Determination of the sleep-wake pattern and feasibility of NREM/REM discrimination using the non-invasive piezoelectric system in rats.

The piezoelectric cage-floor sensors have been used to successfully dissect sleep patterns in mice based on signal features related to respiration and body movements. We studied performance of the piezoelectric system to quantify the sleep-wake pattern in the rat over 7 days of recording compared with a visual electroencephalogram/electromyogram scoring, and under two light/dark (LD12:12 and LD16:8) photoperiods leading to change in the 24-hr sleep characteristics (N = 7 per group). The total sleep time (%/24 hr) over the 7 days recording and hourly sleep time over the last 24-hr recording were not statistically different between methods under the two photoperiods. Both methods detected higher total sleep time with the LD16:8 photoperiod compared with LD12:12 (p < .05), and correlated significantly (p < .001) at light and dark periods during each photoperiod. The accuracies for discrimination of sleep-wake patterns between methods were 81.9% and 84.9% for LD12:12 and LD16:8, respectively. In addition, spectral analysis of the respiratory signal given by piezo during all 10-s periods of the corresponding non-rapid eye movement and rapid eye movement sleep periods recorded by electroencephalogram/electromyogram resulted in selection of 36 features that could be inserted in an automated non-rapid eye movement sleep and rapid eye movement sleep classification, with 90% accuracy with the electroencephalogram/electromyogram visual scoring. The piezo system proved to be a reliable non-invasive alternative to electroencephalogram recording to study total sleep time in rat, with feasibility to discriminate between non-rapid eye movement and rapid eye movement sleep stages. This will be interesting in pharmacological or bio-behavioural studies evaluating sleep patterns or the restorative functions of sleep in the body and the brain.