Medically relevant criteria used in EEG compression for improved post-compression seizure detection

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Informatics

Background:

• Biomedical signals are crucial for diagnosing disorders.
• Lossy compression of these signals risks losing medically relevant information.
• Accurate compression is vital for maintaining diagnostic reliability.

Purpose of the Study:

• To analyze three lossy electroencephalogram (EEG) compression schemes.
• To evaluate compression methods based on biophysical signal models.
• To assess the impact of compression on the reliability of clinical features, specifically for seizure detection.

Main Methods:

• Investigated three distinct lossy EEG compression schemes.
• Employed signal models with varying reliance on EEG production and physiological characteristics.
• Evaluated the performance of seizure detection algorithms after signal compression.

Main Results:

• Compression schemes based on biophysical signal models demonstrated better preservation of clinically significant EEG features.
• The degree of reliance on physiological characteristics influenced compression effectiveness.
• Seizure detection accuracy was maintained effectively by compression schemes incorporating biophysical signal properties.

Conclusions:

• Biophysically-inspired signal models are effective for lossy compression of biomedical signals like EEG.
• These models enhance the reliability of compressed data for diagnostic applications, such as seizure detection.
• Future research should explore advanced biophysical models for optimized biomedical signal compression.