Sleep Paralysis as an Anomalous REM State 

REM and Dreaming: A major distinction of sleep states, for close to a half century, has been accepted between REM and NREM sleep(Aserinsky & Kleitman, 1953; Jouvet, 1967). REM periods are characterized by desynchronized cortical characterized by low-voltage fast EEG patterns with synchronized hippocampal activity characterized by slow (4-8 Hz) theta activity (e.g., Culebras, 1994). It is also widely accepted that dreaming is more common and more vivid during REM than during NREM sleep (Dement & Kleitman, 1957). In addition to the characteristic desynchronized cortical low-voltage fast EEG activity, there are numerous physiological, behavioral, and sensory features associated with REM such as muscle atonia, gating of sensory input, rapid eye and middle ear movements, as well as heart rate and respiration changes (Carskadon & Dement, 1989; Symons, 1993). 

Within REM periods a distinction is sometimes made between a background tonic state (TREM) and bursts of phasic REM (PREM) every 16-120 seconds and lasting from 2-9 seconds (Aserinsky, 1971, Molinari & Foulkes, 1969). Specifically, PREM is characterized by bursts of rapid eye and middle ear movements and characteristic cortical and hippocampal EEG patterns. PREM is associated with, and may be preceded by, ponto-geniculo-occipital EEG waves (PGO spikes in animal preparations) originating in the bilateral, dorsolateral pons and projecting rostrally through the lateral geniculate nucleus and other thalamic nuclei (Hobson, Alexander, Frederickson, 1969). It has been conjectured that the most vivid dreams, or most vivid events within dreams, are associated with PREM (Molinari and Foulkes, 1969). 

REM and SP: SP has also been associated with REM states, particularly with sleep?onset and sleep?offset REM (SOREM) (Nan'no, Hishikawa, & Koida, 1970). In both REM dreams and SP hallucinations a general atonia is maintained during REM by marked and sustained hyperpolarization of the motoneurons (Chase & Morales, 1989). One likely function of the general atonia is the prevention of the physical enactment of the motor components of dreaming. There are at least two major traditional hypotheses concerning the connection between neurophysiological events and visual imagery in dreams. The visual imagery of dreaming may arise either from the direct stimulation of visual areas of the cortex during the PGO spike, in which case the rapid eye movements may reflect attempts to scan the images (Ladd, 1892; Roffwarg, Dement, & Muzio, 1962), or conversely, the mages may be produced by the oculomotor impulses in response to direct stimulation from the gigantocellular pontine reticular field (Hobson & McCarley, 1977; McCarley & Hobson, 1979). 

REM is thought to be generated in the lateral portions of the nucleus reticularis pontis oralis (RPO) immediately ventral to the locus ceruleus in the pontine reticular formation. The neurotransmitters in this region have not been clearly determined, but are neither cholinergic nor monoaminergic. The RPO receives projections from cholinergic regions in the laterodorsal tegmental nucleus (LDT) and the pedunculopontine tegmental nucleus (PPT) as well as from ventromedial portions of the medulla. The RPO, LDT, and PPT are collectively thought to be part of the REM-on neural population (Steriade & McCarley, 1990). These populations are hypthesized to interact with REM-off noradrenergic neurons in the locus ceruleus and seratonergic neurons in the raphe system. These latter populations are most active during waking and least active during REM. Interactions between the REM-on and REM-off populations are thought to control REM onset and offset (Steriade & McCarley, 1990). 
SP may reflect an anomaly of the functioning of the monoaminergic systems and/or their inhibition of the REM-on cholinergic system. Experimental and clinical dissociations have been demonstrated among major components of REM:  namely, PGO activity, atonia, and EEG desynchronization (Hishikawa & Shimizu, 1994). Hishikawa & Shimizu speculate that SP may be produced by hyperactivation of the Sleep-on populations or, they deem more likely, hypoactivation of the Sleep-off populations. That SP may be alleviated by serotonin and adrenergic reuptake inhibitors is taken to be consistent with this hypothesis. Also involved may be suprapontine systems involving the reticular system, including the hippocampus and amygdala. 

REM SP with HHEs differs from REM dreams in that during SP there is little or no blocking of exteroceptive stimulation and there is no loss of waking consciousness. SP with HHEs differs from dream experience in that the sensory cortex may be receiving both externally and internally generated information. The peculiarity of the HHEs in SP may, in part, be a result of the brain's attempts to integrate endogenous cortical arousal originating in the pons with normal sensory input. A similar peculiarity may exist for motor pattern arousal during SP. McCarley and Hobson argue that, during dream generation by internal stimulation of motor programs, we interpret the activity of the pattern generators and their corollary discharge as movement. The lack of peripheral feedback, though not normally necessary for effective control, may contribute to a sense of unreality to the apparent movement and hence to the "bizarreness" of dreams. Pontine activation of motor patterns during SP appears to be less common in SP than in dreams, if subjective reports of illusory movement are to be taken as evidence. Volitional attempts at movement during SP are common, however, and the absence of feedback is most often, though not always, experienced as paralysis rather than illusory movement. Thus it appears that, during SP, the frontal cortex is more sensitive to the absence of feedback than during dreaming. When motor programs are spontaneously activated during SP these might be extremely resistant to coherent interpretation am may be experienced as very unusual bodily states. In concluding sections we will relate more specifically the phenomenology of various HHEs to the underlying neurophysiology.