Symbolic and Compression-Based Complexity in EEG: Benchmarking Frameworks and Thalamic Applications

This work explores the use of symbolic and compression-based complexity measures for analysis of scalp and deep brain EEG signals, with a focus on thalamic EEG. We review symbolizing strategies, entropy based measures (permutation entropy (PE), dispersion entropy(DispEn)) and compression-based metrics (Lempel-Ziv complexity (LZC) and Effort-to-Compress (ETC)) and their applicability to scalp EEG, stereo-electroencephalography (SEEG), and thalamic EEG are studied. Due to constraints in data accessibility, signal nonstationarity and variability in annotated data, adopting these methods for thalamic EEG analysis poses a significant challenge. To address this, a theoretical benchmarking facility is developed comprising of chaotic signal models, surrogate testing, and stochastic models. Further, the effects of preprocessing filtering, segmentation and transforms application are studied with respect to the symbolization process and metric analysis. By integrating concepts from nonlinear dynamical systems, we propose to develop a framework that goes beyond descriptive complexity measures and bridges the gaps between symbol encoding, phase space embedding, and deep brain dynamics. Our goal is to establish complexity analysis of symbolic EEG as a rigorous and clinically relevant framework for characterizing and interpreting brain dynamics, especially for deep brain structures.