Can the neural basis of repression be studied in the MRI scanner? New insights from two free association paradigms.

BACKGROUND

The psychodynamic theory of repression suggests that experiences which are related to internal conflicts become unconscious. Previous attempts to investigate repression experimentally were based on voluntary, intentional suppression of stimulus material. Unconscious repression of conflict-related material is arguably due to different processes, but has never been studied with neuroimaging methods.

METHODS

We used functional magnetic resonance imaging (fMRI) in addition with skin conductance recordings during two free association paradigms to identify the neural mechanisms underlying forgetting of freely associated words according to repression theory.

RESULTS

In the first experiment, free association to subsequently forgotten words was accompanied by increases in skin conductance responses (SCRs) and reaction times (RTs), indicating autonomic arousal, and by activation of the anterior cingulate cortex. These findings are consistent with the hypothesis that these associations were repressed because they elicited internal conflicts. To test this idea more directly, we conducted a second experiment in which participants freely associated to conflict-related sentences. Indeed, these associations were more likely to be forgotten than associations to not conflict-related sentences and were accompanied by increases in SCRs and RTs. Furthermore, we observed enhanced activation of the anterior cingulate cortex and deactivation of hippocampus and parahippocampal cortex during association to conflict-related sentences.

CONCLUSIONS

These two experiments demonstrate that high autonomic arousal during free association predicts subsequent memory failure, accompanied by increased activation of conflict-related and deactivation of memory-related brain regions. These results are consistent with the hypothesis that during repression, explicit memory systems are down-regulated by the anterior cingulate cortex.