Faculty Proceedings, Presentations, Posters, Speeches, Lectures, etc.


Local sensitivity to stimulus orientation and spatial frequency within the receptive fields of neurons in visual area 2 of macaque monkeys



Conference Title

Society of Neuroscience Annual Meeting


Society of Neuroscience


San Diego, California / November 13-17, 2010

Publication Date / Copyright Date



A substantial percentage of neurons in monkey visual area 2 (V2) have been reported to encode combinations of non-collinear orientations and hypothesized to be involved in detecting 'corners' and 'angles' in texture patterns. These previous studies employed spatially restricted grating stimuli at various locations within their receptive-fields (RFs) or a finite set of complex stimuli in order to reveal local sensitivity to stimulus orientation for a given unit. To avoid potential limitations of this approach, we employed a new method to reveal the local sensitivity (subfields) of a unit to stimulus orientations and spatial frequencies without many assumptions (local spectral reverse correlation or LSRC, Nishimoto et al, 2006). To examine the spatial homogeneity of RFs, we isolated activity from 137 V2 neurons in anesthetized macaque monkeys and stimulated each unit with a dense dynamic two-dimensional noise array that covered an area 3 times larger than the cell’s classic RFs. After obtaining sufficient numbers of spikes (median = 16136 spikes; range = 1574 - 77271 spikes), we performed an LSRC analysis by cross-correlating the spike train and the frequency spectra of Gaussian-windowed local stimuli. From the local spectral analysis, we determined the preferred orientation and spatial frequency of each subfield. By shifting the center position of the Gaussian window we obtained the spatial matrix of local selectivity for each neurons. We found that the preferred orientations or spatial frequencies of subfields for a given unit rarely differed from each other and very few units exhibited subfields maps of preferred orientation that might be sensitive to angles or corners in complex contour patterns. There were strong correlations between the average optimal orientation of subfields in a given unit and its preferred orientation determined with sine wave gratings. Although a sampling bias may not be ruled out, the observed high homogeneity of subfield orientations and spatial frequencies in V2 neurons suggests that the detection of angles and corners in complex contour stimuli may require the 'specialized' RF internal organization of individual V2 neurons that was not revealed in this study (e.g., suppressive tuning, Willmore et al, 2010).



This document is currently not available here.

Peer Reviewed