The enigma of long-term forgetting

Memory symptoms are common among people with epilepsy (pwe). Their complaints are often corroborated by impairments on traditional tests of memory which assess learning and retention over delays of up to 30 min. Yet a substantial minority of patients report memory problems which are not identified on standard testing. This discrepancy has sometimes been ascribed to mood disorders. There is, however, an alternative, or complementary, explanation: that established methods of memory testing fail to detect some genuine memory problems. Standard tests, for example, seldom assess autobiographical memory or retention of new information over delays greater than 30 min. A growing literature reports patients with epilepsy, usually arising from the temporal lobes (TLE), who show normal or near normal retention of information over 30 min delays, but increased forgetting over longer delays of days or weeks. This phenomenon has been described as ‘accelerated long-term forgetting’ – or ALF. Whilst there is growing evidence that ALF is measurable and clinically relevant, several key questions remain: (i) Is ALF a widespread phenomenon among pwe? (ii) Does ALF help to explain discrepancies between subjective memory complaints and results of memory testing? (iii) What are the underlying mechanisms of ALF? We consider these questions in turn. (i) Is ALF widespread among pwe? There is now good evidence that ALF occurs among pwe, particularly in TLE. It appears to be especially common in transient epileptic amnesia (TEA), a form of TLE in which episodes of amnesia are the principle manifestation of the seizure disorder. However, some studies in patients with TLE have failed to identify ALF. The explanation for these discrepancies is probably methodological. In studies which train subjects to a high learning criterion, ceiling effects and overlearning can obscure early forgetting, giving rise to the false attribution of ALF; conversely, in studies in which patients and controls start off with differing levels of retention, their forgetting curves are difficult to compare, and ALF may go undetected. Recent studies which take these methodological problems into account have confirmed the occurrence of ALF in TLE. There is much less evidence regarding its occurrence in generalised epilepsy, though an early study suggested that it occurs following ECT. (ii) Does ALF help to explain discrepancies between perceived and objective memory? In people with TEA, two findings suggest that ALF is relevant to memory complaints. Butler et al. reported more marked ALF among patients who spontaneously reported memory loss over weeks compared to those who did not. Butler et al. found that ALF was a significant predictor of perceived memory on the Everyday Memory Questionnaire. However in other epilepsies, the picture has not been so clear. Muhlert et al. found no