Cortical mechanisms of learned spatial-temporal sequence coding

Boston University (Charles River Campus), Boston MA

Investigators

Linked publications & trials

Abstract

The brain continually processes streams of information that are coded at the neural level by temporal sequences of spiking activity. From this activity, the brain is able to extract behaviorally relevant data and form internal representations of the external world used, in turn, to create behavioral output. The brain also uses its internal state to make predictions about how external stimuli will change; these predictions play a critical role in executive behavioral planning. It is not known how this processing is accomplished in the brain. Establishing the relationship between activity sequences, plasticity and the neural coding of these internal representations will greatly inform our understanding of normal brain function and is necessary to understand the cognitive deficits associated with mental disorders. Since animals cannot self-report their cognitive state it is very difficult to explore the high-level neural mechanisms of sequence learning using animal models. Generally speaking, the neocortex is organized according to a single common plan that imparts a characteristic local architecture and there is evidence suggesting that brain regions acquire functional differentiation as a result of their specific inputs. In this framework, visual cortex is ?visual? primarily because it connects to the retina and all regions of cortex are capable of solving similar information processing problems. This suggests that the same basic mechanisms used to learn sequences in ?higher? cortical regions should exist within ?lower? regions as well and leads to the hypothesis that primary sensory areas should contain the mechanisms necessary to locally encode sequence representations. A series of experiments testing this hypothesis demonstrate that it is possible to entrain visual sequences in primary visual cortex with both temporal and spatial precision. This research aims to fully characterize and understand the mechanistic nature of this learning and its consequences for cortical processing. The proposed experiments are designed to test the hypothesis that visual sequence learning is encoded by NMDAR mediated synaptic plasticity between populations of neurons spread across the cortical layers locally within V1 using a combination of electrophysiological observation, 2- photon microscopy, pharmacological and optogenetic manipulation, and computational modeling.