Figure 1.

Recording setup and detecting NREM sleep. A, Sleeping rat along with location of recording locations in the M1 and CB. B, Example of sleep classification in both PSD (top) and temporal spaces. Each dot represents the PSD in γ (30–60 Hz) and δ (0.1–4 Hz) frequency bands during 6 s windows. Epochs were classified into two clusters using k-means clustering: awake/REM (blue) and NREM (red). The average LFP trace is plotted for an example sleep block at the bottom. C, Example of a detected M1 spindle, highlighting the automatic method used for identification. The broadband LFP (top) is filtered in the spindle band and the spindle band components (bottom) are extracted. The spindle band envelope (black line; bottom) must have exceeded an upper threshold (solid black line; bottom) for one sample and a lower threshold (dashed black line; bottom) for at least 500 ms. Transparent shaded gray box represents a detected spindle. D, Same as C showing cerebellar spindle detection. E, Average M1 spindle-triggered waveform (black line) and spectrogram (heat map). F, Average CB spindle-triggered waveform (black line) and spectrogram (heat map).