Please use this identifier to cite or link to this item: https://doi.org/10.1016/j.neuroimage.2020.117310
Title: The individualized prediction of cognitive test scores in mild cognitive impairment using structural and functional connectivity features
Authors: Yu, J.
Rawtaer, I.
Fam, J. 
Feng, L. 
Kua, E.-H. 
Mahendran, R. 
Keywords: Connectome-based predictive modeling
Executive functions
Functional connectivity
Individualized prediction
Memory
Mild cognitive impairment
Structural connectivity
Issue Date: 2020
Publisher: Academic Press Inc.
Citation: Yu, J., Rawtaer, I., Fam, J., Feng, L., Kua, E.-H., Mahendran, R. (2020). The individualized prediction of cognitive test scores in mild cognitive impairment using structural and functional connectivity features. NeuroImage 223 : 117310. ScholarBank@NUS Repository. https://doi.org/10.1016/j.neuroimage.2020.117310
Rights: Attribution-NonCommercial-NoDerivatives 4.0 International
Abstract: Neuropsychological assessments are essential in diagnosing age-related neurocognitive disorders. However, they are lengthy in duration and can be unreliable at times. To this end, we explored a modified connectome-based predictive modeling approach to estimating individualized scores from multiple cognitive domains using structural connectivity (SC) and functional connectivity (FC) features. Multi-shell HARDI and resting-state functional magnetic resonance imaging scans, and scores from 10 cognitive measures were acquired from 91 older adults with mild cognitive impairment. SC and FC matrices were derived from these scans and, in various combinations, entered into models along with demographic covariates to predict cognitive scores. Leave-one-out cross-validation was performed. Predictive accuracy was assessed via the correlation between predicted and observed scores (rpredicted-observed). Across all cognitive measures, significant rpredicted-observed (0.402 to 0.654) were observed from the best-predicting models. Six of these models consisted of multimodal features. For three cognitive measures, their best-predicting models’ rpredicted-observed were similar to that of a model that included only demographic covariates— suggesting that SC and/or FC features did not contribute significantly on top of demographics. Cross-prediction models revealed that the best-predicting models were similarly accurate in predicting scores of related cognitive measures— suggesting their limited specificity in predicting cognitive scores. Generally, multimodal connectomes together with demographics, can be exploited as sensitive markers, though with limited specificity, to predict cognitive performance across a spectrum in multiple cognitive domains. In certain situations, it may not be worthwhile to acquire neuroimaging data, considering that demographics alone can be similarly accurate in predicting cognitive scores. © 2020
Source Title: NeuroImage
URI: https://scholarbank.nus.edu.sg/handle/10635/199330
ISSN: 10538119
DOI: 10.1016/j.neuroimage.2020.117310
Rights: Attribution-NonCommercial-NoDerivatives 4.0 International
Appears in Collections:Staff Publications
Elements

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
10_1016_j_neuroimage_2020_117310.pdf3.74 MBAdobe PDF

OPEN

NoneView/Download

Google ScholarTM

Check

Altmetric


This item is licensed under a Creative Commons License Creative Commons