Adult Category Learning Differences Predicted by a Dynamic Neural Field Theory Account of Information Sampled from the Fovea

Abstract

Here we explore the possibility that the speed of learning is affected by the precision of our sensory estimates for the learned category’s diagnostic feature dimensions. Colour information from a foveated stimulus, if represented as a sample on a metric colour dimension, should faithfully represent differences in the shapes and precision of the estimates as a consequence of the sample size. Differences in the sample variability are expected to have affect on exactly what gets associated during learning. We provide evidence that a manipulation in sub-fovea feature size, 0.18° vs 1.19° of visual angle, influences learning speed. In both conditions the simple colour features are easy to see and we do not detect any gaze differences as measured by total fixation durations and individual feature fixation durations. Learning methods that metrically represent activity on feature dimensions such as Dynamic Field Theory (DFT) may be able to account for this data.


Back to Table of Contents