back to

Seminar: 2/4 - Job Talk: Sensation and Perception Position Psych/HWNI

4:00 to 5:30 PM      at:  

John Serences
Department of Cognitive Sciences and Center for Cognitive Neuroscience, University of California, Irvine
"Population codes, Attention, and Perceptual Coherence"

Sensory neurons are inherently noisy, so repeated presentations of a stimulus never evoke the same pattern of activity twice. To overcome this obstacle, theoretical studies suggest that perception is based on small populations of neurons that pool their output, since averaging reduces noise. However, when multiple objects are simultaneously present in the scene, selective attention is also required to ensure that representations of behaviorally relevant stimuli dominate awareness. In my talk I will discuss a series of fMRI experiments that examine how top-down attentional factors – such as behavioral goals and reward history – shape population response profiles to support coherent perceptual representations. In the first study, an analysis of activation patterns within visual cortex reveals that feature-specific attentional modulations spread to empty regions of the visual field. These spatially global modulations may increase sensitivity to behaviorally relevant features even before a stimulus is presented. Next, we use high-resolution imaging and a novel analysis method to estimate orientation-selective ‘tuning functions’ based on activation levels within small subregions of primary visual cortex (V1). Using these tuning functions as a proxy for feature-selective neural activity, our data suggest that attention increases the gain of neurons tuned to attended features while simultaneously attenuating the gain of neurons tuned to unattended features. Finally, we show that the reward history of a stimulus biases neural activity across the visual hierarchy (including in V1), suggesting that value influences the way in which we represent and perceive stimuli in the surrounding environment. Collectively, our methodological approach provides a new set of analytical tools to characterize response modulations within sensory cortices, and our empirical results reveal several ways in which attention shapes population response profiles to support stable perceptual representations.