To gain a better understanding of how perceptual expertise affects object processing, we trained subjects to be experts at a novel class of homogeneous objects. In contrast to earlier studies, the stimuli, although based on Greebles, were asymmetrical in terms of part arrangement. To assess how object processing changed with expertise, we ran an extensive battery of behavioral tests concurrently with expertise training, including: (1) a same/different task between two sequentially presented images of Greebles where both were either upright or inverted; (2) a same/different task between two simultaneously presented images of Greebles where we varied the presentation time from 50-500 ms; (3) a visual search task where subjects search for a familiar or unfamiliar Greeble among distractors that were either more or less visually similar; (4) an interference task where a Greeble or an object from a non-expert domain (e.g., a car) appeared flanked on both sides by either Greebles or objects from the non-expert class (subjects were told to ignore the flankers and judge whether the target was upright or inverted); and (5) a recognition memory task for familiar versus unfamiliar examples from classes trained at either the individual level (Greebles) or basic level (Fribbles). Based on the assumption that post training configural (holistic) processing is applied automatically to objects from domain of expertise, it was hypothesized that subjects perform differently from their pre-training performance compared to control objects (expertise effects may facilitate or hinder processing). Indeed, this was the case with some tasks showing improved performance and others showing decrements in performance for experts. Such results suggest that perceptual expertise is not a unitary process, but rather recruits a wide variety of mechanisms in the perceptual "tool-kit" that reflect different components of the default processing applied by experts in their domain of expertise.
All Science Journal Classification (ASJC) codes
- Sensory Systems