A New Graph Based Brain Connectivity Measure

• Published in: Advances in Computational Intelligence Vol. 11507; pp. 450 - 459
• Main Authors: Salazar, Addisson, Safont, Gonzalo, Vergara, Luis
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2019; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Brain connectivity; EEG; HMM; Neuropsychological tests; Signal processing on graphs
• ISBN: 3030205177; 9783030205171
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-20518-8_38

This paper presents a new measure of brain connectivity based on graphs. The method to estimate connectivity is derived from the set of transition matrices obtained by multichannel hidden Markov modeling and graph connectivity theory. Analysis of electroencephalographic signals from epileptic patients performing neuropsychological tests with visual stimuli was approached. Those tests were performed as clinical procedures to evaluate the learning and short-term memory capabilities of the patients. The proposed method was applied to classify the stages (stimulus display and subject response) of the Barcelona and the Wechsler Memory Scale - Figural Memory tests. To evaluate the capabilities of the proposed method, commonly used brain connectivity measures: correlation, partial correlation, and coherence were implemented for comparison. Results show the proposed method clearly outperforms the other ones in terms of classification accuracy and brain connectivity structures.