ERP Correlates of Directed Forgetting Effects in Direct and Indirect Memory Tests

Directed forgetting effects on direct (recognition) and indirect (categorisation) memory tests were analysed using electrophysiological (ERPs) and behavioural (RT and accuracy) measures. These effects were evaluated in relation to the two main hypotheses traditionally proposed to explain them (i.e. the differential rehearsal and inhibition hypotheses). In the study phase, 100 words were presented, each followed by either a remember (R) or a forget (F) instruction. In the test phase, half of the subjects performed a recognition task (old vs new words) and the other half did a categorisation task (concrete vs abstract words). In the study phase, ERPs following Rinstruction differed from those elicited by F-instruction, being more positive in the first 400 msec post-stimulus and becoming less positive from 400 to 800 msec. In the test phase, the effects of the forgetting manipulation were quite clear in the recognition test, affecting both ERP (from 200 to 300 and from 500 to 700 msec) and behavioural measures, but these effects did not appear in the categorisation test. The old/new effects were reflected in the behavioural measures corresponding to both tests, but with the ERP measures they appeared only in the recognition task from 500 to 700 msec. The lack of directed forgetting effects in the indirect test is in line with the differential rehearsal hypothesis. However, the forgetting manipulation produced a short-latency effect (200-300 msec) in the ERPs obtained in the recognition test that is difficult to explain based on that hypothesis. This finding suggests the existence of inhibitory mechanisms, although operating only when conscious access to the previous learning episode is required.