Speaker: JONATHAN CANT, Psychology Department, University of Toronto at Scarborough
Title: The Role of Ensemble Processing in Human Visual Cognition
Ebbinghaus Empire Series 2018-2019
Video Recording of Talk
Abstract: Ensemble, set, or summary-statistical processing collectively refer to the visual system’s ability to compress repeating or redundant visual information into a single meaningful metric, such as the average size of a group of objects, or the average emotional expression of a crowd of faces. This is done at the expense of recovering fine detail about any individual item from the set. Given that ensemble processing operates efficiently on large groups of objects, it can be seen as an adaptive mechanism, as it can circumvent the capacity limitation inherent in visual working memory and attention. In this talk I will present data from my laboratory which aims to better understand ensemble processing at multiple levels of visual representation (i.e., from lower-level geometric shapes to mid-level familiar objects and higher-level faces), using a number of different experimental approaches. First, using behavioural psychophysical methods, I will present data investigating whether single-object and ensemble processing are mediated by distinct cognitive mechanisms; how visual features, such as shape and surface properties, contribute to ensemble representations of multiple objects; and, whether the simultaneous extraction of different statistical moments leads to interference in ensemble processing. Next, using functional magnetic resonance imaging (fMRI), I will explore which regions of the brain are especially sensitive to processing ensemble information, what types of visual features are represented in these regions, and how outliers affect the representation of object ensembles across various regions of occipital, temporal, and parietal cortex. Finally, using electroencephalography (EEG), I will present results examining the temporal dynamics of ensemble face processing, using a combination of traditional univariate and more advanced multivariate statistical techniques. Collectively, the results of these studies reveal different processing mechanisms in the visual system for object individuation versus summary-based statistical encoding, and point towards an adaptive role for ensemble processing in human visual cognition. Namely, it can be seen as a resource that has a wider bandwidth than traditional models of visual working memory, and as such can facilitate attentional selection and contribute to aspects of visual processing that require fast and efficient statistical extraction of features, such as texture and scene perception.
For further information please contact Johnny Dubois at firstname.lastname@example.org.
If you require an accommodation due to a disability, please contact the event coordinator OR email email@example.com five days prior to the event. We will work with you to make appropriate arrangements.