Do perceptual biases emerge early or late in visual processing? Decision-biases in motion perception

Abstract

Visual perception is strongly influenced by contextual information. A good example is reference repulsion, where subjective reports about the direction of motion of a stimulus are significantly biased by the presence of an explicit reference. These perceptual biases could arise early, during sensory encoding, or alternatively, they may reflect decision-related processes occurring relatively late in the task sequence. To separate these two competing possibilities, we asked (human) subjects to perform a fine motion-discrimination task and then estimate the direction of motion in the presence or absence of an oriented reference line. When subjects performed the discrimination task with the reference, but subsequently estimated motion direction in its absence, direction estimates were unbiased. However, when subjects viewed the same stimuli but performed the estimation task only, with the orientation of the reference line jittered on every trial, the directions estimated by subjects were biased and yoked to the orientation of the shifted reference line. These results show that judgements made relative to a reference are subject to late, decision-related biases. A model in which information about motion is integrated with that of an explicit reference cue, resulting in a late, decision-related re-weighting of the sensory representation, can account for these results.

Benefits of commitment in hierarchical inference

“…Figure 2b shows the resulting distributions for both strategies given the specific parameter settings of our generative model. Note that the estimate distributions fundamentally differ; the conditioned inference strategy exhibits a characteristic bimodal distribution for θ values close to the category boundary, which matches a range of experimental results (Jazayeri and Movshon, 2007; Zamboni et al, 2016; Luu and Stocker, 2018).…”

“…3e). This corresponds to many of the estimation tasks reported in previous studies (Jazayeri and Movshon, 2007; Zamboni et al, 2016; Luu and Stocker, 2018). Local relative accuracy shows very similar curves (Fig.…”

“…These biases can be thought of as a form of consistency (Brehm, 1956) or confirmation bias (Nickerson, 1998) where the perceptual estimate aligns with and confirms the chosen category (Bronfman et al, 2015; Talluri et al, 2018). Importantly, these biases are not dependent on subjects making an explicit, overt choice of a high-level interpretation; similar biases were observed in tasks that did not require an explicit categorical choice (Wu et al, 2009; Zamboni et al, 2016; Ding et al, 2017), indicating that committing to a high-level interpretation could be an inherent inference strategy. As proposed (Stocker and Simoncelli, 2007), and refined and validated more recently (Luu and Stocker, 2018), the behavioral biases in above examples are remarkably well described by a conditioned observer model .…”

“…For example, by making an category assignment a subject’s subsequent perceptual estimate of a low-level stimulus feature ( e.g. , motion direction (Jazayeri and Movshon, 2007; Zamboni et al, 2016) or visual orientation (Luu and Stocker, 2018; Fritsche and de Lange, 2019)) is biased towards the assigned category on a per trial basis. These biases can be thought of as a form of consistency (Brehm, 1956) or confirmation bias (Nickerson, 1998) where the perceptual estimate aligns with and confirms the chosen category (Bronfman et al, 2015; Talluri et al, 2018).…”

Choices Change the Temporal Weighting of Decision Evidence

“…Recently developed task protocols provide new tools for assessing the impact of choices on the accumulation of subsequent decision evidence. These tasks prompt two successive judgments within the same trial: commonly a binary choice followed up by a continuous estimation [33][34][35][36][37][38] or a confidence [39,40] judgment. Specifically, some tasks prompt binary choice and estimation judgments sequentially, separated by a second evidence stream presented in between [35,39,40,38].…”