CUTTING THE REM NERVE: AN APPROACH TO THE ADAPTIVE ROLE OF REM SLEEP RAMON GREENBERG, M.D., and CHESTER PEARLMAN, M.D.* Extensive research in the psychophysiology of sleep has failed to demonstrate a function for rapid eye movement (REM) sleep about which all or most investigators would agree. Hennevin and Leconte [1] have recently presented a complete review of the various hypotheses for a function of REM sleep that have so far been presented. Their review concludes that the hypothesis that REM sleep is involved in information processing seems the most tenable and experimentally verifiable of the hypotheses about REM sleep. This hypothesis has been presented by a number of authors [1] and yet has failed to gain general acceptance. We would like to begin with this hypothesis and demonstrate how, with some refining, it can help us organize what has seemed to be a mass of conflicting and contradictory experimental findings. InAn Introduction to the Study ofExperimental Medicine, Claude Bernard [2] made some cogent observations about problems in physiologic research which may be applicable to this situation. In one example [2, p. 199] he described how, while looking for one result, he found exactly the opposite. The experiment involved cutting the cervical sympathetic nerve in the rabbit. He expected to find a fall in skin temperature on the side of the lesion. Instead, he found a striking increase in temperature and circulatory hyperactivity. Most relevant to our present discussion was Bernard's observation that he and other investigators had made this lesion before but had never noticed these changes. They had had no hypothesis to focus their attention on the temperature change. "We had the fact under our eyes and did not see it because it conveyed nothing to our mind. However, it could hardly be simpler to perceive, and since I described it, every physiologist without exception has noted and verified it with the greatest ease." We would like to use Bernard's observation as an analogy and to apply this analogy to the phenomena resulting from REM sleep deprivation (REMD) or "cutting the REM nerve." We will try to show that the procedure of REMD (that is, depriving the organism of *Department of Psychiatry, Tufts University School of Medicine. Address: V.A. Hospital , 150 South Huntington Avenue, Boston, Massachusetts 02130. Perspectives in Biology and Medicine · Summer 1974 | 513 the functions of REM sleep), given an organizing hypothesis, leads to similarly verifiable effects which have previously remained hidden in the mass of seemingly contradictory experimental findings. We will try to examine all of the published studies on the effects of REMD. After the seminal discovery that the objective, electrophysiologic signs of REM sleep were associated with dreaming, the expectation arose that the function of dreaming could now be determined. When Dement [3] showed that REMD1 in humans led to an increased pressure for REM sleep, as indicated by a rebound of REM sleep on recovery nights, and some psychologic changes in REMD subjects, this expectation seemed close to fulfillment. Subsequent studies, however, failed to find consistent psychologic changes in REMD subjects [4-8]. This failure led to some disillusionment with the idea that REM sleep or dreaming had an important psychologic function. Was this skepticism warranted, or, as in Bernard's example, were the investigators not looking for the right thing? Perhaps they were influenced by Freud's hypothesis that dreaming served to discharge drive tension. The absence of evidence that REMD regularly produced signs of drive pressure, such as heightened anxiety, supported the conclusion that REMD had no major psychologic effects. On the other hand, studies utilizing the Rorschach or Holzman (an analogue of the Rorschach) tests have consistently indicated that REMD has some psychologic effect in humans. This is in contrast to the finding that repeated use of the Rorschach, without intervening personality change, shows considerable stability [9, pp. 446-460]. In three studies, REMD produced clear changes in these test protocols which differed from those associated with control interruptions of non-REM (NREM) sleep. In our study [10], we found that the changes were specific for each subject. They are best characterized as indications in the REMD projective material of diminishing and shifting of...