Font Size: a A A

Analogous neural coding of shape in vision and touch

Posted on:2010-10-03Degree:Ph.DType:Thesis
University:The Johns Hopkins UniversityCandidate:Yau, Jeffrey MFull Text:PDF
GTID:2448390002478853Subject:Biology
Abstract/Summary:
We are remarkably adept at perceiving objects in our environment. We can appreciate object shape through vision alone (from a distance) and through touch alone (in the dark). Presumably, the visual and somatosensory systems share information to allow for consistent recognition across conditions. Common form processing mechanisms would faciliate crossmodal communications.;In this thesis we tested the hypothesis that vision and touch share common coding formats in representing 2-dimensional shapes. We recorded responses from isolated neurons in primary and secondary somatosensory cortex to straight, curved, and angled contour fragments indented into the distal finger pad an awaking, behaving monkey. Our first goal was to demonstrate that somatosensory neurons at early and intermediate levels of the form processing pathway represent contour fragments like their visual counterparts. For area 3b and area 1 neurons, we characterized response patterns using a simple spatial model and we found these neurons to extract orientation information, much like neurons in primary visual cortex (Hubel and Wiesel, 1968). We quantified response patterns in area 2 and SII and compared these to neural responses in visual area V4 to matched visual stimuli. Individual visual and somatosensory neurons at intermediate levels of processing exhibited clear timing for curvature direction. Response patterns characterizing curvature tuning explained large amounts of response variance in each neural population. The responses of many visual and somatosensory neurons were explained by a curvature model based on linear/nonlinear combinations of orientation-tuning basis functions. This model explicitly related curvature tuning to orientation signals extracted in primary sensory cortex. By fitting temporal weighting functions for this model, we showed that the information represented by V4 neurons transformed over time, transitioning from early selectivity for individual component orientations to later tuning for orientation configurations (curvature).;Overall, the results of this thesis are consistent with parallel shape coding mechanisms in the visual and somatosensory systems. Isomorphic representations at the receptor level are transformed into orientation signals in primary sensory cortex. At intermediate levels, these representations are dynamically integrated and transformed into larger, more complex shape constructs. Parallel processing of shape at early and intermediate stages of the visual and somatosensory perceptual pathways imply the existence of common coding schemes in higher-order areas.
Keywords/Search Tags:Shape, Coding, Visual, Vision, Neural, Neurons, Intermediate, Area
Related items