Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published December 15, 2021 | Published
Journal Article Open

A deep learning based approach identifies regions more relevant than resting-state networks to the prediction of general intelligence from resting-state fMRI

Abstract

Prediction of cognitive ability latent factors such as general intelligence from neuroimaging has elucidated questions pertaining to their neural origins. However, predicting general intelligence from functional connectivity limit hypotheses to that specific domain, being agnostic to time-distributed features and dynamics. We used an ensemble of recurrent neural networks to circumvent this limitation, bypassing feature extraction, to predict general intelligence from resting-state functional magnetic resonance imaging regional signals of a large sample (n = 873) of Human Connectome Project adult subjects. Ablating common resting-state networks (RSNs) and measuring degradation in performance, we show that model reliance can be mostly explained by network size. Using our approach based on the temporal variance of saliencies, that is, gradients of outputs with regards to inputs, we identify a candidate set of networks that more reliably affect performance in the prediction of general intelligence than similarly sized RSNs. Our approach allows us to further test the effect of local alterations on data and the expected changes in derived metrics such as functional connectivity and instantaneous innovations.

Additional Information

© 2021 The Authors. Human Brain Mapping published by Wiley Periodicals LLC. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Issue Online: 17 November 2021; Version of Record online: 29 September 2021; Manuscript accepted: 27 August 2021; Manuscript revised: 26 August 2021; Manuscript received: 28 June 2021. This study was financed by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES)—Finance Code 001 and FAPESP (The São Paulo Research Foundation; grants 2017/02752-0 and 2018/11881-1). Data were provided by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University. Authors also thank James Townsend, and Lyndon White, Peifan Wu, and Julia AutoDiff community for helpful comments on automatic differentiation of the singular value decomposition. The authors declare no potential conflict of interest. Data Availability Statement: All data, imaging, demographical or behavioral, used is provided by the Human Connectome Project main study, HCP Young Adult. Data and details can be obtained at their site, pending approval. See https://www.humanconnectome.org/. Neural networks were implemented in Flux v0.9.0. See https://github.com/FluxML/Flux.jl/tree/v0.9.0. Additional code for analyses was implemented in Julia v1.3.0. See https://github.com/JuliaLang/julia/tree/v1.3.0. Code will be shared upon request.

Attached Files

Published - hbm.25656.pdf

Files

hbm.25656.pdf
Files (3.7 MB)
Name Size Download all
md5:e7191d3c640cae6dc8510d5f5437ca63
3.7 MB Preview Download

Additional details

Created:
August 22, 2023
Modified:
October 23, 2023