| Summary: | <p>Perceptual decisions are of fundamental interest, and it is increasingly recognised
that a sense of confidence supports decision-making. The drift diffusion model
(DDM), provides an excellent account of decisions and response times. It also
features the optimal property of tracking the difference in evidence between two
options. However, the DDM struggles to account for human confidence reports.
Possibly because of this, much confidence research has used non-normative models
of the decision mechanism. Motivated by the idea that perceptual decision-making
will reflect optimal computation, we consider 10 variants of the DDM. Motivated
by the idea that the brain will not duplicate the representation of evidence, in all
model variants confidence is read out from the decision mechanism.</p>
<p>We compare the models to benchmark findings and make 3 qualitative predictions
that we verify in a preregistered study. We develop new computationally cheap
predictions for confidence in DDM accounts. Using these predictions we model
confidence on a trial-by-trial basis, finding that a subset of model variants provide an
excellent account of the precise quantitative effects observed. Specifically, models in
which confidence reflects a miscalibrated Bayesian readout perform best. Focusing
on the idea of a Bayesian readout, we explore the claim that evidence associated
with a selected option is overweighted in confidence. Although this idea appears to
challenge Bayesian confidence, we do not find this overweighting in three previously
collected datasets, and we suggest an explanation for previous results. Finally, we
examine whether the DDM remains normative in settings where rapid learning is
more important than the theoretical possibility of achieving maximal reward rate.</p>
<p>These results support the claim that confidence is based on the decision
mechanism, which is itself optimal. In particular, the DDM with a miscalibrated
Bayesian readout is supported on both theoretical and empirical grounds.</p>
|
|---|