Individual differences in eye movements and perceptual averaging

Poster Presentation: Sunday, May 19, 2024, 8:30 am – 12:30 pm, Pavilion
Session: Attention: Spatial selection 1

William G. Hayward1 (), Herrick Fung2, Janet H. Hsiao3; 1Lingnan University, 2Georgia Institute of Technology, 3The University of Hong Kong

Numerous studies have documented our ability to perceive visual statistical averages over groups of similar objects. Nevertheless, the spatial attentional mechanisms and perceptual computations that constitute the basis of this ensemble representation remain unclear. Here, we explored these issues using a gaze-contingency paradigm. Participants performed a mean orientation estimation task under three conditions: (1) foveal vision only (obscuring the visual field further than 2.2° from fixation), (2) peripheral vision only (obscuring foveal information less than 2.2°), and (3) natural viewing. Both ensemble variance and stimulus duration were manipulated. Behaviorally, we found that participants could achieve similar accuracy as natural viewing when foveal vision was blocked but were impaired when peripheral vision was blocked. These results suggest that elements in the periphery are typically integrated into the ensemble representation, when available. Modeling analysis further revealed individual differences in integration mechanisms, with one group of participants better predicted by models assuming equal weighting of all items (which we liken to a Zoom-lens Model), as compared to another group which was better predicted by models weighted by the spatial distance between their fixations and the corresponding items (akin to a Spotlight Model). Using a Hidden Markov Model-based data-driven eye movement analysis, we found that participants could be clustered into two groups; one group tended to land their fixations on the center region of the ensemble whereas the other group distributed their fixations across the whole ensemble. Combining both analyses, we discovered that there was a weak correlation (r = 0.4) between the two groupings. Specifically, those better predicted by the zoom-lens model also tend to demonstrate a central fixation bias in their pattern of eye movements. Our findings argue against a universal mechanism in explaining perceptual averaging. Instead, individual strategic differences may play a role in determining the nature of perceptual averaging.

Acknowledgements: This research was supported by a Grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (17608519) to W.G.H.