Noninvasive dissection of mouse sleep using a piezoelectric motion sensor

Area of Science:

• Neuroscience
• Sleep Science
• Biomedical Engineering

Background:

• Autonomic control changes influence breathing patterns between REM and NREM sleep.
• Piezoelectric sensors detect respiration and movement, previously used to differentiate sleep and wake states in mice.

Purpose of the Study:

• To develop and validate a noninvasive piezoelectric sensor-based classifier for distinguishing REM and NREM sleep states in mice.

Main Methods:

• Manual scoring of 24-h EEG/EMG recordings in 20 mice.
• An unsupervised classifier analyzed piezoelectric signal features (movement, respiration) to identify three states: Wake, NREM, and REM.
• Comparison of classifier predictions with manual EEG/EMG scores.

Main Results:

• Unsupervised classification achieved high accuracy for Wake (89% sensitivity, 96% specificity) and moderate for REM (73% sensitivity, 75% specificity), but poor for NREM (51% sensitivity, 96% specificity).
• The classifier sometimes confused light NREM sleep with REM due to irregular breathing.
• A supervised classifier improved sensitivities for Wake, NREM, and REM to 90%, 81%, and 67%, respectively, with specificities over 90%.

Conclusions:

• The piezoelectric sensor method effectively differentiates sleep states beyond just wake vs. sleep.
• This approach can aid large-scale genetic screening, drug studies, and other sleep manipulation research.