A NETWORK SIMULATION OF THALAMIC CIRCUIT OPERATIONS IN SELECTIVE ATTENTION

The ability of a thalamic circuit to process information selectively from a spatial location was investigated in neural network models. Starting with the known general structure of the thalamic circuit, we considered three variations of the projections from the inhibitory cells of the reticular nucleus onto the cells of the pulvinar nucleus of the dorsal thalamus. The three circuits were modeled as systems of difference equations, and their operations were simulated by computer-based numerical integration. In all three circuits, when input from a target location was slightly larger than the input from neighboring locations, the time evolution of principal (relay) cell outputs showed substantial selective enhancement at the target location compared with neighboring locations. The selective enhancement effect was produced not only on ascending inputs but also on descending cortical inputs. Simulations separating the lateral inhibitory and feedback-enhancement components of the circuits suggested that the feedback-enhancement component substantially magnified the ability of lateral inhibition to produce a target/surround difference.