Purchase this article with an account.
Jeffrey Y. Lin, Bjorn Hubert-Wallander, Sung Jun Joo, Scott O. Murray, Geoffrey M. Boynton; It’s not just gist! Recognition memory for scrambled scenes with limited attentional resources. Journal of Vision 2012;12(9):1076. doi: https://doi.org/10.1167/12.9.1076.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The visual system can rapidly extract the ‘gist’ of complex visual scenes (Potter, 1975; Potter et al., 2010). Researchers argue that gist perception involves a global analysis of a scene’s spatial properties, such as its spatial layout, without necessarily encoding individual object details (Oliva & Torralba, 2001). Alternatively, we might analyze diagnostic objects in a scene (Stigliani et al., 2010; Quattoni & Torralba, 2009). We sought to determine if participants could extract more than the gist of a scene, even when attentional demands were high and scene information content was low.
In Experiment 1, participants performed a demanding digit counting RSVP task at fixation while simultaneously categorizing a target scene presented briefly behind the fixation task (Cohen & Alvarez, 2011). Participants first reported the number of digits they observed and then reported the category that best matched the target scene. Participants were then asked to select the specific target scene among 5 same-category distractors. Surprisingly, participants correctly selected 50% of the specific target scenes among same-category distractors even when scenes were presented for 67 ms and were masked before and after with random colored blocks (chance performance was 16.7%).
In Experiment 2, we scrambled scenes into 4x4, 7x7 or 10x10 blocks then shuffled them to create target scenes that had disrupted spatial layouts and minimal object information. Even when target scenes were scrambled into 10x10 blocks and presented for 67 ms, participants correctly identified the gist for 66% of the target scenes while performing the digit task. Critically, participants were also able to select the correct unscrambled version of the target scene among 5 same-category distractors 40% of the time. Together, these results suggest that participants are able to encode details beyond the gist of a scene even with limited attentional resources or when the scenes lack spatial layout or object information.
Meeting abstract presented at VSS 2012
This PDF is available to Subscribers Only