537 So.2d 973 (1989)
James A. MORGAN, Appellant,
v.
STATE of Florida, Appellee.
No. 67334.
Supreme Court of Florida.
January 5, 1989.
Robert G. Udell, Stuart, for appellant.
Robert A. Butterworth, Atty. Gen., Joan Fowler Rossin and Eddie J. Bell, Asst. Attys. Gen., West Palm Beach, for appellee.
PER CURIAM.
James A. Morgan appeals his conviction on retrial for first-degree murder and his sentence of death. We have jurisdiction, article V, section 3(b)(1), Florida Constitution, and reverse the conviction and sentence and remand for a new trial. We conclude the trial court erroneously excluded medical expert opinion testimony that was based on a diagnosis which used information obtained from Morgan by hypnosis. We find the recent United States Supreme Court decision in Rock v. Arkansas, 483 U.S. 44, 107 S.Ct. 2704, 97 L.Ed.2d 37 (1987), mandates this result.
Morgan was sixteen years old at the time of the incident, of marginal intelligence, unable to read or write, had sniffed gasoline regularly since he was four, and was described as an alcoholic. He brutally murdered an elderly woman while at her home to mow her yard, after entering the house to telephone his father. Inside the home, appellant killed the woman by crushing her skull with a crescent wrench, stabbing her face, neck, and hands numerous times, and also biting her breast and traumatizing her genital area. According to Morgan, he killed the woman because he thought she was writing his mother about his drinking. There is no dispute over appellant's commission of this homicide; the single issue is appellant's sanity at the time of the offense, and the experts excluded were the only witnesses for appellant on this issue.
This is the third time this cause has been before this Court. In Morgan v. State, 392 So.2d 1315 (Fla. 1981) (Morgan I), we remanded the case because the bifurcated insanity procedure used in that trial had been subsequently held unconstitutional. In Morgan v. State, 453 So.2d 394 (Fla. 1984) (Morgan II), we remanded the case because the trial court denied Morgan an opportunity to present an insanity defense.
In this third trial, Morgan filed notice of his intent to rely on an insanity defense. During the opening statements, his counsel advised the jury that insanity would be his client's defense. The state presented its case, rested, and then moved to prevent Morgan from presenting his expert witnesses, a psychologist and a psychiatrist, on the grounds that their opinions were partially based on statements Morgan made while under hypnosis.[*] The trial court granted the motion based on this Court's decisions in Bundy v. State, 455 So.2d 330 (Fla. 1984) (Bundy I), and Bundy v. State, 471 So.2d 9 (Fla. 1985) (Bundy II), cert. denied, 479 U.S. 894, 107 S.Ct. 295, 93 L.Ed.2d 269 (1986), and the Third District Court's decision in Rodriguez v. State, 327 So.2d 903 (Fla. 3d DCA), cert. denied, 336 So.2d 1184 (Fla. 1976). As a result of the trial court's order, Morgan was precluded from presenting any expert testimony on the issue of insanity.
Morgan proffered the experts' testimony in the jury's absence. The psychologist testified that he met Morgan on three occasions. During the first visit, the doctor stated Morgan was initially reluctant to be honest, but, after encouragement, explained generally his killing of the woman. Morgan stated that: (1) before going to mow, he had been drinking; (2) the victim permitted him to enter her home to call his father; (3) he saw the victim writing and thought she was notifying his mother about his drinking; and (4) he became angry with the victim and hit her in the head with the crescent wrench. According to the psychologist, Morgan could not clearly remember his actions after striking the woman and had no recollection of cleaning up after the attack and leaving the premises.
In the second session, the psychologist performed various psychiatric tests and obtained additional background information. The personal history revealed that Morgan was sixteen years old at the time of the incident, had regularly sniffed gasoline since he was four and, in recent years, on more than a daily basis, and regularly used alcohol to the extent that the psychologist concluded he was a sixteen-year-old alcoholic. Testing revealed that, although he had completed the eighth grade, he could not read or write and was organically brain-damaged and brain-impaired.
After the second session, the psychologist, along with the psychiatrist, decided to hypnotize Morgan to obtain further details concerning the incident. Both doctors testified that hypnosis is a medically accepted diagnostic technique used by mental health professionals. According to both experts, the use of hypnosis facilitates diagnosis by revealing information which might otherwise be unavailable from the unhypnotized patient.
Morgan was hypnotized by the psychologist in the psychiatrist's presence. During a four-hour hypnotic session, Morgan provided more expansive details of his conduct in the killing. Both experts concluded, from their examination of Morgan, his history, and the hypnotic session, that he was insane at the time of the offense under the M'Naughten standard. Both testified they would have been unable to assess Morgan's sanity without utilizing the information from the hypnotic session.
In the penalty phase, the trial court allowed admission of the medical experts' testimony for the jury to consider mental impairment as a mitigating factor. The jury recommended the death penalty by a seven-to-five vote, and, accordingly, the trial judge imposed that sentence.
Morgan raises multiple issues concerning his conviction and sentence. We find dispositive his claim that the trial court erroneously excluded his expert witnesses' testimony during the guilt phase of the trial on grounds their opinions were partially based on statements made while Morgan was under hypnosis.
The trial judge excluded the experts' opinions based on Bundy I, Bundy II, and Rodriguez, holding that, since he could not determine the "reliability of statements procured under hypnosis," the opinions based on hypnotic statements were inadmissible. We do not criticize the trial court's ruling; we recognize hypnotic evidence is a new and evolving area of law. We find that the United States Supreme Court decision in Rock v. Arkansas controls.
In Rock, the defendant was charged with manslaughter of her husband. She could not remember the exact details surrounding the event and was hypnotized by a licensed neuropsychologist in order to refresh her memory. After hypnosis, she was able to recall that at the time of the shooting she had not had her finger on the trigger and the gun had discharged when her husband grabbed her arm during a scuffle. The gun was later found defective and prone to fire when hit or dropped, without the trigger being pulled. At trial, the court limited the defendant's testimony to only those matters remembered and stated prior to her being placed under hypnosis. On appeal, the Supreme Court of Arkansas affirmed the trial court, rejecting the defendant's claim that the limitations on her testimony violated her right to present her defense. The Arkansas court concluded that "the dangers of admitting this kind of testimony outweigh whatever probative value it may have," Rock v. State, 288 Ark. 566, 573, 708 S.W.2d 78, 81 (1986), and followed the view of those states which have held hypnotically refreshed testimony of witnesses inadmissible per se. The question addressed by the United States Supreme Court was "whether a criminal defendant's right to testify may be restricted by a state rule that excludes her post-hypnosis testimony." Rock v. Arkansas, 107 S.Ct. at 2710. In answering this question in the negative, the Court explained: "Just as a State may not apply an arbitrary rule of competence to exclude a material defense witness from taking the stand, it also may not apply a rule of evidence that permits a witness to take the stand, but arbitrarily excludes material portions of his testimony." Id. at 2711. The Court stated that the per se rule does not allow a trial court the opportunity to consider whether post-hypnosis testimony may be admissible in a particular case, concluding that "[t]his rule operates to the detriment of any defendant who undergoes hypnosis, without regard to the reasons for it, the circumstances under which it took place, or any independent verification of the information it produced." Id. at 2712. The opinion noted that many states which have adopted the exclusionary rule did so only as to witnesses and not for a defendant's testimony. The Court cited the California rule contained in People v. Shirley, 31 Cal.3d 18, 723 P.2d 1354, 181 Cal. Rptr. 243, cert. denied, 459 U.S. 860, 103 S.Ct. 133, 74 L.Ed.2d 114 (1982), where that court barred the entire testimony of any witness who had been hypnotized but explicitly accepted a defendant's testimony by stating:
[W]hen it is the defendant himself not merely a defense witness who submits to pretrial hypnosis, the experience will not render his testimony inadmissible if he elects to take the stand. In that case, the rule we adopt herein is subject to a necessary exception to avoid impairing the fundamental right of an accused to testify in his own behalf.
31 Cal.3d at 67, 723 P.2d at 1384, 181 Cal. Rptr. at 273.
We found in Bundy that hypnosis had not received sufficient scientific acceptance to be held reliable as substantive evidence and concluded that "hypnotically refreshed testimony is per se inadmissible in a criminal trial in this state." Bundy II, 471 So.2d at 18. Rock mandates that we recede from the Bundy II rule to the extent it affects a defendant's testimony or statements made to experts by a defendant in preparation of a defense.
Even without the Rock decision, we would conclude that expert testimony in this instance must be allowed. The issue is not whether Morgan's hypnotic statements are reliable testimony to prove the truth of the matter asserted. Rather, the question is limited to whether mental health experts can testify about Morgan's sanity if their opinion is based in part on information received from hypnotic statements obtained through a medically approved diagnostic technique. The evidence sought to be presented here is distinguishable from that of the Bundy cases or the Rock case. In Bundy I and Bundy II, the state sought to introduce statements from hypnotic sessions as direct evidence to prove the truth of the matter by refreshing a witness's recollection. In Rock, the defense attempted to present direct evidence to prove the truth of the matter asserted by refreshing the defendant's recollection.
We note that although Bundy prohibits the offering of hypnotically refreshed testimony as direct evidence, it does not preclude all uses of hypnosis. In Bundy II, this Court stated that "we do not undertake to foreclose the continued use of hypnosis by the police for purely investigative purposes. Any corroborating evidence obtained is admissible in a criminal trial subject to other evidentiary objections." 471 So.2d at 19.
Courts cannot establish accepted medical practices; they can only ensure that accepted methods are properly utilized. We conclude that, even without the United States Supreme Court Rock decision, Morgan should have been permitted to introduce conclusions drawn from medically accepted techniques. Here, his mental health experts were effectively barred from using medically accepted procedures to diagnose him. If courts seek medical opinions, they cannot bar the medical profession from using accepted medical methods to reach an opinion.
The use of hypnosis is an evolving issue and, clearly, some safeguards are appropriate to help assure reliability in the courts. We find it appropriate in the future, when hypnosis may be used to refresh a defendant's memory or by an expert witness to facilitate a medical diagnosis, that reasonable notice be given to the opposing party. Additionally, the hypnotic session should be recorded to ensure compliance with proper procedures and practices. At this time we recede from Bundy II only as it pertains to the defendant as a witness.
In the instant case, there is no doubt that Morgan committed the murder. Rather, the sole issue is his sanity at the time he committed the offense. As reflected in this opinion, Morgan was not able to present evidence on this question.
For the reasons expressed, we vacate Morgan's conviction and sentence and remand the cause for a new trial.
It is so ordered.
EHRLICH, C.J., and OVERTON, McDONALD, BARKETT, GRIMES and KOGAN, JJ., concur.
SHAW, J., concurs in result only with an opinion.
SHAW, Justice, concurring in result only.
I write separately to identify areas of disagreement with the majority opinion. First, the Court in Rock v. Arkansas, 483 U.S. 44, 107 S.Ct. 2704, 97 L.Ed.2d 37 (1987), was very careful to limit its holding to the hypnotically induced testimony of a defendant. Id. 107 S.Ct. at 2712 n. 15. While Rock is highly relevant to the issue of expert testimony based on the examination of a defendant under hypnosis, the decision is not controlling as the majority states. The Rock court recognized that hypnotically recalled testimony ran a substantial risk of unreliability. One of the major safeguards on which the Court relied to cure this unreliability was cross-examination. However, the testimony of an expert witness on the hypnotic recall of a defendant who does not himself take the stand presents major difficulties in cross-examination not present in the Rock context. Because this is an evolving area of law, I would not unqualifiedly hold that "statements made to experts by a [hypnotically enhanced] defendant in preparation of a defense" are always admissible. Op. at 976.
I also do not agree with the majority's unqualified acceptance of "accepted medical practice" as the criterion for determining whether testimony is admissible before a court. I certainly agree that we cannot, and should not, tell the medical community what constitutes accepted medical practice. Quite rightly, the medical profession will determine what medical techniques are best for a patient without reference to legal standards of admissibility of evidence. Conversely, the courts will independently determine what testimony or evidence meets legal criteria for admissibility. These standards are not necessarily congruent. Nothing in Rock suggests that the courts' authority and responsibility to ensure that evidence is sufficiently reliable to warrant admission has been abrogated. Rock only holds that the state may not have a per se rule that excludes hypnotically recalled testimony of a defendant who takes the stand. As the Court said:
The State would be well within its powers if it established guidelines to aid trial courts in the evaluation of posthypnosis testimony and it may be able to show that testimony in a particular case is so unreliable that exclusion is justified.
Rock, 107 S.Ct. at 2714.
Finally, I am convinced that in light of Rock we have no choice but to recede entirely from the per se rule of Bundy v. State, 471 So.2d 9 (Fla. 1985), cert. denied, 479 U.S. 894, 107 S.Ct. 295, 93 L.Ed.2d 269 (1986), that hypnotically refreshed testimony is inadmissible in a criminal trial. The broad message I receive from Rock is that we should not adopt per se rules of exclusion at the beginning of an evolving medical practice, i.e., hypnosis as a tool for memory recall. If we maintain the Bundy per se rule, as slightly modified by the decision here, evolution and experience, as it applies to criminal proceedings, cannot take place in Florida. This and other Florida courts will be bystanders awaiting incremental directions from the United States Supreme Court which may well result in numerous reversals of Florida convictions or, in the case of acquittals, the denial of relevant hypnotically recalled evidence. Instead of a per se rule of exclusion, I would adopt the approach outlined in Judge Ervin's thoughtful examination of the issue in Brown v. State, 426 So.2d 76 (Fla. 1st DCA 1983), of which we spoke approvingly in Bundy v. State, 455 So.2d 330 (Fla. 1984). Briefly, the approach calls for a threshold determination by the trial judge of the reliability and relevance of the hypnotically recalled evidence based on the specific circumstances of the hypnotic session(s), novelty and want of general scientific acceptance being only one facet in the court's relevancy analysis. Thereafter, assuming the evidence is admitted, the jury should be instructed on the potential shortcomings of the technique and the parties should be permitted to attack or defend the technique using available authorities and evidence. In essence, the issue at this stage becomes one of weight and credibility for the jury.
I concur in result because I believe there was sufficient indicia of reliability and relevance in the denied testimony to warrant presenting it to the jury.
NOTES
[*] The trial judge advised both counsel that all foreseen legal issues should be raised at pretrial conference. Although the state knew prior to pretrial conference that Morgan's expert witnesses would use hypnotic statements as a basis for their opinions, the state waited to challenge the opinion testimony until just before Morgan was to present his experts.