Opinion by JUDGE STERNBERG.

The defendant, Barney Quintanar, was charged with manslaughter and criminally negligent homicide. Following a jury trial he was convicted of the latter offense and acquitted of the former. On appeal he asserts that reversal is required because two witnesses who testified against him had been rendered incompetent to testify by pretrial hypnotic sessions seeking to enhance the witnesses' recollections. He also argues that the evidence was insufficient to support his conviction. We find no merit in his contentions and therefore affirm the conviction.

At trial, the prosecution's evidence indicated the following events had occurred. At 3:00 a.m. one morning, the victim and Quintanar became involved in an argument at the apartment of defendant's estranged wife. Shortly thereafter, the victim's brother and two friends arrived, and the victim's brother temporarily broke up the altercation. Thereafter, as the two friends, the victim and his brother left the apartment, defendant leaped at the victim and struck him hard in the face, causing him to fall to the ground.

According to a pathologist, the victim died of a blow to the head. And, defendant later gave a statement to the police in which he conceded knowledge that the victim had told him of his involvement in an automobile accident as a result of which "he would die . . . . if he ever got drunk and fell down and hit his forehead."

The prosecution's two main witnesses had given written statements of their recollections from that evening to the police. Both witnesses recalled the defendant striking the victim some place on the head, though not the exact location. Later in the investigation both witnesses were hypnotized by a police officer for the purpose of aiding their recall of the details. The defendant sought to exclude both witnesses from testifying on the grounds that the hypnosis made them incompetent to testify. Following argument at an in-camera hearing, the trial court allowed them to testify, but limited their testimony to their pre-hypnotic recollection as recorded in their written statements.

I.

The question of whether hypnotically refreshed testimony is admissible in a criminal trial is an open one in Colorado. See People v. District Court, 652 P.2d 582 (Colo. 1982); People v. Angelini, 649 P.2d 341 (Colo.App. 1982). This court has held that an expert's opinion on a subject's mental state based on his observations of the subject during hypnosis is not admissible because hypnosis has not been accepted as a reliable basis for such an opinion. See People v. Diaz, 644 P.2d 71 (Colo.App. 1982), ( cert. denied March 29, 1982). But Diaz does not settle the question of the use of a witness' own hypnotically refreshed testimony.

Three basic approaches have been taken regarding the use of testimony from witnesses whose memories have been refreshed by hypnosis. In some jurisdictions, prior hypnosis affects only the weight of the testimony, not its admissibility. See, e.g., Chapman v. State, 638 P.2d 1280 (Wyo. 1982); People v. Smrekar, 68 Ill. App.3d 379, 385 N.E.2d 848 (1979). In others a per se exclusionary rule is followed. See, e.g., People v. Shirley, 31 Cal.3d 18, 641 P.2d 775 (1982); Collins v. State, 447 A.2d 1272 (Md.App. 1982). And, some admit it subject to certain procedural requirements which seek to insure reliability. See, e.g., State v. Hurd, 86 N.J. 525, 432 A.2d 86 (1981).

The People urge us to adopt a rule of absolute admissibility or, in the alternative, to condition admission on a showing of reliability in each instance. The defendant argues for a per se rule that the witness is incompetent to testify. We conclude that none of these is the appropriate approach. Instead, we hold that testimony of a witness who has been questioned under hypnosis is per se inadmissible as to his recollections from the time of the hypnotic session forward. However, we hold that the witness is not incompetent to testify to pre-hypnosis recollections that have previously been unequivocally disclosed and recorded by tape recording, video tape, or, as here, by written statement.

The reliability of evidence obtained from use of a scientific principle is generally tested under the rule in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923). There a systolic blood pressure deception test was held to be inadmissible because it had not gained sufficient standing and scientific recognition to justify its use. By application of the Frye test, a per se exclusionary rule was adopted concerning evidence of polygraph results in People v. Anderson, 637 P.2d 354 (Colo. 1981).

Although hypnosis is considered reasonably reliable in the fields of psychiatry and psychology, where the interest is in merely stimulating memory, it does not hold this stature in forensics, where the interest is in stimulating accurate memory. Levitt, The Use of Hypnosis to "Freshen" the Memory of Witnesses or Victims, Trial 56 (April 1981); Diamond, Inherent Problems in the Use of Pretrial Hypnosis on a Prospective Witness, 68 Calif. L. Rev. 313 (1980). These recent articles by Levitt, a professor of psychology, and Diamond, a professor of clinical psychiatry, demonstrate that hypnotically induced recollection has not attained the level of reliability among authorities in the field as required by Frye to outweigh the possibilities for prejudice.

Levitt characterizes two salient conditions of a person in a hypnotic trance as hypersuggestibility and hypercompliance. These two characteristics produce a person who is more easily influenced by external sources and more highly motivated to please others than the average person. Also present are the characteristics of reduced use of critical faculties and increased impulsiveness. When these combine with the subject's knowledge that the purpose of the session is to supply missing facts or details, the result is recall that is likely to be a combination of fact and fantasy, the reliability of which the subject is absolutely convinced. Any previous uncertainty in the subject's memory evaporates so that the witness cannot be shaken by cross-examination.

In his article, Diamond points out that even a skilled hypnotist is unable to avoid implanting suggestions in the subject's mind, and the desire to please causes the subject to fill in missing details with fantasy or confabulation (a filling in of gaps in memory by free fabrication, Webster's Third International Dictionary 475 (1976)). Because the subject cannot distinguish between memories "learned" under hypnosis and pre-existing memories, the testimony as a whole appears to be genuine and spontaneous. Nor can the hypnotist or any other expert verify the accuracy of the enhanced memory. Diamond, supra. These observations demonstrate the inadequacy of the adversary system to probe the reliability of the witness' memory and the credibility of his testimony.

In response to the claim that all eyewitness testimony is subject to similar criticism, Diamond points out that witnesses who have not been hypnotized may communicate their awareness of the deficiency in their recall by body language which indicates their lack of confidence, and may be made to concede that their memory is fallible. In contrast, after hypnosis, witnesses become so confident of their memory that they show no hesitancy in relating the "facts." Thus, not only is the recall likely to be inaccurate, but its accuracy cannot be determined either by experts or the witness. Consequently, the trier of fact can do no more than speculate as to the accuracy and reliability of hypnotically-refreshed memory.

We conclude that a jury's ability to observe a witness' demeanor and analyze a witness' ability to perceive, remember, and articulate is so hampered by the hypnotic process that the probative value of such evidence cannot be tested within the framework of the adversary system as is any other device used to refresh recollection. We hold therefore that hypnotically refreshed testimony is inadmissible.

In our view, these problems are not overcome by the procedural safeguards adopted by the New Jersey Supreme Court in State v. Hurd, supra. Of primary concern to that court was the fallibility of all human memory from the time of perception, through the process of interrogation, and continuing to the point of articulation.

In Hurd, supra, a physician and professor of psychiatry testified that the use of hypnosis with certain types of memory loss is less likely to yield confabulation and fantasy than with others, i.e., if the loss resulted from traumatic neurosis rather than a bad memory or inability to observe sufficiently in the first place. He suggested that a reliable procedure in the appropriate situation could produce reasonably accurate results. The Hurd court imposed procedural requirements aimed at providing an adequate record to test the reliability of the procedure and to ensure a minimum level of reliability. In brief, these requirements were that an experienced psychiatrist or psychologist conduct the session; that the hypnotist be independent of either party; that all information given to the hypnotist before the session be recorded in some fashion; that a detailed description of the subject's recollections be obtained prior to hypnosis; that all contact between the hypnotist and the subject before, during, and after the session be recorded; and that only the hypnotist and subject be present during any phase of the session.

We conclude that whatever value these safeguards hold for increasing the reliability of the subject's memory, they do not overcome the problem created by the inability of the trier of fact to assess the witness' credibility. Additionally, this approach requires the court to determine in each case whether the kind of memory loss and the procedure employed would make hypnosis a reasonably reliable means of restoring the witness' memory. See People v. Gonzales, 108 Mich. App. 145, 310 N.W.2d 306 (1981); State v. Palmer, 210 Neb. 206, 313 N.W.2d 648 (1981); Collins v. Superior Court, 132 Ariz. 180, 644 P.2d 1266 (1982); People v. Shirley, 31 Cal.3d 18, 181 Cal.Rptr. 243, 641 P.2d 775 (1982); Commonwealth v. Nazarovitch, 496 Pa. 97, 436 A.2d 170 (1981); State v. Mack, 292 N.W.2d 764 (Minn. 1980); Collins v. State, supra.

The courts in California and Maryland follow this reasoning to its logical extreme and hold that the witness is incompetent to testify on any matter related to the subject discussed under hypnosis. In effect, they agree with Dr. Diamond's view that the witness is "contaminated." And, they conclude that even when the witness is testifying to pre-hypnotic memories, because the subject cannot separate the source of recall, the testimony would not, in fact, be from memory. A witness who cannot testify from memory, they reason, is incompetent.

In Collins v. Superior Court, supra, the Arizona Supreme Court modified its position that the witness was incompetent taken in State v. Mena, 128 Ariz. 226, 624 P.2d 1274 (1981). The court balanced the benefit and risk entailed in the use of hypnosis as an investigatory tool. It concluded that a blanket rule of incompetency was not worth the loss of this tool by the police who would seldom use it if a key witness would be rendered incompetent to testify. Instead, it held that a witness who is hypnotized during the investigatory phase of the case will be permitted to testify with regard to those matters which he was able to recall and relate prior to hypnosis, provided the recollection was recorded in written, tape recorded, or videotaped form.

Similarly, in Commonwealth v. Taylor, 294 Pa. Super. 171, 439 A.2d 805 (1982), Pennsylvania refined its rule that a previously hypnotized witness could not testify as to matters remembered after hypnosis when it allowed testimony by a witness who had identified her assailant prior to hypnosis. Because hypnosis did not add anything substantial, the dangers of fantasy and confabulation were not present. Indiana followed similar reasoning in Strong v. State, 435 Ind. 969, 435 N.E.2d 969 (1982), ruling that the product of hypnosis is not admissible but the witness can testify if the state demonstrates the testimony had independent origin. Minnesota has adopted a rule allowing a witness to testify as to matters previously and unequivocally disclosed to authorities prior to hypnosis, and Michigan has also ruled that a witness may testify about aspects of the case which were remembered prior to hypnosis. See State v. Koehler, 312 N.W.2d 108 (Minn. 1981); People v. Wallach, 110 Mich. App. 37, 312 N.W.2d 387 (1981).

We are persuaded by the reasoning of those courts which allow testimony of pre-hypnotic recollections which have been previously and unequivocally disclosed and recorded. While recognizing the problems in cross-examining the witness regarding this testimony, we, like the Arizona court, conclude the value of hypnosis as an investigative tool outweighs these problems when there is an accurate record of pre-hypnotic recollections.

Turning now to the facts of this case, we note the prosecution's two key witnesses were both hypnotized after they had given written statements to the police. The sessions were conducted by a police officer who received his training in a series of four lectures. The session itself was videotaped, but the pre-induction session was not. Both witnesses had recollections of the events prior to hypnosis which were damaging to defendant. These did not change substantially following hypnosis. The victim's brother remembered a second punch under hypnosis and more details such as facial expressions and clothing. The other witnesses did not remember anything differently following hypnosis. Both witnesses were allowed to testify at trial, and the brother was strictly limited to the facts recorded in his pre-hypnosis statement.

Under these facts, allowing the witnesses to testify was not error. Both witnesses had memories of the event which they were able to recall and relate prior to hypnosis, and these recollections were adequately recorded prior to hypnosis. The hypnosis did not add significantly to these recollections so, as in Taylor, supra, the dangers of fantasy and confabulation did not "contaminate" these recollections. Finally, neither was permitted to testify as to any details brought out by hypnosis so the defendant was not prejudiced by the jury receiving inadmissible testimony.

II.

Contrary to Quintanar's next contention, the People's evidence was sufficient to support his conviction.

The evidence had established that defendant struck the victim with his fist during the course of a heated argument. It showed that the victim perhaps had aroused defendant's jealousy but had not sought a fight. It also showed that defendant was substantially larger than the victim and that he knew about the victim's prior injury.

Considering the standards to be applied when ruling on a judgment of acquittal, we conclude the trial court properly denied defendant's motion. When viewed in the light most favorable to the prosecution, the evidence was substantial enough to support a conclusion by a reasonable mind that defendant was guilty beyond a reasonable doubt. People v. Andrews, 632 P.2d 1012 (Colo. 1981). See also People v. Bennett, 183 Colo. 125, 515 P.2d 466 (1973).

Defendant argues, nevertheless, that the court erred in admitting the portion of his statement in which he admitted the victim had told him of a prior head injury and that a blow to the head could prove fatal. There was no error.

Evidence is relevant if it tends to make the existence of any fact that is of consequence to the determination of the action more or less probable than it would be without the evidence. Colorado Rules of Evidence 401. It is within the special province and competence of the trial court to determine the relevance of evidence at trial. People v. Bynum, 192 Colo. 60, 556 P.2d 469 (1976). Defendant's statement is relevant to the issue of negligence.

Additionally, it is significant that defendant was also charged with, although acquitted of, reckless manslaughter, which charge placed his subjective state of mind directly in issue. That expert testimony did not establish a causative relationship between the prior injury and the victim's death does not affect the relevance of defendant's knowledge of the injury. Therefore, admission of the statement was not error.

The judgment is affirmed.

CHIEF JUDGE ENOCH and JUDGE VAN CISE concur.