A Lateral Inhibitory Spiking Neural Network for Sparse Representation in Visual Cortex
Sparse representation has been validated to be a common phenomenon in many sensory neural systems, but its underlying neural mechanism still remains unclear. This paper proposes a neurally plausible model towards solving this problem. We find that lateral inhibition is the fundamental neural mechanism for sparse representation in the visual cortex, by which cortical neurons not only compete with each other so that the input signal can be represented sparsely but also cooperate with each other to make the representation more accurate. We integrate this result into the matching pursuit framework, a quite suitable solution for neural implementation, to illustrate how an input signal is sparsely represented in V1. Our simulation results show that lateral inhibition can evidently decrease the average squared error in the representation and then the input signal can be sparsely represented very well by the proposed algorithm.
KeywordsInput Signal Visual Cortex Basis Vector Lateral Inhibition Sparse Representation
Unable to display preview. Download preview PDF.
- 3.Barlow, H.: Possible principles underlying the transformation of sensory messages. Sensory Communication, 217–234 (1961)Google Scholar
- 5.Olshausen, B.A.: Principles of Image Representation in Visual Cortex. In: Chalupa, L.M., Werner, J.S. (eds.) The Visual Neurosciences, pp. 1603–1615. MIT Press (2003)Google Scholar
- 6.Ringach, D.L.: Spatial structure and symmetry of simple-cell receptive fields in macaque primary visual cortex. J. Neurophysiol. 88, 455–463 (2002)Google Scholar
- 22.Anderson, J.S., Carandini, M., Ferster, D.: Orientation Tuning of Input Conductance, Excitation, and Inhibition in Cat Primary Visual Cortex. J. Neurophysiol. 84, 909–926 (2000)Google Scholar