Keywords
Normalization, Induction, Surround Suppression, Contextual Integration, Extra-classical effects
Abstract
The perception of a stimulus is largely determined by its surrounding. Examples abound from color (Land and McCann, 1971), disparity (Westheimer, 1986) and motion induction (Anstis and Casco, 2006) to orientation tilt effects (O’Toole and Wenderoth, 1976). Some of these phenomena have been studied individually using monkey neurophysiology techniques. In these experiments, a center stimulus is typically used to probe a cell’s classical “center” receptive field (cRF), whose activity is then modulated by an annular “surround” (extra-cRF) stimulus. While this center-surround integration (CSI) has been well characterized, a theoretical framework which unifies these different phenomena across visual modalities is lacking. Here, we present an extension of a popular cortical inhibition circuit, divisive normalization (Carandini and Heeger, 2011), which yields a computational model that is consistent with experimental data across visual modalities. We have found that a common characteristic of CSI across modalities is a shift in neural population responses induced by surround activity. Typical implementations of the divisive normalization model rely on gain control mechanisms from an ‘untuned’ suppressive pool of cells; that is, the identity of that pool is the same for every cell being suppressed. As such, the circuit cannot account for the selective shift in population response curves observed in contextual effects. In contrast, we show that the addition of an extra-classical suppressive ‘tuned’ pool of cells which selectively inhibits different parts of a population response curve is sufficient to explain complex shifts in population tuning responses. Overall, our results suggest that a normalization circuit based on two forms of inhibition, gain control and selective suppression, captures shifts in population responses associated with CSI and yields a model that seems consistent with contextual phenomena across visual modalities.
Start Date
14-5-2015 3:15 PM
End Date
14-5-2015 3:40 PM
Session Number
04
Session Title
Theory
Towards a Unified Computational Model of Contextual Interactions across Visual Modalities
The perception of a stimulus is largely determined by its surrounding. Examples abound from color (Land and McCann, 1971), disparity (Westheimer, 1986) and motion induction (Anstis and Casco, 2006) to orientation tilt effects (O’Toole and Wenderoth, 1976). Some of these phenomena have been studied individually using monkey neurophysiology techniques. In these experiments, a center stimulus is typically used to probe a cell’s classical “center” receptive field (cRF), whose activity is then modulated by an annular “surround” (extra-cRF) stimulus. While this center-surround integration (CSI) has been well characterized, a theoretical framework which unifies these different phenomena across visual modalities is lacking. Here, we present an extension of a popular cortical inhibition circuit, divisive normalization (Carandini and Heeger, 2011), which yields a computational model that is consistent with experimental data across visual modalities. We have found that a common characteristic of CSI across modalities is a shift in neural population responses induced by surround activity. Typical implementations of the divisive normalization model rely on gain control mechanisms from an ‘untuned’ suppressive pool of cells; that is, the identity of that pool is the same for every cell being suppressed. As such, the circuit cannot account for the selective shift in population response curves observed in contextual effects. In contrast, we show that the addition of an extra-classical suppressive ‘tuned’ pool of cells which selectively inhibits different parts of a population response curve is sufficient to explain complex shifts in population tuning responses. Overall, our results suggest that a normalization circuit based on two forms of inhibition, gain control and selective suppression, captures shifts in population responses associated with CSI and yields a model that seems consistent with contextual phenomena across visual modalities.