Neural dynamics of an extended frontal lobe network in goal-subgoal problem solving

In goal-directed behavior, cognitive control calls for representation of current state, goal state, and current move, often with hierarchical organization into a series of steps or subgoals. In the human brain, a network of "multiple-demand" or MD regions underpins cognitive control. Here we recorded from four putative homologs to human MD regions in the frontal cortex - ventrolateral (vlPFC), dorsomedial (dmPFC), dorsal premotor (dPM) and anterior insula/orbitofrontal (I/O) cortex - as two male monkeys solved a multi-step, on-screen spatial maze. Across regions there was overlap but also wide quantitative variation in encoding fundamental task features. Sensory input and current state were strongly coded in vlPFC, goal most stably in dmPFC, and move most rapidly in vlPFC and dPM. I/O responded during revision of a prepared route. In all regions, an abstract code of problem structure marked progress from problem start to end, with hierarchical separation of progress within and between successive steps. Partial specialisations, we suggest, arise from distinct connectivity, while broad co-recruitment reflects strong information exchange. Activity across this network of partial frontal lobe specialists encodes the core features of a controlling task model - current state, goal state, move and hierarchy.