Purchase this article with an account.
J.Brendan Ritchie, Hans Op de Beeck; Using Neural Distance to Predict Reaction Time for Categorizing Animacy, Shape, and Abstract Properties. Journal of Vision 2018;18(10):1155. doi: 10.1167/18.10.1155.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Previous research has shown that distance to a decision boundary through neural activation space can be used to predict reaction times (RT) on animacy categorization tasks (Ritchie and Carlson, 2016). More specifically, it has been found that this relationship is driven by animate, but not inanimate exemplars. However it has yet to be explored how this relationship is impacted by target visual features, or whether the RT-distance relationship holds for other categorization tasks. Here we tested whether this asymmetry still held when animate and inanimate stimuli were balanced along an orthogonal shape dimension (Bracci and Op de Beeck, 2016). We also tested whether the same RT-distance relationship held when observers performed a shape categorization task, and an abstract object location categorization task that criss-crossed the dimensions of animacy and shape. Using human fMRI (N = 15), and focusing on shape and object category-selective regions of visual cortex, we correlated neural distance from a classifier decision boundary with observer RTs on the animacy, shape, and location tasks. In line with previous findings we found a negative correlation between RT and distance for the animacy task, and also the shape task, but the same relationship was not observed for the location task. This negative finding suggests that the neural coding for these other properties of objects might depend on a different mapping between activation space and categorization behavior.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only