Generative forecasting of brain activity enhances Alzheimer's classification and interpretation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Gao, Yutong, Calhoun, Vince D., Miller, Robyn L.
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913568458276864
author Gao, Yutong
Calhoun, Vince D.
Miller, Robyn L.
author_facet Gao, Yutong
Calhoun, Vince D.
Miller, Robyn L.
contents Understanding the relationship between cognition and intrinsic brain activity through purely data-driven approaches remains a significant challenge in neuroscience. Resting-state functional magnetic resonance imaging (rs-fMRI) offers a non-invasive method to monitor regional neural activity, providing a rich and complex spatiotemporal data structure. Deep learning has shown promise in capturing these intricate representations. However, the limited availability of large datasets, especially for disease-specific groups such as Alzheimer's Disease (AD), constrains the generalizability of deep learning models. In this study, we focus on multivariate time series forecasting of independent component networks derived from rs-fMRI as a form of data augmentation, using both a conventional LSTM-based model and the novel Transformer-based BrainLM model. We assess their utility in AD classification, demonstrating how generative forecasting enhances classification performance. Post-hoc interpretation of BrainLM reveals class-specific brain network sensitivities associated with AD.
format Preprint
id arxiv_https___arxiv_org_abs_2410_23515
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Generative forecasting of brain activity enhances Alzheimer's classification and interpretation
Gao, Yutong
Calhoun, Vince D.
Miller, Robyn L.
Machine Learning
Neurons and Cognition
Understanding the relationship between cognition and intrinsic brain activity through purely data-driven approaches remains a significant challenge in neuroscience. Resting-state functional magnetic resonance imaging (rs-fMRI) offers a non-invasive method to monitor regional neural activity, providing a rich and complex spatiotemporal data structure. Deep learning has shown promise in capturing these intricate representations. However, the limited availability of large datasets, especially for disease-specific groups such as Alzheimer's Disease (AD), constrains the generalizability of deep learning models. In this study, we focus on multivariate time series forecasting of independent component networks derived from rs-fMRI as a form of data augmentation, using both a conventional LSTM-based model and the novel Transformer-based BrainLM model. We assess their utility in AD classification, demonstrating how generative forecasting enhances classification performance. Post-hoc interpretation of BrainLM reveals class-specific brain network sensitivities associated with AD.
title Generative forecasting of brain activity enhances Alzheimer's classification and interpretation
topic Machine Learning
Neurons and Cognition
url https://arxiv.org/abs/2410.23515