Recall and recognition of verbal paired associates in early Alzheimer's disease.

The primary impairment in early Alzheimer's disease (AD) is encoding/consolidation, resulting from medial temporal lobe (MTL) pathology. AD patients perform poorly on cued-recall paired associate learning (PAL) tasks, which assess the ability of the MTLs to encode relational memory. Since encoding and retrieval processes are confounded within performance indexes on cued-recall PAL, its specificity for AD is limited. Recognition paradigms tend to show good specificity for AD, and are well tolerated, but are typically less sensitive than recall tasks. Associate-recognition is a novel PAL task requiring a combination of recall and recognition processes. We administered a verbal associate-recognition test and cued-recall analogue to 22 early AD patients and 55 elderly controls to compare their ability to discriminate these groups. Both paradigms used eight arbitrarily related word pairs (e.g., pool-teeth) with varying degrees of imageability. Associate-recognition was equally effective as the cued-recall analogue in discriminating the groups, and logistic regression demonstrated classification rates by both tasks were equivalent. These preliminary findings provide support for the clinical value of this recognition tool. Conceptually it has potential for greater specificity in informing neuropsychological diagnosis of AD in clinical samples but this requires further empirical support.

Recall and recognition measures of paired associate learning in healthy aging.

Associate-recognition has received little attention as a potential clinical tool for detecting early Alzheimer's disease (AD). As an important preliminary stage to investigating the paradigm's diagnostic utility, we designed and administered a verbal associate-recognition task to healthy elderly participants (n = 62) and compared their performance to that on traditional cued-recall PAL. In both test conditions, the stimulus list comprised of a mixture of highly imageable and less imageable word pairs. Overall, performance on the associate-recognition task was superior to that on the cued-recall analogue. This 'recognition advantage' was not attributable to the higher baseline or chance guessing rate in the associate-recognition condition, as the size of the recognition advantage varied across learning trials and stimulus imageability. In comparison to performance on the imageable stimuli, performance on the less imageable stimuli was poor in both associate-recognition and cued-recall conditions. Across the delay, performances were more likely to drop in the cued-recall condition than the associate-recognition condition. These results suggest that verbal associate-recognition may be clinically efficacious and better tolerated in elderly populations than traditional cued-recall paradigms. Although these results are encouraging, further research is required to examine the utility of associate-recognition in clinical populations, particularly early AD.