Purchase this article with an account.
Amy Guthormsen, Michael Ham, Brenna Fearey, Luis Bettencourt, John George; Effect of Target/Non-Target Similarity on the Timecourse of Visual Object Recognition: An ERP investigation. Journal of Vision 2012;12(9):819. doi: 10.1167/12.9.819.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Some have argued that, because visual object recognition occurs so fast (<150 ms.), it must be essentially a feed-forward system. A study commonly cited as support reveals divergence between brain responses to visual targets and those to non-targets occurring around 150 ms after stimulus onset (Thorpe et al. 1996). However, it has been argued that visual targets (animals) differed in systematic ways from the non-targets (natural scenes with no animal). In this study, we investigated the effect of target /non-target similarity on the timecourse of ERP divergence. We recorded EEG data while participants did a binary-choice target recognition task in which the target was always either cat or dog (varied within-subject). Non-target stimuli included dogs, cats, and natural scenes with no animal. Dissimilar non-target waveforms diverged from target waveforms with a strikingly similar timecourse (roughly 150 ms post stimulus onset) and scalp distribution to that reported by Thorpe and colleagues. The difference between similar non-target and target waveforms differed. Both the time of divergence (350 ms) and the scalp distribution of target – non-target differences (central-parietal positivity) were reminiscent of a P300 effect. To explore the generality of this effect, we repeated the experiment with a different stimulus set, in which the target was always a male or a female face, and stimuli included male and female faces as well as images of non-face objects. Dissimilar non-targets and targets diverged with a similar timecourse and scalp distribution as seen in the cat/dog experiment. However, the difference between similar non-targets and targets had a more localized scalp distribution. These data are consistent with the theory that visual object recognition consists of a fast, feed-forward process that is adequate to make broad stimulus discriminations, and a slower process necessary to resolve detailed, specific representations, which may rely on feedback between visual areas.
Meeting abstract presented at VSS 2012
This PDF is available to Subscribers Only