Brain Potentials Before and After Rapid Eye Movements: an Electrophysiological Approach to Dreaming in REM Sleep

SLEEP PHYSIOLOGY Brain Potentials Before and After Rapid Eye Movements: an Electrophysiological Approach to Dreaming in REM Sleep Keiko Ogawa, MSc; Hiroshi Nittono, PhD; Tadao Hori, PhD Department of Behavioral Sciences, Faculty of Integrated Arts and Sciences, Hiroshima University, Higashi-Hiroshima, Japan Study Objective: This study examined hypotheses regarding dreaming in rapid eye movement (REM) sleep by comparing brain potentials related to rapid eye movements in REM sleep with those in wakefulness. Design: Within participants. Setting: Data were collected in a sleep laboratory. Participants: Fifteen healthy university students who reported having dreams frequently. Measurements and Results: Rapid eye movements in REM sleep were recorded during natural nocturnal sleep. Saccades in wakefulness were recorded during a self-paced visual search task. The presaccadic negativity before and the lambda response after eye movements were examined. It was assumed that the presaccadic negativity reflects voluntary readiness activity before eye movements, and the lambda response reflects visual information processing after saccades in wakefulness. Brain potentials were averaged, time-locked to the onset and offset of INTRODUCTION IN RAPID EYE MOVEMENT (REM) SLEEP, THE EXPERIENCE OF DREAMING IS MORE FREQUENT, MORE VIVID, AND CLEARER THAN IN NON-REM SLEEP.1 Many researchers have investigated the relationship between dreaming and rapid eye movements.1-4 The “activation-synthesis hypothesis”2 and the “sensory image-free association hypothesis”3 both suggest that rapid eye movements occur spontaneously without voluntary control and generate a dream image via the activation of relevant cortical areas. The activation-synthesis hypothesis has evolved into Hobson’s activation/input source/neuromodulation model (AIM), which describes the brain-mind state in a 3-dimensional space: activation, input source, and mode of information processing.4 Another type of hypothesis, “scanning hypothesis,” suggests that rapid eye movements reflect the dreamer’s voluntary watching of visual imagery in a dream. Recent brain imaging studies have made it possible to document the activation of the brain related to rapid eye movements in REM sleep.4-8 Rapid eye movements during REM sleep are analogous in shape to saccades during wakefulness.1,9-12 Saccades are fast ocular movements between visual fixations. Investigation of brain activity associated with rapid eye movements may reveal Disclosure Statement This was not an industry supported study. Drs. Ogawa, Nittono, and Hori have indicated no financial conflicts of interest. Submitted for publication August 2004 Accepted for publication May 2005 Address correspondence to: Tadao Hori, PhD, Department of Behavioral Sciences, Faculty of Integrated Arts and Sciences, Hiroshima University, 1-7-1 Kagamiyama, Higashi-Hiroshima, 739-8521, Japan; Tel: 81 82 424 6580; Fax: 81 82 424 0759; E-mail: SLEEP, Vol. 28, No. 9, 2005 1077 eye movements for the presaccadic negativity and the lambda response, respectively. In wakefulness, the presaccadic negativity occurred at the centroparietal site. However, no presaccadic negativity was found during REM sleep. Lambda-like responses (P1r, P2r) were observed in REM sleep over the parietooccipital site, as were the lambda responses (P1, P2) in wakefulness. Conclusions: The finding that rapid eye movements are initiated without preparation but elicit some neural activity in the cortical visual area suggests that rapid eye movements may trigger dream images. Keywords: REM sleep; rapid eye movements; saccades; presaccadic negativity; lambda response; dreaming. Citation: Ogawa K; Nittono H; Hori T. Brain potentials before and after rapid eye movements: an electrophysiological approach to dreaming in REM Sleep. SLEEP 2005;28(9): 1077-1082. how dreams occur in REM sleep. In studies using hemodynamic brain imaging techniques, there is controversy as to whether rapid eye movements share the same mechanism of generation as saccades.7,8 Although brain imaging techniques have high spatial resolution, their limited temporal resolution is not sufficient to monitor changes in brain activity before and after eye movements. An electrophysiologic approach with higher temporal resolution is useful for such an investigation. In the present study, we compared rapid eye movements during REM sleep and saccades during wakefulness in terms of brain potentials before and after eye movements. Brain potentials before and after saccades have been reported in electrophysiologic studies of wakefulness. The presaccadic negativity (PSN) occurs about 600 milliseconds before the saccade, with maximal amplitude over the centroparietal region.13-16 It is similar to the readiness potential and reflects the voluntary readiness activity of eye movements. A saccade on the side opposite the appearance of a target is called an antisaccade.17 On the other hand, a prosaccade is a saccade on the same side as the target.17 Antisaccades require voluntary control over reflexive behavior. It is known that the PSN is significantly greater for antisaccades than prosaccades, which suggests that the PSN reflects voluntary control over saccades. However, it remains unclear whether the PSN occurs prior to rapid eye movements in REM sleep. LaBerge et al18 suggested that lucid dreamers could intentionally signal the presence of dreams. If rapid eye movements occur voluntarily, as in this case, then the PSN should appear prior to rapid eye movements. Following saccades, positive cerebral potentials, called the lambda response, appear at occipital sites.19-21 The lambda response is assumed to correspond to visual potentials after fixation.22 Previous studies have suggested that rapid eye movements elicit similar brain potentials called the lambda-like response, although there is no visual input in REM sleep.10-12 This finding An Electrophysiological Approach to Dreaming—Ogawa et al outer canthi of both eyes and from above and below the left eye with a time constant of 3.2 seconds. The horizontal electrooculogram was also recorded with a time constant of 0.032 seconds for picking the onsets of saccades and rapid eye movements. Electromyograms were recorded from the mentalis muscles with a time constant of 0.032 seconds. Interelectrode impedances were below 5 kΩ. The entire recording was made by Ag/AgCl electrodes affixed to the scalp with collodion or surgical tape. A high-cut filter of 100 Hz was used. The sampling rate was 250 Hz. The EEG was re-referenced to linked earlobes offline. suggests that some visual information processing activity may occur during REM sleep, as it does during wakefulness. In the present study, we recorded the lambda-like response, time-locked to the offset of rapid eye movements rather than the onset. Because the lambda response in wakefulness occurs timelocked to the offset of saccades, our approach should more clearly reveal the nature of the lambda-like respon (...truncated)