Generating Synthetic Task-based Brain Fingerprints for Population Neuroscience Using Deep Learning

Abstract

Task-based functional magnetic resonance imaging (tb-fMRI) provides valuable insights into individual differences in the neural basis of cognitive functions because it links specific cognitive tasks to their evoked neural responses. Yet, it is challenging to scale to population-level data due to its cognitive demands, variations in task design across studies, and a limited number of tasks acquired in typical large-scale studies. Here, we present DeepTaskGen, a convolutional neural network (CNN) approach that enables us to generate synthetic task-based contrast maps from resting-state fMRI (rs-fMRI) data. Our method outperforms several benchmarks, exhibiting superior reconstruction performance while retaining inter-individual variation essential for biomarker development. We showcase DeepTaskGen by generating synthetic task images from the UK Biobank cohort, achieving competitive or greater performance compared to actual task contrast maps and resting-state connectomes for predicting a wide range of demographic, cognitive, and clinical variables. This approach will facilitate the study of individual differences and the generation of task-related biomarkers by enabling the generation of arbitrary functional cognitive tasks from readily available rs-fMRI data.