Automatic detection of rapid eye movements by discrete wavelet transform

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Accurate detection of rapid eye movements (REM) is crucial for sleep studies.
• Traditional methods for REM detection can be labor-intensive and subjective.
• Automated analysis of electrooculogram (EOG) data offers potential for objective and efficient sleep staging.

Purpose of the Study:

• To develop and validate an automated method for detecting REM using Discrete Wavelet Transform (DWT).
• To assess the efficacy of the Haar wavelet for REM detection in electrooculogram (EOG) signals.
• To evaluate the method's ability to simultaneously detect artifacts and its computational efficiency.

Main Methods:

• Electrooculogram (EOG) data from normal sleep recordings were analyzed.
• Discrete Wavelet Transform (DWT) with a Haar function was applied to 8-second segments.
• Phase shifting of the analyzing wavelet was employed to optimize REM detection.
• Simultaneous detection of body movement artifacts was incorporated.

Main Results:

• The DWT method achieved a high detection rate of 96% for REM.
• The Haar wavelet proved effective due to its similarity to the REM waveform.
• Body movement artifacts were concurrently identified by the same method.
• The computational time for analyzing 30 minutes of EOG data was only 11 seconds.

Conclusions:

• The developed DWT-based method provides an accurate and efficient approach for automated REM detection.
• This technique offers a significant improvement over traditional manual scoring methods.
• The simultaneous artifact detection capability enhances the robustness of sleep analysis.