Fig 1.
Schematic representation of the beads task.
(A) Example of the 10th trial of one sequence. Two bags are displayed which contain either 80 black and 20 white beads, or the converse. Beads are drawn sequentially with replacement. Each of the five sequences consists of 20 drawn beads and the result of each draw, i.e. the color of the bead, remains displayed on the left side of the screen. Within 10 seconds, participants have to indicate their certainty about the bag of origin. They do so by dragging the marker on a visual scale either to the left or the right side. This slider is reset to the center after every trial. (B) At the end of each sequence, feedback about beads seen, choices made and the actual bag of origin of all beads is provided.
Fig 2.
Correlations between observed and predicted probabilities across the task.
For each participant, a Spearman correlation was calculated between the participant’s subjective probability ratings and those predicted by the volatility model with the best fitting volatility parameter. Single points represent the corresponding correlation coefficient (ρ) for each participant and are colored by size of the corresponding volatility estimate (with clusters of high and low volatility based on a bimodality analysis reported in the results section).
Fig 3.
Schematic representation of the visual working memory task.
(A) A trial starts with the display of a fixation cross and participants initiate the presentation of the sample shape by clicking on it. (B) The sample shape is then presented for one second, followed by a fixation cross (C). Clicking on it initiates the recall phase (D) in which 30 shapes are presented in a circular arrangement (note that in the example above, only 15 of the 30 shapes, enlarged, are shown for better visibility). (E) The participants now click onto the shape that most resembles what they remember, and set a capture area surrounding it. (F) They receive feedback by being shown the same shape as during the sample phase, correctly placed in the array of shapes. If it is included in the capture area they selected (as in the example above), the excessive part of that capture area is highlighted in red.
Table 1.
Sample demographics per group (total sample size = 86).
Fig 4.
Boxplots per group for all main variables.
Beads task (BT) variables are initial certainty (A), untransformed disconfirmatory belief updating (B) and estimated volatility (C). Variables from the visual working memory task (VWMT) include error (D) and proportion of hits (E). NT = neurotypically developing individuals, ASD = individuals with autism spectrum disorder, SCZ = individuals with schizophrenia. All BT variables are expressed as probabilities, average error is expressed in degrees, and proportion of hits is the proportion of trials where the capture area included the target. Points represent single participants.
Table 2.
Descriptive summary statistics of the two tasks per group, effect size and p-value from the conducted ANOVAs (total sample size = 86).
Fig 5.
Scatterplot of disconfirmatory belief updating (untransformed) and estimated volatility from the beads task (BT).
Rho and p display the results of a Spearman correlation conducted across the total sample. Regression lines are fitted for each group for illustrative purposes only. NT = neurotypically developing individuals, ASD = individuals with autism spectrum disorder, SCZ = individuals with schizophrenia.
Table 3.
Model comparison.
Fig 6.
Posterior-predictive distributions of the winning model for all three groups.
(A) for neurotypically developing individuals (NT), (B) for individuals with autism spectrum disorder (ASD), (C) for individuals with schizophrenia (SCZ). Colored lines are posterior means of the posterior predictive distributions, shaded areas are the 5% and the 95% percentile. Black lines are the actual data. Vertical lines are the estimated volatility values for each participant based on the Bayesian volatility model as described in section ‘Measures: Beads task’. NT = neurotypically developing individuals, ASD = individuals with autism spectrum disorder, SCZ = individuals with schizophrenia.