Premium
P4‐309: MULTIMODAL‐CNN: IMPROVED ACCURACY OF MRI‐BASED CLASSIFICATION OF ALZHEIMER'S DISEASE BY INCORPORATING CLINICAL DATA IN DEEP LEARNING
Author(s) -
Jo Taeho,
Nho Kwangsik,
Risacher Shan L.,
Yan Jingwen,
Saykin Andrew J.
Publication year - 2018
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2018.07.132
Subject(s) - convolutional neural network , deep learning , artificial intelligence , computer science , pipeline (software) , pattern recognition (psychology) , machine learning , programming language
negatively associated to global cognition (lateral temporal, fusiform and occipital) but not to episodic memory, while FDG showed strong and widespread positive associations with both global cognition and episodic memory. Hippocampal FDG mediated the effect of AV1451 on global cognition in the AD group. In the Abneg-MCI group, FDG mediated the AV1451 effect on both global cognition and episodic memory in parahippocampal, lateral temporal and frontal regions. Results for other cognitive domains will be also presented. Conclusions: The direct and FDG-mediated effects of AV1451 on cognition had sharply contrasting patterns across diagnostic groups, suggesting that tau pathology may lead to cognitive deficits via different mechanisms depending on presence or absence of Ab. Multimodal PET studies may contribute to understanding pathophysiological mechanisms by which cerebral tau deposits lead to cognitive dysfunction in neurodegenerative disorders.