Demographic and clinical characteristics of the included subjects.

Model-agnostic analyses of the 4-Armed Bandit Task.

Note. Violin plots showing probability to stay after (A) win, (B) loss, (C) both, (D) neither. CTR: healthy control; REL: participants with a first-degree relative with depression; REM: remitted depressed participants; MDD: currently depressed participants.

Group differences in computational parameters for the 4-Armed Bandit task.

Note. (A) Reward learning rate; (B) Punishment learning rate; (C) Reward sensitivity; (D) Punishment sensitivity; (E) Lapse. Note: CTR: healthy controls; REL: participants with at a first-degree relative with Depression; REM: remitted depressed participants; MDD: currently depressed participants. *p<.05 **p<.01***p<.001.

Parameter estimates and group comparisons on the 4 parameters in the “lapse” model for the 4-Armed Bandit task and the “ra_prospect” model in Gambling task.

Association between factor scores and punishment learning rate.

Note. Points represent individual participants. Solid lines show linear regression fits with 95% confidence intervals (shaded). Left panel: anhedonia factor scores (r = -.18, p=.035); right panel: apathy factor scores (r = -.18, p=.033).

Model-agnostic analyses of the Gambling task.

Note. Violin plots of the gambling task showing probability to take gamble choice under gain-only gamble (A) and mixed gamble (B) conditions in Study 2. CTR: healthy controls; REL: participants with at a first-degree relative with depression; REM: remitted depressed participants; MDD: currently depressed participants.

Group differences in computational parameters for the Gambling task.

Note. (A) Risk aversion; (B) Loss aversion; (C) Inverse temperature. CTR: healthy controls; REL: participants with at a first-degree relative with depression; REM: remitted depressed participants; MDD: currently depressed participants.*p<.05

Model and prior fits

Four-armed bandit and gambling task procedures.

Note. (A) Example trial of the four-armed bandit task. On each trial, participants chose one out of four bandits and received one out of four possible outcomes: reward (green token), punishment (red token), neither reward nor punishment (empty box) or both reward and punishment (red and green token). (B) On each gambling task trial, participants chose between a 50–50 gamble and a sure (guaranteed amount of points) option. Trials were either mixed gambles (50–50 chance of winning or losing points or sure option of 0 points) or gain-only trials (50–50 chance of winning or receiving nothing or sure gain).

Group difference in parameters for the lapse model, using a single prior for the 4-arm bandit task

Note. (A) Reward learning rate; (B) Punishment learning rate; (C) Reward sensitivity; (D) Punishment sensitivity; (E) Lapse. (F) Decay rate. Note: CTR: Control Participants; REL: Participants with at a first-degree relative with Depression; rMDD: Remitted depressed participants; MDD; Currently depressed participants.

Group difference in parameters of the ra_prospect model using single prior in the Gambling task.

Note. (A) Risk aversion; (B) Loss aversion; (C) Inverse temperature. Note: CTR: Control Participants; REL: Participants with at a first-degree relative with Depression; rMDD: Remitted depressed participants; MDD; Currently depressed participants.