How do humans learn from raw sensory experience? Throughout life, but most obviously in infancy, we learn without explicit instruction. We propose a detailed biological mechanism for the widely embraced idea that learning is driven by the differences between predictions and actual outcomes (i.e., predictive error-driven learning). Specifically, numerous weak projections into the pulvinar nucleus of the thalamus generate top–down predictions, and sparse driver inputs from lower areas supply the actual outcome, originating in Layer 5 intrinsic bursting neurons. Thus, the outcome representation is only briefly activated, roughly every 100 msec (i.e., 10 Hz, alpha), resulting in a temporal difference error signal, which drives local synaptic changes throughout the neocortex. This results in a biologically plausible form of error backpropagation learning. We implemented these mechanisms in a large-scale model of the visual system and found that the simulated inferotemporal pathway learns to systematically categorize 3-D objects according to invariant shape properties, based solely on predictive learning from raw visual inputs. These categories match human judgments on the same stimuli and are consistent with neural representations in inferotemporal cortex in primates.

No abstract

Recent work has shown that abstract, non-spatial relationships between entities or task states are organized into representations called cognitive maps. Here we investigated how cognitive control enables flexible top-down selection of goal-relevant information from multidimensional cognitive maps retrieved from memory. We examined the relationship between cognitive control and representational geometry by conducting parallel analyses of fMRI data and recurrent neural network (RNN) models trained to perform the same task. We found both 2D map-like representations in a medial temporal lobe and orbitofrontal cortical network and simultaneous 1D orthogonal representations of relevant task dimensions in a frontoparietal network, supporting representational stability and flexibility, respectively. These representational motifs also emerged with distinct temporal profiles over the course of training in the RNN. We further show that increasing control demands due to incongruence (conflicting responses) between current task-relevant and irrelevant dimensions produce warping along the context-invariant axis in subjective representations, and the degree of warping further accounts for individual differences in cognitive control. Together, our findings show how complementary representational geometries balance generalization and behavioral flexibility, and reveal an intricate bidirectional relationship between cognitive control and cognitive map geometry.

No abstract