A simple and efficient algorithm operating with linear time for MCEEG data compression.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Computational Neuroscience

Background:

• Increasing use of electroencephalograph (EEG) signals necessitates data compression for lower power and storage.
• Existing compression methods face challenges in latency, hardware dependency, anomaly detection, and signal reconstruction.

Purpose of the Study:

• To propose a novel, computationally simple spatial pseudo codec (SPC) for multichannel EEG (MCEEG) compression.
• To achieve lossy to near lossless compression with efficient data handling.

Main Methods:

• MCEEG signals are normalized, then processed in parallel for integer and fractional parts.
• Spatial domain encoding is used for the integer part, while the fractional part is coded as pseudo integers.

Main Results:

• The SPC achieved an average compression ratio (CR) of 3.16 with an average PRD of 2.72.
• The algorithm exhibits O(n) complexity, with encoding/decoding times of 0.3 ms and 0.04 ms per sample.
• Performance is comparable to fDCT and tensor decomposition methods.

Conclusions:

• The proposed SPC is a feasible compression scheme for MCEEG recording, archiving, and BCI systems.
• Offers a balance of compression efficiency, signal recovery, and computational simplicity.