DREAMGATE

Sleep is crucial for our global cognitive functioning, but its exact functions and mechanisms are still poorly understood. Cognitive studies of sleep typically rely on linking electrophysiological changes measured during sleep with behavioral and neural changes collected in tasks performed during wakefulness. What concomitantly happens in the mind of sleeping subjects is often ignored, certainly because it is virtually inaccessible. Yet, major advances in the understanding of human behaviors have resulted from an integrated approach that combines both neural and cognitive measures of their ongoing mental processes. The goal of my proposal is to provide real-time measures of the cognitive processes occurring within sleep. To prompt real-time access to the sleeping mind, I will use auditory stimulation in people with unique sleep peculiarities: lucid dreamers, who may provide first-person accounts of dreaming experiences during REM sleep (WP1), and sleepwalkers whose overt behaviours may enable to objectively visualize ongoing cognitive processes during non-REM (NREM) sleep (WP2). In WP1, I aim to provide the building blocks for establishing an online dialogue between an experimenter and a dreamer: I will test whether lucid dreamers are able to hear auditory inputs from the outside, understand and discriminate them, and appropriately respond with measurable outputs. Dreamers usually take their dreams as if they were a veridical perception, but lucid dreamers have the rare ability to be aware that they are dreaming while in the midst of REM sleep, and signal it to an experimenter. So far, research has been limited to one-way communication, with no online intervention of the experimenter. In WP1, we will test the possibility for two-way communication during dreaming. We will ask lucid dreamers with narcolepsy to perform a lexical decision task during five daytime naps. They will have to contract different facial muscles (zygomatic and corrugators) in response to verbal stimuli (either a word or a pseudo-word) while monitored with polysomnography. We hypothesize that lucid dreamers will be able to hear and understand auditory inputs, and then answer appropriately with facial muscular contractions while staying in a dreaming state. In WP2, I aim to offer a quantifiable, real-time measure of memory processing during NREM sleep by using memory-related cues to influence which behaviours sleepwalkers will exhibit. Because sleepwalking is characterized by complex behaviors emerging from NREM sleep, it could provide an observable window into mental activity in NREM sleep. While neuronal memory reactivations have largely been demonstrated in sleeping rodents, evidence for the existence of such reactivations in humans is at best indirect. In WP2, we will use targeted memory reactivation (TMR), a tool that allows to control which individual memories are reactivated on a trial by trial basis. TMR consists in associating sensory cues with a specific learning, then re-applying these cues during subsequent sleep to trigger the reactivation of the corresponding learning. Sleepwalkers will be trained on a modified version of the serial reaction time task: they will perform as fast and accurately as possible a sequence of gestures in response to auditory cues. Our goals are to: i) show that playing the auditory cues during NREM sleep triggers a behavioral replay of the learned ‘choreography’ in sleepwalkers, ii) quantify, with a gesture recognition algorithm, how the sleep gestures differ from the wake ones (speed, accuracy, sequence fidelity), and iii) test whether this evoked replay is accompanied by a congruent dream. My project proposes to overcome traditional difficulties stalling dream research, -namely the lack of access and experimental control over dreams-, and may open a new era of investigations into sleep cognition.