Multichannel EEG compression using predictive coding on DCT-transformed signals

Electroencephalography (EEG) signals generate high-dimensional data, creating challenges in storage and transmission. This paper presents an efficient EEG compression method combining Discrete Cosine Transform (DCT), Predictive Coding, Max Absolute Scaling, and Huffman Coding to enhance compression while maintaining signal integrity. The proposed approach first applies DCT to transform EEG signals into a compact frequency-domain representation, concentrating energy into fewer coefficients. Predictive Coding is then employed across DCT-transformed signals from different EEG electrodes, exploiting spatial correlations to further reduce data redundancy. To optimize encoding efficiency, Max Absolute Scaling refines floating-point precision, preparing the data for Huffman Coding, which ensures efficient entropy compression. Experimental results demonstrate that the method achieves an 81% compression ratio under full-pass conditions and up to 90% under low-pass conditions, while maintaining a high Peak Signal-to-Noise Ratio (PSNR) of 41.61 dB. The approach is validated using experimental EEG recordings, proving its effectiveness in real-time processing, AI-driven analysis, and large-scale healthcare monitoring. By efficiently compressing multichannel EEG data, the method reduces storage and transmission costs, making it highly suitable for resource-constrained environments.