Flexible adjustment of anticipations in human outcome processing.

To sense whether thoughts refer to current reality or not, a capacity called orbitofrontal reality filtering, depends on an orbitofrontal signal when anticipated outcomes fail to occur. Here, we explored the flexibility and precision of outcome processing in a deterministic reversal learning task. Healthy subjects decided which one of two colored squares hid a target stimulus. Brain activity was measured with high-density electroencephalography. Stimuli resembling, but not identical with, the target stimuli were initially processed like different stimuli from 210 to 250 ms, irrespective of behavioral relevance. From 250 ms on, they were processed according to behavioral relevance: If they required a subsequent switch, they were processed like different stimuli; if they had been declared potential targets, they were treated like true targets. Stimuli requiring a behavioral switch induced strong theta activity in orbitofrontal, ventromedial, and medial temporal regions. The study indicates flexible adaptation of anticipations but precise processing of outcomes, mainly determined by behavioral relevance.