Decision bounds are adaptive to dynamic task conditions

Classical models of decision-making, including drift-diffusion and other accumulate-to-bound models, assume that the decision rule for committing to a decision based on accumulated evidence is determined prior to observing the evidence. These predefined rules are typically modeled as thresholds, or bounds, on the accumulated evidence (“decision variable”) that are either fixed [1-2] or collapsing over time [3], in both cases serving to balance decision speed (emphasized when bounds are low and reached quickly) and accuracy (emphasized when bounds are high and allow for more evidence to be accumulated). However, a predefined rule is effective only insofar as it correctly anticipates the nature and quality of incoming evidence, which can be difficult in dynamic, real-world environments. Our goal is to understand if and how decision rules are adjusted to account for unpredictable changes in evidence quality between and within decisions. Our study uses a combination of normative theory and human and monkey psychophysics. The theory uses dynamic programming to determine the appropriate moment-to-moment adjustments to the decision bound, both within and across decisions, that maximize reward rate over a finite time horizon [4]. Guided by this theory, we designed a novel commitment task that, unlike traditional paradigms that estimate the decision variable and decision rule as latent parameters, presents the decision variable explicitly as a visual stimulus, allowing us to measure the decision bound directly on each trial. We present preliminary data from 19 human subjects performing the commitment task under varying evidence signal-to-noise ratios (SNRs) between and within trials. We also provide preliminary behavioral data from a monkey performing the commitment task with between-trial SNR modulation, along with a random dots discrimination task to assess cross-task generalization. Taken together, our findings provide strong preliminary evidence that decision bounds adjust adaptively to environmental dynamics.