Memory as assessed by recognition and reading time in normal and memory-impaired people with Alzheimer's disease and other neurological disorders.

Three experiments are reported that examine the dissociations in performance on two tests of retention, speeded reading and recognition in young people, home-dwelling and institutionalized elderly people, and people with severe memory disorders. In Experiment 1, subjects read sentences in normal and in geometrically transformed script at initial presentation, 1-2 hr later, and again 4-14 days later. On the latter two occasions, they were required to distinguish old sentences that they had read previously from new ones. In general, the young and elderly subjects could distinguish old from new sentences at the short delay, and all but the institutionalized elderly people could do so at the long delay. Retention as measured by reading speed typically paralleled recognition performance in that those items that were recognized best were read most quickly. The dissociation between these two tests of retention is seen only in people with memory disorders. Although these people could not distinguish old from new items even at short delays (most could not even remember having seen any sentences), their retention as assessed by reading time was similar to that of the other groups. Old sentences were read most quickly, indicating retention of item-specific information, and reading time of new sentences improved, indicating the acquisition and retention of a general skill. Experiment 2 examined what type of item-specific information was retained. Young, elderly, and memory-disordered subjects studied weakly associated word pairs and sentences. A few minutes later they were tested on both recognition and speeded reading of old, new, and recombined pairs and sentences, the last being those in which words from a studied pair or sentence were recombined with words from other such pairs or sentences. The results of both retention tests indicated that young and elderly people could distinguish old from new and recombined items. People with memory disorders, however, again failed on recognition but performed normally on speeded reading. Like the other two groups, they read old items faster than either recombined or new items. In Experiment 3, similar results were obtained even when the word pairs were constructed using randomly associated items. The results of all three experiments suggest that on implicit tests of memory, such as speeded reading, people with memory disorders can be shown to have formed and retained new associations despite failing utterly on explicit tests, such as recognition, that require conscious recollection of a previous episode.(ABSTRACT TRUNCATED AT 400 WORDS)