Purchase this article with an account.
Marlene Poncet, Leila Reddy, Michele Fabre-Thorpe; Presentation time does not affect superordinate-level advantage in ultra-rapid categorization. Journal of Vision 2012;12(9):809. doi: https://doi.org/10.1167/12.9.809.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
A recent study (Mace et al., 2009) challenged the prevalent idea of a basic-level advantage in categorization. They showed that the processing time required to access basic visual categories (such as a dog) is longer than that needed to access superordinate visual categories (such as an animal). Mack & Palmeri (2011), among others, argue that these results can be explained as a consequence of the brief stimulus presentation time (25ms) used in the Mace et al. study. Visual information gathered from such fast presentations could be minimal, degraded, and emphasize coarse visual information. In this study, we tested whether the superordinate-level advantage (SLA) found with briefly flashed images could be reproduced with longer stimulus presentation times. Participants were asked to perform a go/no-go visual categorization task at either the superordinate (animal/non-animal) or the basic (bird/non-bird animal) level. Images were presented for 25ms, 250ms or 500ms. Reaction times were faster for superordinate-level categorization than for basic-level categorization at all presentation durations (SLA at 25ms = 55±10ms; 250ms = 34±4ms; 500ms = 37±6ms; one-way repeated measures ANOVA: F(2,24) = 2.43; p>0.1). The same SLA was also found when a yes/no response was required instead of a go/no-go (presentation time 250ms; SLA for yes/no = 29±10ms; go/no-go = 32±6ms; paired t-test: t(8)=0.22, p>0.8). Our results clearly show that the superordinate category is accessed faster than the basic category, regardless of the presentation time or the type of response. Thus, presentation time cannot explain the basic-level advantage found in other studies. However, other factors, such as the kind of stimuli (isolated animal or animal in a natural scene), the kind of task (naming or detection), might be involved. These possibilities must be tested in further experiments.
Meeting abstract presented at VSS 2012
This PDF is available to Subscribers Only