Emotional and Neutral Sounds for Neurophysiological Prognostic Assessment of Critically Ill Patients With a Disorder of Consciousness

The evaluation of the neurological outcome of intensive care unit (ICU) patients with a disorder of consciousness (DOC) is a major medical, ethical and economic issue. These DOC are essentially related to a direct anoxo-ischaemic (post-cardiac arrest), traumatic or even vascular (caused by a hemorrhagic or ischemic vascular accident) cerebral aggression. The techniques currently available, whether neurophysiological (electroencephalogram (EEG) and evoked potentials (EP)), neuroradiological or biological, only allow an approximate evaluation for a large number of aetiologies and patients (Obadi. EEG and EPs have the advantage of being feasible at the patient's bedside, with a precise spatial-temporal resolution of the cerebral capacities to integrate sensory stimulation. If some neurophysiological tests have an imperfect predictive capacity, event-related potentials, (ERPs) with "oddball paradigm" seem to be a promising method. During their realizations by exposing the subject to listening to a deviant and rare auditory stimulus within other frequent stimuli, a first neurophysiological response is generated 150 to 200ms after the stimulation called "mismatch negativity" (MMN), then a second response to 300ms of stimulation called "P3a" is generated. The latter would reflect the orientation of a subject's attention towards the deviant stimulus and could predict arousal. Some recent data report that a P3 response obtained by exposing the subject to a stimulus with expressive and emotional value, such as the patient's own first name, could improve the prognostic value of this neurophysiological tool (Fischer et al, Holeckova et al). Indeed, the neural processing of expressive voices involves a greater number of subcortical and cortical regions than neutral sounds (Schirmer and Kotz). Moreover, some data suggest that the use of a "subject own name" (SON) auditory stimulus pronounced by a familiar voice (FV) compared to an unfamiliar voice (NFV) could improve the prognostic value of P3 or even the use binaural sounds with a three-dimensional effect as "looming" or "receding" sounds, these hypotheses having never been evaluated in DOC patients. The investigators hypothesize that cortical and subcortical activation is more complex and intense in response to emotional than to neutral sounds, and that obtaining a P3a response generated by sounds expressive type SON pronounced by a familiar voice (FV) would have a prognostic value greater than the P3 response induced by the SON with an unfamiliar voice for wakefulness prediction of DOC patients; The investigators will also test the hypothesis that the prognostic value of the MMN response generated by sounds with randomly varied motion in their 3D auditory field (e.g. looming or receding sources) is higher than those generated by neutral sounds.