A distributed brain network predicts general intelligence from resting-state human neuroimaging data

Abstract

Individual people differ in their ability to reason, solve problems, think abstractly, plan and learn. A reliable measure of this general ability, also known as intelligence, can be derived from scores across a diverse set of cognitive tasks. There is great interest in understanding the neural underpinnings of individual differences in intelligence, since it is the single best predictor of long-term life success, and since individual differences in a similar broad ability are found across animal species. The most replicated neural correlate of human intelligence to date is total brain volume. However, this coarse morphometric correlate gives no insights into mechanisms; it says little about function. Here we ask whether measurements of the activity of the resting brain (resting-state fMRI) might also carry information about intelligence. We used the final release of the Young Adult Human Connectome Project dataset (N=884 subjects after exclusions), providing a full hour of resting-state fMRI per subject; controlled for gender, age, and brain volume; and derived a reliable estimate of general intelligence from scores on multiple cognitive tasks. Using a cross-validated predictive framework, we predicted 20% of the variance in general intelligence in the sampled population from their resting-state fMRI data. Interestingly, no single anatomical structure or network was responsible or necessary for this prediction, which instead relied on redundant information distributed across the brain.peer-reviewed)

The Bayesian polyvertex score (PVS-B): a whole-brain phenotypic prediction framework for neuroimaging studies

“…Using simulations and empirical data, we have shown that incorporating effect sizes across all vertices, using the PVS method, explained more variance in behavior than thresholding vertices based on significance level. Consistent with association patterns reported in other task-based fMRI studies (Chang et al, 2015;Gonzalez-Castillo et al, 2012) and for other brain phenotypes (Dubois et al, 2018;Palmer et al, in prep;Smith et al, 2015), the explanatory effect of these functional brain phenotypes on the behaviors tested was widespread and distributed across the cortex. Our results provide an empirical evidence that it is important to account for the whole brain pattern of association when studying brain-behavior relationships.…”

Intrinsic neural circuitry of depression in adolescent females

“…Furthermore, future work should examine the generalizability of these neural network models to depression in adolescent males. In addition, recent evidence has suggested that task‐based connectivity models outperform resting‐state connectivity models in predicting fluid intelligence (Greene, Gao, Scheinost, & Constable, ), and resting‐state connectivity models were associated with higher‐order latent‐factors stronger than lower‐order cognitive task scores (Dubois, Galdi, Paul, & Adolphs, ). This indicates that engaging networks in targeted domains can improve the predictability of individual differences.…”

Task-Based Functional Connectomes Predict Cognitive Phenotypes Across Psychiatric Disease

“…A similar strategy has been successfully employed to create a phenotypic measure of intelligence across individual measures of crystallized ability, processing speed, visuospatial ability and memory (Dubois et al, 2018). To maintain separate train and test groups, for each iteration, each PCA was limited to training datasets and these PCA coefficients were subsequently applied to the test dataset (see Figure 1.C).…”

“…We performed a series of predictive and explanatory analyses of the brain's functional connectome . Functional connectomes have been shown to be unique to an individual (Finn et al, 2015), stable over a lifespan (Horien, Shen, Scheinost, & Constable, 2019), and predictive of clinical and cognitive traits in novel subjects (Dubois, Galdi, Paul, & Adolphs, 2018;Rosenberg et al, 2016). Though connectomes are typically based on resting-state fMRI, task-based fMRI has been shown to improve the prediction of individual cognitive traits and more clearly delineate brain-behavior relationships (Greene, Gao, Scheinost, & Constable, 2018).…”