If you have any problems related to the accessibility of any content (or if you want to request that a specific publication be accessible), please contact us at firstname.lastname@example.org.
Spatiotemporal Color Representations: High-density EEG and Psychophysics
AuthorVanston, John Erik
AdvisorCrognale, Michael A.
AltmetricsView Usage Statistics
A fundamental aim in neuroscience is to understand the way that information is represented and processed in the brain. Studying the visual system has been a fruitful approach as principles of neural network function and computational strategies discovered first in the visual system have been shown to generalize to other brain regions. Color perception is an important part of our visual experience of the world, and guides behavior in many ways. Studying how the visual system encodes spectral information to produce color vision has yielded valuable insights into the brain’s structure and function. However, much remains unknown about how the brain encodes chromatic information, and how these neural representations are related to perception and behavior. Psychophysics and electroencephalography (EEG) have been used to show that the visual pathways carrying color information have distinct spatiotemporal characteristics. These techniques complement each other, as each reveals different aspects of visual function. Psychophysics allows for the quantification of human perception, but provides indirect inference about the underlying brain activity. EEG directly measures the summed electrical potentials generated by neural activity. In addition to high temporal resolution, EEG can be leveraged to provide a topographic representation of activity across the scalp. Multivariate pattern analysis (MVPA) can be applied to this spatiotemporal data to reveal the information contained in patterns of activity across cortex. Pattern analysis of EEG color responses would be invaluable in understanding how visual information is represented across time and space in the brain. Furthermore, a comparison of these EEG responses with chromatic discrimination (a fundamental color-based behavior) would provide a crucial link between brain activity and behavior. In the current study, high-density EEG was used to measure chromatic visual evoked potentials (cVEPs) to stimuli presented in either onset-offset or steady-state temporal modes. cVEPs were measured for test stimuli that modulated along eight chromatic directions, and for comparison stimuli that contained no chromatic contrast. A pattern classification algorithm was used to determine whether test stimuli could be distinguished from comparison stimuli based on a variety of spatiotemporal parameters, including waveforms and frequency spectra. The ability of the algorithm to accurately classify stimuli was compared to psychophysical chromatic detection and discrimination. When the time course of activity or frequency spectrum at a given set of electrodes was used as a criterion, the best performance was seen at occipital electrodes, with accuracy decreasing for more anterior sites. The effectiveness of the spatial distribution of activity was dependent upon the paradigm, with raw amplitude (derived from onset data) being a more effective cue for classification than response magnitude at the stimulus frequency (derived from steady-state data). Classification thresholds were highly variable, but tended to be several times larger than detection or discrimination thresholds, and were generally not well described by ellipses. The current study finds that spatiotemporal responses evoked by chromatic stimuli can, in many cases, be distinguished from those evoked by achromatic stimuli. However, the discrimination thresholds for the pattern classification algorithm used here were, on the whole, higher than human behavioral thresholds for the same task.