Traces of expectancy breaches in sequential events: Differences between destabilized and adapted predictions

The fMRI-study aimed to investigate how and by the help of which cortical areas the frontomedian cortex succeeds in prediction of events despite continually and rapidly changing events. A former study observed the frontomedian cortex to be activated if a sequential event could be predicted well, whereas a reduction of that activation was present when prediction was destabilized by a preceding change of that event. If this reduction of frontomedian activity is a neural signature of destabilization only or of adaptation to altered events as well should be clarified here. Furthermore, we hypothesized parietal or temporal areas to reflect adaptation to altered events. In a pre-fMRI session subjects (n = 21) learned a sequence of 24 digits. During the fMRI session this sequence was repeatedly presented digit by digit. Unexpectedly digits were omitted and subjects had to indicate these sequential changes by key presses. Half of all omissions were repeated (double omissions) and half were not (single omissions). If events can be predicted well, i.e. in non-destabilized events (NDE) the frontomedian cortex was activated. This activity was reduced in destabilized events after single omissions (DES) as known but almost reached the activity level as in NDE in destabilized events after double omissions (DED). Thus, the frontomedian activity reduction might only be a signature of destabilization. DED compared to DES revealed an activation increase in temporal and frontal areas. This might be a cortical signature of breaking through a beginning adaption and of "repairing the known event syntax" after repeated event-alterations.