Schedule of Events | Search Abstracts | Symposia | Invited Symposia | Poster Sessions | Data Blitz Sessions

Poster C109

Multi-Layer Extreme Learning Machine for Classification of Subjective Cognitive Decline and Neurodegenerative Disease Stages using White Matter Data

Poster Session C - Sunday, April 14, 2024, 5:00 – 7:00 pm EDT, Sheraton Hall ABC

Nishant Chauhan1, Hyun Woong Roh2, Sang Joon Son2, Chang Hyung Hong2, Dongha Lee1; 1Cognitive Science Research Group, Korea Brain Research Institute, Daegu, Republic of Korea, 2Department of Psychiatry, Ajou University School of Medicine, Suwon, Republic of Korea

Neurodegenerative disorders, characterized by the accumulation of brain proteins, can lead to cognitive impairment and disability, posing diagnostic and functional challenges. White matter (WM), housing vital neural circuits and glial cells, can sustain damage, resulting in clinical symptoms that affect daily life independence. Recent advances in imaging enhance accessibility to brain MRI and amyloid PET exams. Traditionally focusing on grey matter, this study underscores white matter's significance and potential in neurodegenerative disease diagnosis. This study utilized a Multi-Layer Extreme Learning Machine (MLELM), an advanced machine learning technique, to analyze data from 455 participants in the Biobank Innovations for Chronic Cerebrovascular Disease with Alzheimer’s Disease Study (BICWALZS), encompassing individuals with 49 subjective cognitive decline (SCD), 260 mild cognitive impairment (MCI), 103 alzheimer's disease (AD), and 43 vascular dementia (VD). For each participant, we constructed the WM populational connection label Map (pCLM) using the DARTEL toolbox and the international consortium for brain mapping (ICBM) template. This map was generated by overlaying T-2- FLAIR and amyloid PET data, which were co-registered to the T1-weighted MRI image. Subsequently, we created an ensemble dataset for analysis. The results of the pairwise comparisons using the MLELM revealed statistically significant differences between SCD and MCI, SCD and AD, as well as SCD and VD, showcasing the effectiveness of the model in distinguishing these conditions. While no statistically significant differences emerged between MCI and AD, MCI and VD, and AD and VD, the study demonstrated proficiency in accurately classifying individuals with SCD from other cognitive stages.

Topic Area: METHODS: Neuroimaging


CNS Account Login


April 13–16  |  2024