Task designs

In Study 1 (Fig. 3a), participants were presented with a simple discrimination rule for 4 s (e.g. yellow shapes = left button response and orange shapes = right button response) followed by a sequence of coloured shapes. There were 4 compound stimuli per rule, constructed from 2 exemplars per dimension. There was no feedback post response. Stimuli were presented in randomised order at a rate of 1 per 1.7 s with 1/3 of trials showing fixation as opposed to a stimulus, which allowed activation during discriminations to be estimated relative to fixation. Presentation continued for 3 min, subsequent to which a new rule was presented; therefore, activations related to rule learning were not confounded by the total time spent in scanner or on task. Rules always changed across dimensions; i.e. if one rule related to shape then the next related to colour and all exemplars were replaced when the rules changed; this design ensured that the previously learned stimulus–response mappings did not have to be overridden. There were a total of four rule slides, each followed by a 3 min sequence of discriminations. Study 2 (Fig. 3b) used a variant on the design of Study 1 with the same rules and stimuli (Fig. 3c). However, the participants had to derive the rule based on feedback as opposed to explicit instruction with a rule slide. Feedback was presented centrally on the screen as either the word ‘Correct’ in green or ‘Incorrect’ in red after a random 50% of trials, which allowed activations related to rule novelty and feedback to be estimated separately. The duration of each block was reduced to 2.5 min based on the rapid learning effects observed in Study 1. Behavioural outlier values (defined as > 2.5 SDs from the mean) were winsorised within condition for both studies to ensure they did not distort the results.

a) In Study 1 participants learnt novel discrimination rules from explicit instruction. Initially, a slide was presented with a discrimination rule. Subsequently, a sequence of coloured shapes was presented for 3 min and the participant was required to respond with the relevant button press as quickly and accurately as possible. After 3 min a new rule slide was displayed followed by another sequence of coloured shapes. b) In Study 2, there were no rule slides. Instead, feedback indicating whether the previous response was correct or incorrect was presented randomly after 50% of trials. Therefore, participants were required to derive the discrimination rules by a process of trial-and-error. c) The two studies used the same stimulus sets. d) In Study 1, response times followed a non-monotonic decrease when stimulus–response rules were being learnt from instruction. Specifically, there was a rapid decrease in RT from stages one to three, followed by a small increase in RT then a more gradual decrease. e) In Study 2, a similar non-monotonic decrease was also evident in RTs when stimulus-response rules were being learnt by exploration with feedback. (***p < 0.001, **p < 0.01, *p < 0.05 two tailed significance).