Category Learning Stretches Neural Representations in Visual Cortex

Abstract

We review recent work that shows how learning to categorize objects changes how those objects are represented in the mind and the brain. After category learning, visual perception of objects is enhanced along perceptual dimensions that were relevant to the learned categories, an effect we call dimensional modulation (DM). DM stretches object representations along category-relevant dimensions and shrinks them along category-irrelevant dimensions. The perceptual advantage for category-relevant dimensions extends beyond categorization and can be observed during visual discrimination and other tasks that do not depend on the learned categories. fMRI shows that category learning causes ventral stream neural populations in visual cortex representing objects along a category-relevant dimension to become more distinct. These results are consistent with a view that specific aspects of cognitive tasks associated with objects can account for how our visual system responds to objects.

Extended categorization of conjunction object stimuli decreases the latency of attentional feature selection and recruits orthography-linked ERPs

“…The earlier onset, more sustained SN to trained stimuli was quite similar to the result of (Smid et al, 1999), who observed an earlier onset SN to salient, easily perceived target features than smaller, less salient visual shape features. This suggests that the attention effect we observed was driven by an increased ability to perceptually discriminate between the features of the trained stimuli and links our result with previous demonstrations of task independent increases in discriminability resulting from category learning (Folstein, Palmeri, Van Gulick, & Gauthier, 2015).…”

Occipitotemporal Representations Reflect Individual Differences in Conceptual Knowledge

“…Category training is known to induce changes in both perceptual (Folstein et al, 2012; Op de Beeck et al, 2003; Goldstone et al, 1996; Goldstone, 1994; Gureckis & Goldstone, 2008) and neural sensitivity (e.g., Dieciuc et al, 2017; Folstein et al, 2013, 2015; Li et al, 2007; Sigala & Logothetis, 2002). In two datasets, we demonstrate that occipitotemporal stimulus representations covary with the attentional parameters derived from formal categorization theory.…”

The Latent Structure of Dictionaries

“…What is needed in order to ground either pictures or words in their referents is a mechanism that can recognize sensorimotor categories. Modeling the capacity for the learning, recognition, and representation of sensorimotor categories has become a rich and fertile field (Ashby & Maddox, ; De Vega et al., ; Folstein et al., ; Lupyan, ; Maier et al., ; Meteyard et al., ; Pezzulo et al., ). In a dual‐coding sensorimotor/symbolic model it is the sensorimotor module that needs to connect words to their referents.…”

“…The more concrete of these categories, and hence the words that name them, can be learned directly through trial‐and‐error sensorimotor experience, guided by feedback that indicates whether an attempted categorization was correct or incorrect. The successful result of this learning is a sensorimotor category representation—that is, a feature‐detector that enables the learner to categorize sensory inputs correctly, identifying them with the right category name (Ashby & Maddox, ; Folstein, Palmeri, Van Gulick, & Gauthier, ; Hammer, Sloutsky, & Grill‐Spector, ). A grounding set composed of such experientially grounded words would then be enough (in principle, though not necessarily in practice) to allow the meaning of all further words to be learned through verbal definition alone.…”