Bayesian confidence in optimal decisions
The optimal way to make decisions in many circumstances is to track the difference in evidence collected in favour of the options. The drift diffusion model (DDM) implements this approach, and provides an excellent account of decisions and response times. However, existing DDM-based models of confid...
| Main Authors: | , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
American Psychological Association
2024
|
| _version_ | 1826317823378456576 |
|---|---|
| author | Calder-Travis, J Charles, L Bogacz, R Yeung, N |
| author_facet | Calder-Travis, J Charles, L Bogacz, R Yeung, N |
| author_sort | Calder-Travis, J |
| collection | OXFORD |
| description | The optimal way to make decisions in many circumstances is to track the difference in evidence collected in
favour of the options. The drift diffusion model (DDM) implements this approach, and provides an excellent
account of decisions and response times. However, existing DDM-based models of confidence exhibit certain
deficits, and many theories of confidence have used alternative, non-optimal models of decisions. Motivated by
the historical success of the DDM, we ask whether simple extensions to this framework might allow it to better
account for confidence. Motivated by the idea that the brain will not duplicate representations of evidence, in
all model variants decisions and confidence are based on the same evidence accumulation process. We compare
the models to benchmark results, and successfully apply 4 qualitative tests concerning the relationships between
confidence, evidence, and time, in a new preregistered study. Using computationally cheap expressions to model
confidence on a trial-by-trial basis, we find that a subset of model variants also provide a very good to excellent
account of precise quantitative effects observed in confidence data. Specifically, our results favour the hypothesis
that confidence reflects the strength of accumulated evidence penalised by the time taken to reach the decision
(Bayesian readout), with the penalty applied not perfectly calibrated to the specific task context. These results
suggest there is no need to abandon the DDM or single accumulator models to successfully account for confidence
reports. |
| first_indexed | 2024-03-07T08:17:57Z |
| format | Journal article |
| id | oxford-uuid:f5696994-8d91-472e-a7a0-f1637c5c6573 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2025-03-11T17:00:01Z |
| publishDate | 2024 |
| publisher | American Psychological Association |
| record_format | dspace |
| spelling | oxford-uuid:f5696994-8d91-472e-a7a0-f1637c5c65732025-03-10T10:04:31ZBayesian confidence in optimal decisionsJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:f5696994-8d91-472e-a7a0-f1637c5c6573EnglishSymplectic ElementsAmerican Psychological Association2024Calder-Travis, JCharles, LBogacz, RYeung, NThe optimal way to make decisions in many circumstances is to track the difference in evidence collected in favour of the options. The drift diffusion model (DDM) implements this approach, and provides an excellent account of decisions and response times. However, existing DDM-based models of confidence exhibit certain deficits, and many theories of confidence have used alternative, non-optimal models of decisions. Motivated by the historical success of the DDM, we ask whether simple extensions to this framework might allow it to better account for confidence. Motivated by the idea that the brain will not duplicate representations of evidence, in all model variants decisions and confidence are based on the same evidence accumulation process. We compare the models to benchmark results, and successfully apply 4 qualitative tests concerning the relationships between confidence, evidence, and time, in a new preregistered study. Using computationally cheap expressions to model confidence on a trial-by-trial basis, we find that a subset of model variants also provide a very good to excellent account of precise quantitative effects observed in confidence data. Specifically, our results favour the hypothesis that confidence reflects the strength of accumulated evidence penalised by the time taken to reach the decision (Bayesian readout), with the penalty applied not perfectly calibrated to the specific task context. These results suggest there is no need to abandon the DDM or single accumulator models to successfully account for confidence reports. |
| spellingShingle | Calder-Travis, J Charles, L Bogacz, R Yeung, N Bayesian confidence in optimal decisions |
| title | Bayesian confidence in optimal decisions |
| title_full | Bayesian confidence in optimal decisions |
| title_fullStr | Bayesian confidence in optimal decisions |
| title_full_unstemmed | Bayesian confidence in optimal decisions |
| title_short | Bayesian confidence in optimal decisions |
| title_sort | bayesian confidence in optimal decisions |
| work_keys_str_mv | AT caldertravisj bayesianconfidenceinoptimaldecisions AT charlesl bayesianconfidenceinoptimaldecisions AT bogaczr bayesianconfidenceinoptimaldecisions AT yeungn bayesianconfidenceinoptimaldecisions |