Robust associations between white matter microstructure and general intelligence

Improved estimation of general cognitive ability and its neural correlates with a large battery of cognitive tasks

Elucidating the neural mechanisms of general cognitive ability (GCA) is an important mission of cognitive neuroscience. Recent large-sample cohort studies measured GCA through multiple cognitive tasks and explored its neural basis, but they did not investigate how task number, factor models, and neural data type affect the estimation of GCA and its neural correlates. To address these issues, we tested 1,605 Chinese young adults with 19 cognitive tasks and Raven's Advanced Progressive Matrices (RAPM) and collected resting state and n-back task fMRI data from a subsample of 683 individuals. Results showed that GCA could be reliably estimated by multiple tasks. Increasing task number enhances both reliability and validity of GCA estimates and reliably strengthens their correlations with brain data. The Spearman model and hierarchical bifactor model yield similar GCA estimates. The bifactor model has better model fit and stronger correlation with RAPM but explains less variance and shows weaker correlations with brain data than does the Spearman model. Notably, the n-back task-based functional connectivity patterns outperform resting-state fMRI in predicting GCA. These results suggest that GCA derived from a multitude of cognitive tasks serves as a valid measure of general intelligence and that its neural correlates could be better characterized by task fMRI than resting-state fMRI data.

Investigating robust associations between functional connectivity based on graph theory and general intelligence

Previous research investigating relations between general intelligence and graph-theoretical properties of the brain's intrinsic functional network has yielded contradictory results. A promising approach to tackle such mixed findings is multi-center analysis. For this study, we analyzed data from four independent data sets (total N > 2000) to identify robust associations amongst samples between g factor scores and global as well as node-specific graph metrics. On the global level, g showed no significant associations with global efficiency or small-world propensity in any sample, but significant positive associations with global clustering coefficient in two samples. On the node-specific level, elastic-net regressions for nodal efficiency and local clustering yielded no brain areas that exhibited consistent associations amongst data sets. Using the areas identified via elastic-net regression in one sample to predict g in other samples was not successful for local clustering and only led to one significant, one-way prediction across data sets for nodal efficiency. Thus, using conventional graph theoretical measures based on resting-state imaging did not result in replicable associations between functional connectivity and general intelligence.