This work is supported by NIH EY12925 and was presented as a poster at the 2002 meeting of the Society for Neuroscience.
Download our poster here [pdf].
We are also exploring this mechanism of attention with behavioral experiments. For example, we have recently found that attending to a given feature (such as color or direction of motion) in one location in visual space facilitates performance on the same feature for stimuli in other spatial locations. These psychophysical results have been published in Vision Research [pdf] .
More recently, we used functional MRI in humans to evaluate the role of competing stimuli in non-spatial, feature-based attention. Here, we repeated our feature-based attention motion experiment without the overlapping stimulus to assess the role of distractors. Two stimuli, one attended and one ignored, were presented to the left and right of fixation. As in our original experiment, we again found that responses in multiple cortical visual areas (V1, V2, V3, V4, V3A, MT+) to the unchanging, ignored stimulus increased when subjects attended the matching direction of motion in the attended stimulus. However, the magnitude of the effect was smaller in the experiments without distractors. Thus, the presence of distracting features within an attended field increases the strength of the global feature-based attention effect. This work will be presented at the 2003 Vision Sciences Society. Download our poster here [pdf].
This work is now published in Neuron (pdf)
Much progress has been recently made in understanding the relationship between perceptual decisions and responses of single sensory neurons (e.g. Parker and Newsome, 1998), but it remains unclear how perceptual decisions depend on the way information is represented across a population of neurons. We studied the relationship between population responses, as measured by fMRI, and visual discrimination thresholds using a simple model of neuronal population responses. FMRI responses and psychophysical discrimination thresholds were acquired from subjects viewing moving sinusoidal grating stimuli and performing either a speed or a contrast discrimination task at various baseline contrast levels. FMRI responses were found to increase monotonically with contrast in area V1, but saturated at low contrasts in area MT+. Perceptually, increases in baseline contrast lead to an increase in contrast discrimination thresholds, but to a decrease in speed discrimination thresholds. Our neuronal population model explains this relationship between the fMRI and psychophysical data by exploiting the different neural representations for contrast and speed in different visual areas. Specifically, our modeling results show that speed discrimination thresholds are most consistent with responses in area MT+ and contrast discrimination thresholds are most consistent with responses in area V1. Our study demonstrates the critical significance of modeling the underlying population code in understanding the relationship between fMRI responses in visual cortical areas and visual perception.
This work is now under review.The interpretation of rapid event-related fMRI experiments, in which hemodynamic responses to events overlap in time, generally assumes a linear relationship between the fMRI response and the underlying neuronal events (Boynton et al., 1996). However, it has been shown that when two identical visual stimuli are presented closely in time, the estimated BOLD response to the second stimulus is significantly smaller than the response to a single stimulus alone (Dale and Buckner, 1997; Huettel and McCarthy, 2000). This apparent nonlinearity could be caused by either nonlinearities in neurovascular coupling, or it could simply be the result of neuronal adaptation. To test for neuronal adaptation, we measured event-related BOLD responses in primary visual cortex to pairs of contrast modulated grating stimuli in which the orientation of the stimulus pairs were either the same or orthogonal. If the apparent nonlinearity described above is caused by neuronal adaptation, then we might not expect to find this nonlinearity in the orthogonal pair condition because the two stimuli should excite two relatively separate subpopulations of neurons across the orientation tuned cells in primary visual cortex.
As expected from previous results, we found a smaller response to the second stimulus in the same-orientation pair condition. However, contradicting the neuronal adaptation hypothesis, we also found a similarly small response to the second stimulus in the orthogonal pair condition. The amount of suppression of the response to the second stimulus recovered as a function of stimulus separation in a similar manner for both the same-orientation and the orthogonal pair conditions. These results suggest that the nonlinearity in the hemodynamic response measured with paired stimuli is the result of saturation in the neurovascular coupling process. This has implications for the analysis of rapid event-related fMRI experiments. These results also raise questions about event-related fMRI experiments that presumably use neuronal adaptation as a tool for studying the properties neuronal subpopulations.
This work will be presented as a poster at the 2003 conference on Human Brain Mapping, and is supported by the Dana Foundation.We are currently exploring the neural basis of a specific sub-type of synesthesia (grapheme-color synesthesia) using fMRI. These subjects associate colors with graphemes (numbers and letters) and are interesting not only because of their novelty, but also because understanding synesthesia may lead to a better understanding about neuronal connections in the normal brain. Our main fMRI experiment involves an experimental condition designed to localize brain areas that are preferentially active to grapheme stimuli. This experiment is a variation of an earlier study by Pesenti et al. (2000) which compared brain responses to grapheme symbols to nonsense symbols. This PET study found activity in the fusiform gyrus in control subjects (which is an area also associated with object and face recognition). Our hypothesis is that like the control subjects, synesthetic subjects should show activity to grapheme stimuli in the fusiform area, but unlike control subjects, responses to grapheme stimuli should be also found in earlier retinotopic areas associated with color processing.
Preliminary results indicate that synesthetic subjects may possess unusually strong neuronal connections between the fusiform gyrus and V1. These connections may send signals either cortically from the fusiform area back to V1, or down through the lateral geniculate nucleus (LGN) and back up into V1. We are currently recruiting further control subjects and synesthetes to confirm this effect.
This work will be presented at the 2003 Vision Sciences Society.The luminance and color of surfaces in natural scenes are relatively independent under certain linear transformations, with the luminance of a surface providing little information about the color of that surface, and vice versa. However, differences in luminance between two locations in a natural scene remain strongly associated with differences in color. We used the statistics of the spatio-chromatic structure of natural scenes as the priors for a Bayesian model that decides whether or not two points within an image fall on the same surface. This model provides a biologically plausible algorithm for surface segmentation that models observer segmentations well.
This work is now published in the Journal of the Optical Society of America, A (pdf)"Suppose a man born blind, and now adult, and taught by his touch to distinguish between a cube and a sphere Suppose then the cube and the sphere placed on the table, and the blind man made to see Query: whether by his sight, before he touched them, he could distinguish and tell which is the globe, which is the cube?" Despite the philosophical and psychological interest of Molyneaux's question, cases of adult sight restoration are so rare that even now little is known about perceptual experience after long-term visual deprivation. To address this question, we used psychophysics and functional magnetic resonance imaging to characterize visual processing in a subject who had been blind from the age of 3 to 43. We found several consequences of long-term visual deprivation, including a shift in the tuning of neurons towards very low spatial frequencies, impairments in form processing, object agnosia, and prosopagnosia. Using fMRI we demonstrated that these deficiencies were consequent upon neural changes in visual striate and extrastriate cortex. In contrast to these difficulties with form perception, motion processing was relatively undisturbed by deprivation. Consistent with this dissociation, cortical areas responsible for motion processing (MT complex) showed stronger and more organized fMRI activation than form processing areas (V1-V4).
This work is currently in review.Using functional magnetic resonance imaging (fMRI), visual stimuli have been shown to activate regions of auditory cortex in deaf but not in hearing subjects (Finney, et al, Nature Neurosci, 2001) [pdf]. Thus auditory deprivation leads to cross-modal plasticity, whereby deaf subjects recruit auditory cortex for the processing of visual stimuli. Early deafness in humans may also lead to compensatory plasticity in remaining intact modalities (Bavelier, et al, J. Neurosci, 2000 [pdf], 2001 [pdf]).
To test this further, we are applying fMRI to compare the size and responsiveness of different areas of visual cortex, including V1, V2, V3, V4 and the MT/MST complex (MT+), in deaf and hearing subjects. Using standard functional techniques to define retinotopic areas and MT+, we found no significant differences in the size of retinotopic areas or MT+ between groups. Overall activation and the effect of visual attention was measured using a lateralized motion stimulus. Moving dots were presented within a 10-degree aperture in either the right or left peripheral visual field. In separate blocks, subjects performed a dimming task on either the moving dots (attend-motion) or on a stationary fixation square (ignore-motion). Contralateral responses were larger in the attend-motion condition than in the ignore-motion condition, while there was little activation and no effect of attention in the ipsilateral visual field. There were no differences, however, in overall activation or in attentional effects between deaf and hearing subjects within occipital cortex or MT+. These results suggest only limited compensatory plasticity within the visual cortex of the deaf.
This work was presented at the November 2002 meeting of the Society for Neurocience.
|
| WebMaster | Salk Information |