additionally, Gayle White had a very close view of his assailant, under good lighting, for nearly 30 minutes.24

We are mindful that the testimony of Gayle White is the sole evidence linking Marvin Craig White to the crime,25 but we are not persuaded from the totality of the circumstances presented that his testimonial evidence was unreliable and that its admission was error. We note that the Louisiana Court of Appeal found "that no suggestion was made during the hypnosis session"; that Gayle White "identified the defendant's picture without hesitation" and "expressed no doubts when he picked the defendant out of a lineup"; and gave a description at the crime scene consistent with his post-hypnotic description.26 6] That court concluded: "It does not appear that the session affected White's ability to identify his assailant."27 Those findings of fact by the state court are presumed correct.28 Considering this statutorily-required deference, and the foregoing factual balancing, we conclude that hypnosis did not alter Gayle White's memory so as to taint Marvin Craig White's trial by depriving him of an opportunity to confront his accuser or by denying his due process rights to a fair trial.

The judgment of the district court is AFFIRMED.