Conference on Cognitive Computational Neuroscience

The cholinergic system has been associated with learning, but also with cognitive decline in dementia, aging and injury. Yet, to date, no computational models have been put forward to explain how the cholinergic system contributes to both learning and cognitive decline. Here we introduce a model that combines a recently proposed model of cortex-wide hierarchical credit assignment with an adaptive cholinergic module inspired by deep learning rules. According to this model, the cholinergic system opens a cortex-wide gate for learning, but its end effect is controlled by local adaptive processes. We show that adaptive cholinergic modulation leads to rapid cortex-wide learning of sensory discrimination tasks when compared with non-adaptive models. This form of learning results in a constant redistribution of plasticity across the cortex making task-encoding sparser. Consequently, adaptive networks become more robust to perturbations. In addition, we demonstrate that global mechanisms are not sufficient for obtaining rapid and robust learning, suggesting the need for a tight cholinergic interaction with cortical circuits. Overall, our work provides a novel theoretical framework for cellular-systems neuroscience with which to link cholinergic cortical modulation to health and disease.