Recognition memory for words and faces in the very old.

Abstract

The assessment of very elderly people is hindered by a scarcity of normative and reliability data for non-verbal memory tests. We tested the suitability of Warrington's Recognition Memory Test (RMT) for use with the elderly. The RMT consists of verbal (Recognition Memory for Words, RMW) and non-verbal (Recognition Memory for Faces, RMF) subtests. The facial recognition test was used in the standard format and a Dutch-language version of the word recognition test was developed using low frequency (10 or less/million) monosyllabic words. Eighty-nine subjects, varying in age from 69 to 93, were tested with the RMF. Means and SD are provided for three age groups (69-79, 80-84 and 85-93). Forty-five consecutive subjects were tested both with the RMW and the RMF. Recognition memory for words was better than recognition memory for faces in this sample. Moderate correlations (0.30-0.48) were found between RMT and WAIS Vocabulary and Raven's Coloured Progressive Matrices scores. Warrington's RMT was well tolerated, even by very elderly adults. The standardization data for the elderly over 70 add to the usefulness of this test of verbal and non-verbal episodic memory.