Research

Information Processing in the Visual System

I try to understand how the visual system in the mammalian brain processes information about the image presented to the eyes. In particular, I am interested in the processing of texture borders, borders in the image that are not defined by a difference in luminance or color, but by a difference of the structure of a texture. Analyzing these kind of stimuli is a highly complex task that the brain can solve, but we don't understand exactly how.

While the processing of 'simple' stimuli that are defined by changes in luminance (i.e. 'gratings') can be analyzed by applying linear filters to the image, the processing of texture borders requires some kind of non-linear processing. Just the output of simple filters is not enough, but these outputs have to combined in a complex way to receive a signal that tells the brain: 'there is a texture border in this image at this location and of this orientation'.

In my current project I try to understand this nonlinear processing better by presenting a checkerboard stimulus with each square filled by a luminance grating. If two neighboring squares are filled with the same type of grating, no texture border is perceived, but if the gratings differ, there is a perceivable border present. For the assignment of what kind of grating is displayed in which square, I use the m-sequence technique which produces pseudorandom binary sequences. This allows me to easily compute the correlation of the input signal with the response of single neurons in the visual cortex and measure non-linear responses in both space and time.