Finding

khalvati-kiani-rao-2021.psychometric.direction-choice-proxy.accuracy-no-sure-target.kiani-shadlen-m2

Khalvati K, Kiani R, Rao RPN. Bayesian inference with incomplete knowledge explains perceptual confidence and its deviations from accuracy. Nature Communications, 2021. · Macaque random-dot motion confidence wagering task · psychometric

Observed curve and fits

Joint-fit companions

Other findings jointly fit by the same model fits.

Fit diagnostics

Caveats on these fits (4 terms)

Each tag below is attached to one or more candidate fits for this finding. The chip column in the table flags which row carries which tag.

aggregate DDM RT approximation
DDM likelihood uses aggregate chronometric constraints rather than a full trial-level response-time likelihood.
figure/source data
Fit is based on paper figure source data or summarized values, not a full public trial table.
motion-duration RT proxy
Chronometric evidence comes from motion/stimulus duration summaries rather than observed reaction-time timestamps.
target-coded choice proxy
Choice side was reconstructed from accuracy or target-coded rows, so it should not be treated as an observed motor choice.
Variant AIC Δ AIC RMSE Caveats
logistic-4param winner 28654.3 0.0 0.0109
  • figure/source data
  • target-coded choice proxy
sdt-2afc 28664.7 10.4 0.0142
  • figure/source data
  • target-coded choice proxy
ddm-vanilla 28686.5 32.3 0.0205
  • aggregate DDM RT approximation
  • figure/source data
  • motion-duration RT proxy
  • target-coded choice proxy
Full diagnostics (logL · n · params · max abs error)
Variant logL n Free params Max |error| Predicted points
logistic-4param -14323.1 25,300 4 0.0173 5
sdt-2afc -14330.3 25,300 2 0.0215 5
ddm-vanilla -14340.3 25,300 3 0.0398 5

Residuals

Fit parameters and provenance

logistic-4param.khalvati-kiani-rao-2021.psychometric.direction-choice-proxy.accuracy-no-sure-target.kiani-shadlen-m2

dirty fit

bias=5.780 · lower_lapse=0.118 · slope=16.107 · upper_lapse=0.028

Method
scipy.optimize.minimize
Commit
3ec74b4
Predicted points
5
BIC
28686.8

Fitted with scipy.optimize.minimize via behavtaskatlas.model_fits.logistic-4param; success=True; CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH

sdt-2afc.khalvati-kiani-rao-2021.psychometric.direction-choice-proxy.accuracy-no-sure-target.kiani-shadlen-m2

dirty fit

criterion=0.046 · d_prime=0.030

Method
scipy.optimize.minimize
Commit
3ec74b4
Predicted points
5
BIC
28681.0

Fitted with scipy.optimize.minimize via behavtaskatlas.model_fits.sdt-2afc; success=True; CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH

ddm-vanilla.khalvati-kiani-rao-2021.psychometric.direction-choice-proxy.accuracy-no-sure-target.kiani-shadlen-m2

dirty fit

boundary=0.023 · drift_per_unit_evidence=2.083 · non_decision_time=0.242

Method
scipy.optimize.minimize
Commit
3ec74b4
Predicted points
5
BIC
28711.0

Fitted with scipy.optimize.minimize via behavtaskatlas.model_fits.ddm-vanilla; success=True; CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH

Observed points

Target-coded motion strength p_right n
3.200 0.5268 5,091
6.400 0.5359 5,077
12.800 0.6327 4,980
25.600 0.7843 5,071
51.200 0.9236 5,081

Take it with you

Cover sheet

Self-contained Markdown for citation, slides, or notebooks

The cover sheet pins the finding to the atlas commit and includes the observed points, fit ranking, caveats, and provenance — everything needed to drop into a paper or notebook without losing the trail back to the deploy.