Modeling search guidance: Three parameters for characterizing performance in different types of visual search.

The accuracy of search guidance is reflected in the probability of fixating distractors within a search display as a function of their similarity to the target (e.g., Stroud et al. , JEP: HPP, 2012). Strong search guidance produces high fixation rates to distractors that match the target on a given dimension (e.g., color), and low fixation rates to distractors that do not match the target. When guidance is absent, fixation rates are equal across all distractors, regardless of similarity to target. We modeled fixation rates across different levels of target similarity using a sigmoid function. Three parameters are necessary and sufficient to fit the function across a range of color-search tasks: single-target search, dual-target search, search plus a working-memory task, and search plus shape discrimination (modeled data: Stroud et al., ACP, 2011; Stroud et al., VSS, 2011; Stroud et al. , JEP: HPP, 2012; Menneer et al., BPS, 2014). These parameters are as follows. (1) Unguided fixation rate (u): Even in guided search, participants exhibit a baseline fixation rate to distractors that are maximally different to the target. u defines the ratio of unguided fixations to guided fixations. (2) Selectivity (s): Participants with high selectivity (e.g., for target-similar colors) exhibit a guidance curve that drops off steeply for distractors less similar to the target, while low selectivity produces a shallow drop-off. (3) Target region (t): Some tasks produce high fixation rates to the target color only, but others produce high fixations rates to a range of colors similar to the target color. t allows adjustment of the region in color space that receives high fixation rates, and determines the level of dissimilarity at which the fixation rate begins to drop off steeply. The estimates for these three parameters demonstrate how guidance changes across different types of search. Meeting abstract presented at VSS 2015.