Dementia-related disorders—including Alzheimer’s disease (AD), vascular dementia (VD), and mild cognitive impairment (MCI)—remain among the most challenging neurological conditions due to their overlapping clinical characteristics and complex neurocognitive manifestations. The difficulty in distinguishing these disorders continues to hinder therapeutic interventions, underscoring the need for analytical tools capable of revealing subtle but meaningful differences in brain network organization.
This book presents an in-depth exploration of brain functional network analysis as a means to address these challenges. Leveraging phase-based mutual information as the connectivity measure, we examine the functional architecture of the brain across healthy controls and individuals diagnosed with MCI, AD, and VD. Using a carefully selected reactive frequency band, we identify consistent patterns of hypoconnectivity across the dementia groups, providing insights into shared and divergent network alterations.
Two complementary computational frameworks form the core of this work: a frequent subgraph mining approach for the extraction of recurrent network motifs, and a minimum spanning tree (MST) method for structural characterization of both unweighted and weighted brain networks. The frequent subgraphs derived from dementia-specific networks are further applied for classification tasks, while MST-based features and additional graph-theoretic measures are used to delineate disease-specific network signatures.
- Arbeit zitieren
- Abdulyekeen T. Adebisi (Autor:in), Kalyana C. Veluvolu (Autor:in), 2026, EEG-Based Computational Neuroengineering for Dementia Diagnosis, München, GRIN Verlag, https://www.grin.com/document/1689170
