Compression-Enabled Joint Entropy Estimation for Seizure Detection on Human Intracortical Electroencephalography

We believe this clinical study analyzed the largest volume of continuous, multi-day intracortical neuroelectrophysiology for quantitative methods-with 2,900+ recording hours (420,000+ electrode-hours, amounting to 30+ TB of data). It is also the first demonstration of compression-based multidimensional estimation in a biological or clinical application. By computing an ensemble effect without linear assumptions or parametric modeling, ICR offers a model-free solution to the classically combinatorially intractable problem of high-dimensional joint entropy; its application likely extends beyond epilepsy to other domains of biomedical signal processing.