Neural Networks, IEEE - INNS - ENNS International Joint Conference on (2000)
July 24, 2000 to July 27, 2000
Kazushi Murakoshi , Tokyo Institute of Technology
Kiyohiko Nakamura , Tokyo Institute of Technology
This research presents a biological realistic model of the primary visual area to account for how line drawings are identical by synchrony and asynchrony of coincident spikes of neural populations, which originate from the neural network with strong circular and weak mutual excitatory connections. This model assumes that in a column 1) a simple cell and a end-stopped cell of the same orientation selectivity have mutual excitatory connections, 2) all end-stopped cells have mutual excitatory connections, and 3) simple cells of the different orientation selectivity have inhibitory connections via an inhibitory neuron. These assumptions led to the results that within some hundreds of milliseconds time range the cross-correlograms of two line segments in the cross and T-junction stimuli showed asynchrony while those in the right angle stimulus showed synchrony. The results of synchrony and asynchrony indicate early local segmentations of line drawings in the psychophysically realistic time scale.
K. Nakamura and K. Murakoshi, "Synchrony and Asynchrony of Coincident Spikes of Neural Populations Depending on Line Drawings," Neural Networks, IEEE - INNS - ENNS International Joint Conference on(IJCNN), Como, Italy, 2000, pp. 5442.