delivered the opinion of the court:
Defendant, Fernando Zayas, was convicted in the circuit court of Cook County on three counts of murder. *286The court sentenced him to three concurrent terms of life imprisonment without the possibility of parole. The appellate court affirmed the convictions. (159 Ill. App. 3d 554, 568.) We granted defendant’s petition for leave to appeal (107 Ill. 2d R. 315). We reverse, holding that a witness other than the defendant himself may not offer testimony to the extent that it is enhanced through hypnosis.
The trial court found defendant guilty of the shooting deaths of Miguel Vargas, Luis Cuaresma and Ruben Gutierrez. These three young men were brutally murdered on the front porch of a building on Chicago’s northwest side in the early morning of July 2, 1983. In presenting its case, the State relied principally upon the testimony of four individuals. Defendant challenged the veracity of these witnesses and presented evidence of his own to support his asserted alibi defense.
One of the State’s principal witnesses was Socorro Roldan. Roldan, a former gang member with a criminal record, testified that Zayas admitted the killings at a party they both attended. This party was supposedly held the evening following the killings at Zayas’ girlfriend’s house, and was for the purpose of celebrating the murders. Roldan also testified that Zayas stated that he and three friends, including one named Jose ‘‘Baby” Rodriguez, took Juanita Rodriguez’s car without her knowledge to aid in the commission of the murders. Juanita is the sister of Jose. Roldan stated that the vehicle was a light-blue, two-door car. The State contends that Zayas’ motive for the murders was to avenge the death of his friend “Chi Chi,” whose name Zayas had tattooed on his arm. The State was unable to offer any other testimony to corroborate Roldan’s assertion that this party occurred or that Zayas confessed to the murders.
Roldan testified that he first related this story to the police on July 30, 1983, almost a month after the mur*287ders, after having been shot by a fellow gang member. The police later tried to secure a confession on tape by having Roldan elicit a statement from Zayas while he visited Roldan at the hospital. Zayas did not make such a confession, and in fact denied responsibility, but the State emphasized Zayas’ statement that someone was “tricking” on him, indicating that someone was informing the police of the events for which Zayas was allegedly responsible.
The State also called Carlos Vargas, the brother of Miguel Vargas, one of the victims. Vargas testified that he had knowledge of Zayas’ association with the “Latin Disciples” street gang and Miguel’s association with the “Insane Unknowns” street gang. Vargas further testified that he was on the premises at the time of the shooting and that he saw Zayas firing a gun at the victims. This testimony was corroborated by the testimony of Julia Tiro, who was the cousin of Luis Cuaresma, another victim, and girlfriend of Miguel Vargas, and Ruby Mateo, who was Carlos Vargas’ girlfriend.
On cross-examination, defendant’s attorney emphasized the fact that Vargas told a substantially different story to the police immediately following the incident. That is, during the initial investigation Vargas told the police that he was in the kitchen cooking when he heard the shots fired, and that he saw nothing. This story was also initially corroborated by Julia Tiro and Ruby Mateo. Vargas stated that he was frightened, which is why he did not come forward with his revised story until July 13,1983.
Another principal witness upon whom the State relied was Timothy McGovern. McGovern, who was 14 years old at the time of the shooting, testified that he saw Zayas firing a gun at the victims. McGovern stated that he was on his way home from his girlfriend’s house at the time, and that his house was almost directly across the street *288from where the incident occurred. McGovern testified that he witnessed the shooting from across the street, near his home. He was later able to pick Zayas’ picture out of a book and then identified Zayas in a line-up.
On cross-examination, defendant’s attorney questioned McGovern concerning why he had not related this story to the police until August 16, 1983, six weeks after the murders, when the police were investigating the shooting of his brother. McGovern also stated that he was afraid. Defendant’s attorney additionally questioned McGovern concerning the details of the events McGovern claimed he witnessed that evening. Zayas later presented the testimony of an investigator in an attempt to impeach McGovern’s testimony by demonstrating that the scenario McGovern described was implausible.
The State also put Detective Michael Atkins on the stand. Detective Atkins testified that, shortly after the murders, he responded to a call which stated that shots had been fired. He stated that when he and his partner approached the scene, however, noticing that other units had already responded and that the street had been secured, they continued on. Detective Atkins testified that they then received another call indicating that two females of Latin ancestry might be involved. He stated that, as he and his partner proceeded, a car containing four males pulled in front of them but that there was nothing unusual about this car apart from the fact that one of its occupants continually looked over his shoulder. Detective Atkins further testified that he then pulled along side another car which four people occupied, two of whom were women appearing to be of Latin ancestry, but that, upon getting a close look, he concluded that they were likely too old to have been involved.
Detective Atkins stated that he then returned to the scene of the shooting where some other officers informed him that a car similar to the car containing the *289four males that he saw earlier might have been involved. It is not clear from the record why these officers suspected such a car. He testified that, upon reflection, he was able to describe the car as a light-blue Plymouth Se-bring with a possible license plate number of XND 405.
On July 22, 1983, Dr. Bennett Braun, a psychiatrist certified in hypnosis, hypnotized Detective Atkins. The purpose of this hypnosis was to aid Detective Atkins in recalling the license plate number of the vehicle which he partially described earlier. Over defendant’s objection, Detective Atkins testified that he recalled under hypnosis that the license plate number of the vehicle was NXJ 402. The actual license plate number of the car belonging to Juanita Rodriquez was NXJ 240.
Defendant contends on appeal that the testimony of Detective Atkins, insofar as he related to the jury that which he recalled under hypnosis, was inadmissible. The appellate court agreed with defendant that this hypnotically induced testimony was inadmissible, but held that any error in its admission was harmless. We agree with the appellate court that the trial court erred in admitting this testimony, but we disagree with the appellate court’s conclusion that this error was harmless.
This court has not yet addressed the question of whether hypnotically enhanced testimony is admissible. In People v. Wilson (1987), 116 Ill. 2d 29, 42-49, this court outlined the state of the law in other jurisdictions but, because this precise issue was not before the court at that time, we did not decide it. The present case does, however, provide this court the opportunity to decide the question.
Hypnosis is, to say the least, an enigma. Its therapeutic uses range from curing smoking to helping athletes deal better with stress. Its use in criminal investigations has increased dramatically in recent years, moreover, *290and its successes are almost legendary. It has not, however, been a stranger to failure.
The American Medical Association has defined hypnosis as the following:
“ ‘[A] temporary condition of altered attention in the subject which may be induced by another person and in which a variety of phenomena may appear spontaneously or in response to [verbal] or other stimuli. These phenomena include alterations in consciousness and memory, increased susceptibility to suggestion, and the production in the subject of responses and ideas unfamiliar to him in his usual state of mind.’ ” (Note, The Admissibility of Hypnotically Refreshed Testimony, 20 Wake Forest L. Rev. 223, 224 (1984), quoting Council on Mental Health, Medical Use of Hypnosis, 168 J.A.M.A. 186, 187 (1958).)
As this rather imprecise definition indicates, the nature of hypnosis remains largely a mystery, and can be defined only by its symptoms and manifestations. The various theories explaining hypnosis, moreover, are often in conflict and may never be fully understood. (Diamond, Inherent Problems in the Use of Pretrial Hypnosis on a Prospective Witness, 68 Calif. L. Rev. 313, 317 (1980).) Because no one really understands it, hypnosis plays a rather dubious role in judicial proceedings.
There are essentially three different rules that this court could adopt regarding the admissibility of hypnotically refreshed recollections. There is certainly no consensus as to which is most appropriate. As the New York Court of Appeals aptly stated in People v. Hughes (1983), 59 N.Y.2d 523, 540, 453 N.E.2d 484, 493, 466 N.Y.S.2d 255, 264, “the law is in a state of flux and there is no rule which will entirely satisfy all demands of logic, policy and practicality.”
The first approach, originally espoused in Harding v. State (1968), 5 Md. App. 230, 246 A.2d 302, is a rule of per se admissibility which places the burden on the jury to *291determine the witness’ credibility. The second approach, which the New Jersey Supreme Court adopted in State v. Hurd (1981), 86 N.J. 525, 432 A.2d 86, places the burden on the trial judge to determine whether law enforcement officials have complied with an elaborate list of procedural safeguards. The third rule, which has been labelled as the “emerging consensus” (Zani v. State (Tex. Crim. App. 1988), 758 S.W.2d 233, 240 n.5), is a rule of per se inadmissibility which relieves trial judges of the burden of determining the quality of such evidence and relieves jurors of the responsibility of determining its credibility, tasks which neither are particularly well suited to perform because of the nature of this evidence.
The rule of per se admissibility has sparsely been followed since 1980. In fact, even in Maryland, where this rule was born, courts no longer follow it. (See, e.g., State v. Collins (1983), 296 Md. 670, 464 A.2d 1028.) Because of its many flaws, this court has no interest in reviving it.
The basis of the rule set forth in Harding was that attorneys have an opportunity to cross-examine previously hypnotized witnesses and may present their own expert witnesses to rebut the testimony of the experts that the proponents of the evidence examined. This rationale is flawed in several respects. First, previously hypnotized witnesses are virtually immune from effective cross-examination. That is, having been hypnotized, the subject gains complete confidence in his “restored” memory, forgets how it was “restored,” and is unable to differentiate between that which he was able to recall before hypnosis and that which the hypnosis elicited. Any information that the hypnotist suggested, or the subject fantasized or confabulated, becomes part of the subject’s “restored” memory, moreover, and the subject is likewise unable to differentiate between this data and that which he was able to independently recall prior to being hypnotized. Ruffra, Hypnotically Induced Testi *292 mony: Should It Be Admitted, 19 Crim. L. Bull. 293, 297 (1983).
Another problem associated with merely allowing the jury to weigh the credibility of a previously hypnotized witness is that jurors are usually unable to properly do so. The public has been barraged with misinformation concerning hypnosis and has resultantly been led to incorrectly believe that it provides a panacea for lost memory. (Mickenberg, Mesmerizing Justice: The Use Of Hypnotically-Induced Testimony In Criminal Trials, 34 Syracuse L. Rev. 927, 958 (1983).) While some studies show that, when properly administered, hypnosis can increase one’s quantitative memory, these studies also demonstrate that subjects tend to fill in forgotten details with details they concoct and are later unable to discern between the two. See generally Mickenberg, at 945 n.101, citing Orne, The Use and Misuse of Hypnosis in Court, 27 Int’l J. of Clinical & Experimental Hypnosis 312, 319 (1979).
The second alternative calls for the development of a strict set of procedural safeguards. The New Jersey Supreme Court set forth an elaborate set of procedural safeguards in Hurd. While these checks provide the trial court with an opportunity to effectively determine whether any conspicuous suggestiveness transpired during the hypnotic session, it does little to cure other problems inherent in the process.
One problem with allowing hypnotic testimony subject to procedural safeguards is that these checks do not detect whether the hypnotist unintentionally or subconsciously implanted suggestions in the mind of the subject. In fact, hypnosis by its very nature requires the hypnotist to suggest answers to the subject. The subject, moreover, will inevitably fill in missing details and fantasize about events that never in fact happened. Procedural safeguards do nothing, furthermore, to eliminate *293the problems associated with bolstered witness confidence and erroneous juror perception. Finally, this second alternative puts considerable strain on a trial court in that the court is forced to hear testimony from several competing experts and then make a rather sophisticated determination of whether the probative value of the testimony will outweigh its prejudicial effect.
Because of the shortcomings of these two rules, the third standard, that of per se inadmissibility, has gained wide-spread acceptance. (See Contreras v. State (Alaska 1986), 718 P.2d 129, 133 n.14 (and cases cited).) It properly applies the standard set forth in Frye v. United States (D.C. Cir. 1923), 293 F. 1013, which held that a court should not allow into evidence a “scientific” test which the relevant scientific community has not recognized as reliable. This court, in holding polygraph evidence inadmissible, applied the standard set forth in Frye:
“ ‘Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is difficult to define. Somewhere in this twilight zone the evidential force of the principle must be recognized, and while courts will go a long way in admitting expert testimony deduced from well-recognized scientific principle or discovery, the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs.’ ” People v. Baynes (1981), 88 Ill. 2d 225, 241, quoting Frye, 293 F. at 1014.
See also People v. Jordan (1984), 103 Ill. 2d 192, 208.
Application of the Baynes polygraph analysis to the issue of whether hypnotically induced testimony is admissible leads inevitably to the conclusion that such evidence should not be admitted.
In Baynes, this court noted the inaccuracy of the polygraph test. So too, studies indicate that hypnosis *294elicits incorrect recall in varying degrees. Additionally, the Baynes court acknowledged the inevitable intrusion upon the proper functioning of the jury by the admission of evidence which “is likely to be shrouded with an aura of near infallibility, akin to the ancient oracle of Delphi.” (Bayres, 88 Ill. 2d at 244.) Jurors seem to demonstrate the same misunderstanding and false sense of confidence in hypnotic evidence as they have historically shown with regard to polygraph evidence. Mickenberg, at 958.
The supreme court of Alaska, in Contreras, 718 P.2d at 135, provides four persuasive reasons why the Frye standard is particularly well suited to handle this question:
“(1) [T]he standard is judicially manageable; (2) the standard saves judicial time and resources; (3) the standard assures that juries will not be misled by unproven, unsound ‘scientific’ procedures, thus safeguarding the court’s truth-finding role; and (4) the standard assures fairness and uniformity of decision-making.”
The Contreras court also appropriately recognized that hypnosis is a technique which elicits scientific evidence and cannot properly be distinguished from other forms of scientific evidence simply because the subject provides the testimony rather than the “scientist.” That is, having been hypnotized, the subject’s memory, and therefore his testimony, becomes the product of his prehypnosis memory, his memory as genuinely augmented by hypnosis, his fantasy and confabulation, and the suggestion of the hypnotist. The subject, the hypnotist, or any other objective observer cannot later decipher these various components.
Frye mandates that trial courts not admit evidence which the relevant scientific community has not generally accepted as reliable. The relevant scientific community not only does not generally accept that hypnotically induced recall is accurate, but the legal and scientific lit*295erature is now replete with articles imploring courts to reject such evidence because of its many flaws. (See, e.g., Mickenberg, 34 Syracuse L. Rev. 927; Ruffra, 19 Crim. L. Bull. 293; Diamond, 68 Calif. L. Rev. 313; Alderman & Barrette, Hypnosis on Trial: A Practical Perspective on the Application of Forensic Hypnosis in Criminal Cases, 18 Crim. L. Bull. 5 (1982).) It is not imperative, moreover, that these scientists who doubt the accuracy of hypnotically induced recall are proven to be correct. What is important is that many doubt its accuracy in restoring one’s memory.
As such, we find that because its reliability is suspect, and it is not amenable to verification due to the fact that even the experts cannot agree upon its effectiveness as a memory-restorative device, a witness’ hypnotically induced testimony, other than that of the defendant, is not admissible in Illinois courts. The fact that information elicited through hypnosis can be corroborated, moreover, does not permit a court to admit it as evidence. We are not concerned with whether evidence is plausible but, rather, whether evidence is reliable. The trial judge emphasized that he found no suggestion in the videotaped hypnotic session and that the information elicited through hypnosis would be corroborated by Roldan’s testimony. This seems to indicate that if Detective Atkins would have recalled a license plate number that does not exist, the judge might not have admitted the evidence for the purpose of impeaching Officer Atkins’ testimony. This is incongruous.
One argument that some have raised in opposition to a per. se rule excluding hypnotically enhanced testimony is that it runs contrary to modern evidence rules which would allow the jury to hear all the evidence and draw its own conclusions, unless the trial judge determined that certain evidence was excessively prejudicial, and that this rule usurps the exclusive power of trial judges *296in making such decisions which they are better equipped to make. (See, e.g., Note, Rock v, Arkansas: Hypnosis & the Prejudice Rule — Your Memories May Not Be Your Own, 21 J. Marshall L. Rev. 409 (1988).) While we agree this court should be cautious when intruding upon the truth-finding functions of trial courts, we think this rule will enhance rather than hinder that process. As indicated earlier, the probative value of any hypnotically enhanced testimony is questionable since most scientists doubt its accuracy. Furthermore, procedural safeguards do not cure these problems. Neither trial judges, nor juries, moreover, are well suited to take such evidence for what it may be worth because of erroneous public perception. Admission of such evidence at the discretion of trial judges would also result in inconsistent results and would strain the judicial process.
The State maintains that the only problem with which this court should concern itself with regard to hypnotically induced testimony is suggestiveness, and that no such suggestiveness is found in this case. First, as indicated earlier, suggestiveness is only one of a myriad of problems that hypnosis presents when offered as testimony in court. To say that suggestiveness is absent in this case, moreover, is a misstatement, for suggestion is an essential component of any hypnotic process. The fact that Dr. Braun was unaware of the license plate number is not determinative. The possibility that Dr. Braun suggested the actual license plate number is but one of many problems this court finds with the process. Some of the other problems that the trial court did not address, and could not properly address, include whether Officer Atkins received some suggestion of the actual license plate number before hypnosis, for instance from a picture or police report, the effect of the hypnosis on Officer Atkins regarding whether the State could properly cross-examine him, and the effect of the introduction of *297hypnotic evidence on the jury by bolstering jury confidence in Detective Atkins’ testimony and, correspondingly, Socorro Roldan’s testimony. Because it could not properly make these determinations, this court finds that the trial court erred by allowing such testimony.
This ruling is not inconsistent with our prior holding in Wilson which allowed a witness to testify as to her prehypnotic recall. The proponent of prehypnotic recall evidence, however, will bear the burden of establishing that the testimony of the previously hypnotized witness is based solely upon that witness’ independent, prehypnotic recall.
The ruling in this case is also consistent with the recent Supreme Court ruling in Rock v. Arkansas (1987), 483 U.S. 44, 97 L. Ed. 2d 37, 107 S. Ct. 2704. There, the Supreme Court ruled that a defendant may, by right, introduce hypnotic evidence because the defendant’s right to testify on his own behalf, among other rights, outweighs the State’s interest in excluding prejudicial evidence, and certain procedural safeguards will protect the integrity of the court. Our case, however, involves a witness for the State, not the defendant himself wishing to testify. As such, there are no constitutional implications and, therefore, Rock does not affect the decision in this case. While the Supreme Court did implicitly acknowledge that hypnotic evidence can have some probative value, its decision does not preclude this court from adopting a per se rule excluding the hypnotically enhanced testimony of a witness other than the defendant.
The appellate court came to this same conclusion, but found that any error in admitting this evidence was harmless. For several reasons, we find that, without this evidence, the jury could have reached a verdict of not guilty.
First, Officer Atkins’ testimony placed defendant near the scene of the crime at the time of the incident. *298Without the testimony concerning the license plate number, Detective Atkins essentially could state only that he saw a light-blue car with a possible license number of XND 405 containing four men with olive colored skin. Second, the fact that Detective Atkins submitted to hypnosis bolstered his confidence in his testimony and impressed the jury that his memory was flawless. Finally, Detective Atkins’ hypnotically enhanced testimony corroborated the testimony of Socorro Roldan, probably the State’s most damaging witness. This provided the jury with confidence in Roldan’s testimony, whose veracity was otherwise somewhat suspect.
With the foregoing in mind, this court concludes that, without the hypnotically aided testimony of Officer Atkins, the jury quite possibly could have reached a different verdict. As it was, the jury deliberated over a day, was repeatedly instructed pursuant to People v. Prim (1972), 53 Ill. 2d 62, and reached its verdict at 6:30 p.m. on Good Friday. For these reasons, the defendant’s convictions are reversed and the cause is remanded for a new trial.
We understand that other criminal cases in Illinois have involved the use of hypnosis. This ruling, however, will not affect those cases that have been finally determined on direct appeal. Our holding on the above issue makes it unnecessary to consider other questions raised on this appeal.
Appellate court reversed; circuit court reversed; cause remanded.