Capturing individual differences in cognitive abilities is central to human neuroscience. Yet our ability to estimate cognitive abilities via brain MRI is still poor in both prediction and reliability. Our study tested if this inability was partly due to the over-reliance on 1) non-task MRI modalities and 2) single modalities. We directly compared predictive models comprising of different sets of MRI modalities (e.g., task vs. non-task). Using the Human Connectome Project (n=873 humans, 473 females, after exclusions), we integrated task-based functional MRI (tfMRI) across seven tasks along with other non-task MRI modalities (structural MRI, resting-state functional connectivity) via a machine-learning, stacking approach. The model integrating all modalities provided unprecedented prediction (r=.581) and excellent test-retest reliability (ICC>.75) in capturing general cognitive abilities. Importantly, comparing to the model integrating among non-task modalities (r=.367), integrating tfMRI across tasks led to significantly higher prediction (r=.544) while still providing excellent test-retest reliability (ICC>.75). The model integrating tfMRI across tasks was driven by areas in the frontoparietal network and by tasks that are cognition-related (working-memory, relational processing, and language). This result is consistent with the parieto-frontal integration theory of intelligence. Accordingly, our results sharply contradict the recently popular notion that tfMRI is not appropriate for capturing individual differences in cognition. Instead, our study suggests that tfMRI, when used appropriately (i.e., by drawing information across the whole brain and across tasks and by integrating with other modalities), provides predictive and reliable sources of information for individual differences in cognitive abilities, more so than non-task modalities.
Test–retest reliability of a functional MRI anticipatory anxiety paradigm in healthy volunteers
