Autonomous Mechanism of Internal Choice Estimate Underlies Decision Inertia

Abstract

Our choice is influenced by choices we made in the past, but the mechanism responsible for the choice bias remains elusive. Here we show that the history-dependent choice bias can be explained by an autonomous learning rule whereby an estimate of the likelihood of a choice to be made is updated in each trial by comparing between the actual and expected choices. We found that in perceptual decision making without performance feedback, a decision on an ambiguous stimulus is repeated on the subsequent trial more often than a decision on a salient stimulus. This inertia of decision was not accounted for by biases in motor response, sensory processing, or attention. The posterior cingulate cortex and frontal eye field represent choice prediction error and choice estimate in the learning algorithm, respectively. Interactions between the two regions during the intertrial interval are associated with decision inertia on a subsequent trial.

Reinforcement Learning in Multidimensional Environments Relies on Attention Mechanisms

“…Although this model does not explicitly involve attention processes or hypothesis testing, it emulates both: the decay of weights of unchosen features allows the model to focus learning on the weight of one consistently chosen feature. At the same time, the decay implements a form of a "choice kernel" that allows the model to better predict future choices based on the repetition of actions at the level of features (Lau and Glimcher, 2005;Schönberg et al, 2007;Wilson and Niv, 2011;Seymour et al, 2012;Akaishi et al, 2014). However, the superior performance of the fRLϩdecay model cannot be wholly attributed to a choice kernel, as simply adding a feature-level choice kernel to the fRL model (without weight decay) improved the fit of the model compared with the fRL model ( p Ͻ 10 Ϫ7 , paired t test), but was still inferior to the fRLϩdecay model ( p Ͻ 10 Ϫ9 , paired t test; results not shown).…”

Post-decisional accounts of biases in confidence

“…This is assuming that evidence is integrated without bias. However, several studies have shown that post-decisional processing could be biased, and so could distort confidence judgments [17,26,[31][32][33][34]. For example, evidence for the chosen option could be overweighed (i.e., accumulated at a larger rate than the unchosen options as in [31]) leading to an increase in confidence that is not based on objective evidence; this is known as "confirmation bias" [26].…”

“…Other sources of bias include serial dependencies; i.e., conditions in which choices made in the recent past influence upcoming decisions [32][33][34]. Such bias has been seen in a low level task, orientation judgement, in which participants' choices were significantly biased toward orientations reported in the previous trials even though the stimuli changed randomly trial-by-trial [33].…”

Low attention impairs optimal incorporation of prior knowledge in perceptual decisions

“…Uncued trials, then, would have been more affected by adaptation and possibly prevented subjects from incorporating priors into their responses in a way not related to our model. A similar worry is that subjects could have tried to balance their responses in uncued trials, thereby avoiding the most frequent response in the most frequent and salient trials (i.e., cued trials) and subsequently reducing the (more optimal) shift in their response bias (Akaishi, Umeda, Nagase, & Sakai, 2014).…”