Quantifying the Similarity of Multiple Multi-Dimensional Point Processes by Integer Programming with Application to Early Diagnosis of Alzheimer’s Disease from EEG

A novel approach is proposed to quantify the similarity (or “synchrony”) of multiple multi-dimensional point processes. It is based on a generative stochastic model that describes how two or more point processes are related to each other. Inference in that model is carried out by integer programming. As an application, the problem of diagnosing Alzheimer’s disease (AD) from multi-channel EEG recordings is considered. The proposed method seems to be more sensitive to AD induced perturbations in EEG synchrony than classical similarity measures.