Research Interests

My research interests concern the characterization of brain mechanisms involved in human visual information processing. I am interested in understanding the specific neural computations involved in transforming visual input into the rich visual representations we effortlessly experience. To this end I utilize psychophysical measures aimed at quantifying performance in visual perception, memory, and attention tasks coupled with the brain imaging techniques of Electroencephalography (EEG), Magnetoencephlography (MEG), and Functional Magnetic Resonance Imaging (fMRI). In addition, I am actively involved in methodological developments relating to the integration of frequency domain EEG (frequency tagging) and high-resolution anatomical MRI for the purpose of assessing the dynamic properties of cortical networks. In these projects I have been in the fortunate position to carry out my research in supportive and forward thinking environments and collaborations.

Current Research Projects:

• Neural Mechanisms of Texture Segmentation: SSVEP Investigations
  The ability of segment regions of the visual scene into coherent objects that are different from their surrounds is a fundamental process of visual perception. We utilize a "frequency tagging" approach to investigate the brain mechanisms underlying scene segmentation. Tagging involves the presentation of visual stimuli with periodic 'tags' (luminance flicker or similar modulation) applied to distinct regions of the visual stimuli. These frequency tags can easily be detected in the Fourier spectrum of EEG collected as subjects view our stimuli. When combined with high-resolution MRI, source localization procesures can provide a detailed picture of the brain dynamics of neural populations involved in processing the visual stimulus. Using this "frequency tagging" method, we have been able to characterize distinct cortical networks mediating the processing of figure and background regions of textured visual stimuli. In particular, a recurrent network of retinotopically defined ventral visual areas (V1, V2v, V3v, LOC) including the lateral occipital cortex has been identified that is selectively active in response to the figure (object) region of these displays. This network is invariant with respect to the defining texture cues used to define the region, as well as the spatial extent and driving frequency of the input stimuli. A separate network, extending from the primary visual cortex, through the dorsal visual pathway (V2d, V3d, MT+) is observed at frequencies corrosponding to harmonics of the background frequency. The identification of these networks fits well into the known framework of visual cortical specificity (ventral "what" and dorsal "where" streams) and provides a platform from which to evaluate explicit stimulus configural relationships and their role in figure-ground assignment.
  
  
• Visual Motion Mechanisms: MEG Investigation
  It has been demonstrated that the human visual system carries out several distinct motion computations. In particular, three statistical classes of motion mechanisms has been described psychophysically; first-order luminance defined motion, second-order motion derived from contrast or textures cues in the absance of luminance information, and third-order motion computed based on the relative salience of regions within a motion display. Each of these are subtle but differentiable processes that are carried out by distinct physiological mechanisms. The aim of this project is to localize the explicit brain nuclei mediating these speicific motion computations. To this end, we employ carefully constructed and calibrated direction reversing motion displays, that uniquely stimulate the individual motion systems. Through the activation and selective adaptation of these independent motion mechanisms we have been able to characterize an extrastriate dorsal system responsible for second-order motion perception. This system is largely spatially overlaping with the first-order system but exhibits a consistent temporal delay associated with an extra rectification stage postulated to be necessary for second-order signals to be interperable by first-order mechanisms.
  
  
• What Bistable Images Reveal About Figure-Ground Processing
  Bistable percepts, such as the classic Rubin Face and the Necker cube, are hypothesized to arise from uncertainty in the figure-ground relationship within an image. That is, the observer can interpret either of the two distinct regions of the image as figure, while the other is interpreted as background. Both figure-ground segmentations can be made with the information available in the display. By applying a luminance flicker, "Frequency Tag" to either of the figure or ground regions of the image we have identified correlates of figure-ground segregation in high resolution EEG recordings. We observed steady state responses at the flicker frequency which were consistently modulated with the reported percept ('Face' or 'Vase'), presumably reflecting changes in the figure-ground organization of the image. Click the Rubin Illusion icon below for a .jpg of our recent Neuroscience 2003 poster.
  
  
  
  
• The Psychophysical Theory of Amplification
  The theory of amplifications predicts that the visibility of one grating is determined, in part, by the contrast of contiguous grating. The degree of amplification is a function of the relative contrast of the two graitng and their spatial frequencies. Using a texture-slant discrimination procedure we quantifed the amout of amplification over a range of spatial and temporal conditions. In effect, we measure the amount in which one patch improves the visibility of adjacent patchs through fine contrast discriminations. This principle of amplification was further applied as a technique for calibration useful in obtaining pure visual displays free of unwanted contaminations which are in turn of use in a host of other psychophysical procedures. Click HERE for a .pdf of our texture amplification abstract. Click HERE for a .pdf of our texture amplification abstract.