Fractionating the human frontoparietal cortex: combining meta-analytic and real-time optimization approaches (360G-Wellcome-209139_Z_17_Z)

Disruptions in the same set of frontal and parietal brain regions are seen across a striking range of psychiatric and neurological conditions. This network of regions has been referred to as multiple-demand (MD) system and can be divided into at least two closely coupled subnetworks. However, despite extensive research efforts, the specific functional mechanism each subnetwork supports remains poorly understood using available neuroimaging technology. To overcome these limitations, I have recently developed a novel technique based on real-time neuroimaging and machine learning: Neuroadaptive Bayesian Optimization (NaBO). The key goal of this fellowship is to develop a complementary approach that leverages the strength of large-scale, automated meta-analyses and NaBO to obtain a fine-grained functional mapping between MD subnetworks and the cognitive processes they support. This approach will exploit the wealth of data generated by neuroimaging to date (meta-analysis) for defining a prior model of how cognitive functions relate to MD subnetworks and then refine this model in unprecedented detail (NaBO). The resulting model will be validated using behavioural assessment. Advancing our understanding of these subnetworks in normal brain function is an important first step for developing targeted clinical interventions and informing the design of sensitive diagnostic test batteries.