Confidence in visual motion discrimination is preserved in individuals with schizophrenia

Abstract

Background: Metacognition is the set of reflexive processes that allows humans to evaluate the accuracy of their mental operations. Metacognitive deficits have been described in people with schizophrenia using mostly narrative assessment, and they have been linked to several key symptoms.

Methods: We assessed metacognitive performance objectively by asking people with schizophrenia or schizoaffective disorder (n = 20) and matched healthy participants (n = 21) to perform a visual discrimination task and report their confidence in their performance. Metacognitive performance was defined as the adequacy between visual discrimination performance and confidence.

Results: Bayesian analyses revealed equivalent metacognitive performance in the 2 groups, despite a weaker association between confidence and trajectory tracking during task execution among people with schizophrenia. We reproduced these results using an evidence accumulation model, which showed similar decisional processes in the 2 groups.

Limitations: These results from a relatively small study sample cannot be generalized to other perceptual and nonperceptual tasks. To meet this purpose, ecological tasks are needed. As well, the role of antipsychotic medication and design deserves greater attention in the future.

Conclusion: We found similar decisional and metacognitive capabilities between people with schizophrenia and healthy controls in a visual discrimination task.

Fig. 1

Experimental paradigm and behavioural performance. (A) Experimental paradigm. Participants were presented with a random-dot kinetogram stimulus moving rightward or leftward and asked to report motion direction by moving the mouse cursor toward a circle presented at the top left or top right of the screen (first-order response). Then, participants reported the confidence they had in their response by moving a cursor on a visual analogue scale (second-order response). Exemplar mouse trajectory and confidence ratings are shown in red. (B) Posterior distribution density of the M-Ratio for the control (green) and schizophrenia groups (orange). The coloured lines at the bottom of the plot represent the 95% highest posterior density intervals. (C) Mixed-effects logistic regression between first-order accuracy and standardized confidence. (D) Mixed-effects linear regression between standardized reaction times and confidence. In panels C and D, regression lines and 95% confidence intervals around them represent the model fit. Although the model took continuous variables as input, for illustrative purposes we plotted dots and error bars that represent mean ± 95% confidence interval over participants after rounding standardized confidence (C) and reaction times (D). The size of each dot is proportional to the number of represented trials.

Fig. 2

Trajectory tracking. (A) Single-trial mouse trajectories leading to the first-order response in case of a change of mind (red) or no change of mind (black) in the control (left panel) and patient groups (right panel). (B) Average first-order accuracy and confidence in the presence (red) and absence (black) of a change of mind in the control (left panels) and patient groups (right panels). (C) Goodness of fit (R²) and slope (β) of the linear fit between vertical and horizontal mouse positions as a function of confidence quantile (low: red, medium: orange, high: green). Large dots represent average estimates, error bars represent the 95% confidence intervals. Small dots represent individual estimates. (D) Average velocity (upper panel) and acceleration (lower panel) from first mouse movement onset as a function of confidence quantile (low: red, medium: orange, high: green). Of note, velocity may be non-null before movement onset as it was defined as a function of the maximal velocity in a given trial (see Appendix 1). Shaded areas represent the 95% confidence intervals. Grey bars represent samples for which confidence covaried significantly with kinematics (p < 0.05, corrected for false discovery rate).