FASTER: an unsupervised fully automated sleep staging method for mice.

Area of Science:

• Neuroscience
• Computational Biology
• Sleep Science

Background:

• Sleep staging is crucial for animal research but is bottlenecked by slow, subjective, and error-prone human analysis of EEG/EMG data.
• Automating sleep staging is essential for high-throughput and objective large-scale animal sleep studies.

Purpose of the Study:

• To develop an unsupervised, automated method for accurate and efficient sleep staging in mice.
• To overcome the limitations of manual sleep scoring in research settings.

Main Methods:

• Developed Fully Automated Sleep sTaging method via EEG/EMG Recordings (FASTER).
• Employed nonparametric density estimation clustering on comprehensive EEG/EMG power spectra.
• Validated the method on mice with drug-induced and genetically modified sleep patterns.

Main Results:

• FASTER achieved over 90% accuracy in all tested mouse groups.
• The automated method processed 24-hour data in 10 minutes on a laptop, significantly faster than human raters.
• FASTER demonstrated high accuracy even with perturbed sleep patterns.

Conclusions:

• FASTER provides a less subjective, high-throughput solution for animal sleep staging.
• This automated method dramatically improves the quality and throughput of sleep staging.
• FASTER is poised to advance quantitative and comprehensive animal sleep research.