A selective impairment of recollection following medial-temporal lobe damage: behavioural and electrophysiological data

Wilding, E. L., and Aggleton, J. P.
School of Psychology, Cardiff University, UK
E-mail: wildinge@cardiff.ac.uk

Patient (KN), a 44 year old man, suffered 50% bilateral shrinkage of the hippocampus and amygdala following meningitis. In contrast to his preserved IQ (WAIS PIQ 116, VIQ 136) he shows a severe, persistent loss of episodic memory. This is demonstrated by his failure to recall prose passages, word lists, and complex figures, so that on the WMSr he has an Attention Concentration Index of 123, yet a General Memory Index of 88, and Delayed Recall Index of 50 (at floor). Likewise, on the recall components of the Doors and People Test he performs extremely poorly (name recall 1%ile, shape recall 5%ile). In contrast, KN appears more normal on tests of recognition memory (e.g. Warrington Recognition Memory test, words 48/50, 75%ile; Doors and People Test, name recognition 25%ile). Consistent with this he performs at normal levels on the Calev recognition task (20/24) but is very impaired on the Calev recall task (7/24). Performance scores on these tasks are expected to be comparable.

Evidence from two different sources indicates that KN is able to perform recognition memory tasks on the basis of familiarity. First, while scalp-recorded event-related potentials (ERPs) differentiated correctly identified old and new test items, the ERP signature of recollection did not form part of this differentiation. Second, estimates of familiarity and recollection derived from an ROC analysis reveal a strong bias toward using familiarity. These findings accord with dual-process models of recognition memory in which familiarity and recollection are distinct components. The former may depend on the parahippocampal region while the latter depends on the hippocampus.