P02-S Automatic analysis of sleep architecture and circadian rhythms for assessment of disorders of consciousness

Abstract

Disorders of consciousness (DOC) are among the major challenges of contemporary medicine. Currently, clinical diagnosis of DOC is based upon behavioural scales like Coma Recovery Scale-Revised, which involves visual assessment of patient’s reactions. However, lack of communication with DOC patients, their potential sensory impairments, and fluctuating levels of attention cause high rates of misdiagnoses in clinical assessment. Since DOC are related with severe alterations of both sleep-wake patterns as well as sleep architecture, parameters supplementing assessment can be derived from methods like actigraphy and polysomnography. However, these recordings in the case of DOC patients differ significantly from standards derived for healthy subjects, so standard methods of their analysis usually fail. To overcome the limitations of standard sleep scoring, we proposed a continuous analysis of sleep architecture in terms of EEG profiles, based upon automatic matching pursuit detection and parameterization of sleep transients like sleep spindles and slow waves. In this presentation we briefly recall the framework, software availability, and results of application of this approach to the analysis of overnight polysomnographic recordings of pediatric patients from “The Alarm Clock” model hospital for children with severe brain damage, combined with an automatic analysis of actigraphic recordings of these patients.