Sleep paralysis (SP) involves temporary immobility at sleep onset or offset, while individuals are awake and aware of their surroundings (American Academy of Sleep Medicine 2001; Hishikawa Reference Hishikawa, Guilleminault, Dement and Passouant1976). SP arises from anomalous intrusion of REM-related motor inhibition and hallucinoid imagery into waking consciousness (Hishikawa & Shimizu Reference Hishikawa and Shimizu1995; Takeuchi et al. Reference Takeuchi, Miyasita, Sasaki, Inugami and Fukuda1992). Across diverse samples and measures, SP hallucinations reliably fit three factors (Cheyne & Girard Reference Cheyne and Girard2004; Reference Cheyne and Girard2007): An ominous felt presence forms a core intruder experience, accompanied by multisensory hallucinations. Incubus experiences instantiate the intruder as perched on the chest, suffocating the experiencer or committing physical or sexual assault. Vestibular-motor experiences comprise illusory movement, vestibular sensations, and more blissful out-of-body phenomena. Despite the history reviewed by Llewellyn that supports contributions of memory to dreaming, SP hallucinations are generally modeled as epiphenomena of REM-induced activation of amygdala and emotional brain, sensory, and associational posterior brain regions, with reduced prefrontal monitoring (Cheyne & Girard Reference Cheyne and Girard2007; Reference Cheyne and Girard2009). Llewellyn offers an interesting perspective that SP hallucinations might also reflect elaborative encoding.

SP experiences appear to be consistent with the mnemonic principles applied by Llewellyn to REM dreaming: dramatic/bizarre association, visual complexity, imagination, emotional arousal, narrative, embodiment, organization, movement, and spatial association. Nonetheless, note that SP hallucinations and dreams diverge in some ways. Although dream reports and SP hallucinations are primarily visual and auditory, the latter involve more substantial tactile, physiological, and vestibular-motor experiences. Whereas dream imagery is entirely endogenous, unconstrained by sensory input, SP hallucinations present a unique interface between perception, imagination, and expression of internal representations (Girard et al. Reference Girard, Martius and Cheyne2007). Half of SP experiencers maintain their ability to open their eyes and perceive their surroundings . Hallucinations overlaid on this environment are perceived with a vivid sense of external reality. SP hallucinations are also distinguished as more terrifying, vivid, complex, and multisensory than other forms of hypnagogic imagery (Cheyne et al. Reference Cheyne, Rueffer and Newby-Clark1999). The potential to test memory for aspects of the real environment and how these are integrated with memory for hallucinatory experiences offers an interesting prospect for investigation. It would also be interesting to code and compare SP narratives with reports of dreams and waking episodic memories.

Novel binding of episodic memories to form hyperassociative dream scenes depends on an underlying commonality. As in Llewellyn's example, fear provides such an overarching theme for dream narrative and is especially common and extreme during SP episodes (Sharpless et al. Reference Sharpless, McCarthy, Chambless, Milrod, Khalsa and Barber2010). Cheyne and Girard (Reference Cheyne and Girard2007) proposed that a threatening felt presence forms a core delusion from which SP hallucinations elaborate. Experiencers often rate the terror of intruder and incubus hallucinations “off the scale.” This fear is thought to arise from REM activation of the amygdala and extended threat-activated vigilance system in the context of waking consciousness while experiencing ominous hallucinations, helplessly paralyzed, typically supine in the dark. Cheyne and Girard (Reference Cheyne and Girard2007) suggest that REM initiation of this vigilance system also offers a model for thematic organization of dreams more generally. Incorporating Llewellyn's proposal, it might be worthwhile to consider the amygdala's role in emotional enhancement of recollection (Sharot et al. Reference Sharot, Delgado and Phelps2004) and that, via the hippocampus (Sharot et al. Reference Sharot, Verfaellie and Yonelinas2007), fear might serve an integrative mnemonic function.

Llewellyn proposes that encoding of recent memories is enhanced via hyperassociations with emotionally salient remote memories. Aspects of SP have been associated with a history of trauma (Abrams et al. Reference Abrams, Mulligan, Carleton and Asmundson2008) and, in at least one case, related to remote memory of childhood abuse (McNally & Clancy Reference McNally and Clancy2005). Such cases may provide insight into the nature of hyperassociative binding. Although core elements are consistent, incubus assaults take various culturally specific instantiations such as old-hag attacks, alien abductions, spirits, and demons (Cheyne et al. Reference Cheyne, Rueffer and Newby-Clark1999), sometimes with fatal consequences (Adler Reference Adler2011). Llewellyn differentiates nonconscious episodic hyperassociations bound during REM from conscious semantic associations formed in waking and NREM sleep. The unique mix of REM and waking in SP may evoke not only episodic associations, but integrate semantic representations. SP narratives might then reflect a product of elaborative encoding of prior experiences and instantiations of culturally relevant schema. It might be informative to explore the extent to which SP experiences are guided by cultural fables and, conversely, whether elaborative encoding during SP shapes memories that influence cultural accounts.

What individual and contextual factors give rise to the threatening intruder versus blissful vestibular-motor experiences? How do personal encounters and navigational experience relate to dreaming and SP associations involving persons and places? Llewellyn reminds that self-identity relies on coherent autobiographical memory. Whereas dreams involve first-person perspective from an embodied, agential self, vestibular-motor hallucinations include out-of-body experiences. Perhaps elaborative encoding of movement-related memories in the form of typically nonconscious hyperassociations during anomalous waking paralysis challenges the continuity of body and “self.” Spatial attributes of intruder and vestibular-motor hallucinations have been linked to intrinsic biases in sensory and motor function (Girard & Cheyne Reference Girard and Cheyne2004; Girard et al. Reference Girard, Martius and Cheyne2007). These findings suggest that place associations might not relate solely to hyperassociations with remembered places, but depend in part on individual differences in sensorimotor functions.

Given the conscious state during SP episodes and ability of individuals to recall their experiences vividly, future research should probe relations between the waking “night-mare” and episodic memories. Elaborative encoding predicts that cuing this associative network should enhance retrieval of linked memories. On the other hand, disruption of normal differentiation between waking and stages of sleep can impair memory and cognition (Terzaghi et al. Reference Terzaghi, Ratti, Manni and Manni2012). Although SP is considered common among cognitively intact individuals, relations between SP and cognitive abilities lack systematic study. Whereas Llewellyn highlights the hippocampus and binding of recent and remote memories, the cognitive domains involved are also implicated in more extended brain networks supporting not only memory of the past but also episodic simulation of imagined future events (Addis et al. Reference Addis, Pan, Vu, Laiser and Schacter2009). Thus, processes underlying elaborative encoding might function not only to strengthen episodic memory, but also to create novel imaginative scenarios for adaptive activation and honing of neural systems, such as those involved in threat detection (Boyer & Bergstrom Reference Boyer and Bergstrom2011).