Table 1.
Socio-demographic information and results from a neuropsychological battery for the 33 healthy subjects who participated in experiment.
Fig 1.
(A) Trial structure. Each trial consisted of choices at two steps. Step 1 involved the first choice between two abstract gray stimuli (Chinese characters, not known to German subjects). The chosen stimulus was framed with red color in the center-top of the screen for 1.5s. Subsequently, subjects were presented with another stimulus pair in step 2. The second choice was rewarded with money (20 cents) or nothing. (B) The transitions from step 1 to step 2 remained fixed, with 70% and 30% of all trials as respectively common and rare transitions. The reward probabilities for each stimulus in step 2 changed independently between 25% and 75%, based on Gaussian random walks with reflecting boundaries [10]. Win probabilities varied, therefore, as a function of the trial number.
Table 2.
Computational model parameters: maximum a posteriori (MAP) estimates (bounded model-scale).
Fig 2.
Computational model parameter estimates.
The model parameter βGD, estimating the weight of the goal-directed systems to behavioral control, is displayed as a function of working memory capacity (Digit Span score) for the break conditions gaming (green points and regression lines) versus music (red).