Generalizable predictive modeling of semantic processing ability from functional brain connectivity

Semantic processing (SP) is one of the critical abilities of humans for representing and manipulating meaningful and conceptual information. Neuroimaging studies of SP typically collapse data from many subjects, but both its neural organization and behavioral performance vary between individuals. It is not yet understood whether and how the individual variabilities in neural organizations contribute to the individual differences in SP behaviors. Here we aim to identify the neural signatures underlying SP variabilities by analyzing individual functional connectivity (FC) patterns based on a large-sample Human Connectome Project (HCP) dataset and rigorous predictive modeling. We used a two-stage predictive modeling approach to build an internally cross-validated model and to test the models generalizability with unseen data from different HCP sub-populations and task states as well as other out-of-sample datasets that are independent of the HCP. FC patterns within a putative semantic brain network were significantly predictive of individual SP scores summarized from five semantic tasks. This cross-validated predictive model can be used to predict unseen HCP data. The model generalizability was enhanced with FCs in language tasks than resting state and other task states and was better for females than males. The model constructed from the HCP dataset can be generalized to two independent cohorts that participated in different semantic tasks. FCs connecting to the Perisylvian language network show the most reliable contributions to predictive modeling and the out-of-sample generalization. These findings contribute to our understanding of the neural sources of individual differences in SP, which potentially lay the foundation for personalized education and improve intervention practice for patients with SP and language deficits.