Ponto-Medullary Interactions in the Generation of REM Sleep

There has been cons iderable uncertainty about the loca l i za t i on and anatomical extent of the brainstem regions crit ical for REM sleep control. While recording experiments have shown that many brainstem cell groups change discharge rate and pattern in REM sleep, the key experimental evidence on the local izat ion of areas critical for REM sleep control must be derived from manipulations, with stimulation and lesion techniques, of the areas suspected of control l ing REM sleep. The loss of REM sleep a f te r brainstem lesions may be due to nonspecific factors. This is particularly true in medullary regions, where st imulation and lesions interfere with the regulation of respiration and blood pressure. Therefore, the medullary role in behavioral state control has remained largely unknown. Recently identif ied anatomical connections between lateral pontine regions and the medulla have increased interest in this role (see Jones, Sakai, this volume). Wh i le the loss of REM sleep may be a result of nonspeci f ic factors, the presence of REM sleep after the removal of a brain area is defini t ive evidence that the removed area is not required for generating this state. Therefore, we have performed a series of brainstem transection studies in the hope that we might find positive evidence bearing on the contribution of both pontine and medullary regions to the pattern of physiological events and periodicities that constitutes REM sleep. Figure 1 i l lustrates the key t ransect ion experiments in the local izat ion of REM s leep control mechanisms. Transection at levels A and B prevent REM sleep signs in the rostral portion of the brain, but a l l ow all of the brainstem s igns of REM sleep caudal to the cut, including rapid eye movements, extreme miosis, neck muscle atonia and normal durations and periods of recurrence for this REM sleep-l ike state (3,5,21,). Work in cats wi th spinal t ransect ions (Fig. 1C) and in humans with spinal injury has shown that the spinal cord makes no essential contribution to the brainstem signs of REM sleep (1,13). This posit ive evidence