Abstract

We employed a highly demanding experimental associative learning test (the AFE-T) to explore memory functioning in Preclinical Alzheimer’s Disease stage 1 (PreAD-1) and stage 2 (PreAD-2). The task consisted in the learning of unknown object/name pairs and our comprehensive setup allowed the analysis of learning curves, immediate recall, long-term forgetting rates at one week, three months, and six months, and relearning curves. Forty-nine cognitively healthy subjects were included and classified according to the presence or absence of abnormal CSF biomarkers (Control, n=31; PreAD-1, n=14; PreAD-2, n=4). Control and PreAD-1 performances on the experimental test were compared by controlling for age and education. These analyses showed clear learning difficulties in PreAD-1 subjects (F=6.98; p=0.01). Between-group differences in long-term forgetting rates were less notable, reaching statistical significance only for the three-month cued forgetting rate (F=4.83; p=0.03). Similarly, relearning sessions showed only statistical trends between the groups (F=3.22; p=0.08). In the whole sample, significant correlations between CSF Aβ42/tau ratio and the AFE-T were found, both in the total learning score (r=0.52; p<0.001) and in the three-month cued forgetting rate (r=-0.38; p<0.01). Descriptive subanalyses involving PreAD-2 suggested greater learning and recall difficulties in these subjects when compared with the PreAD-1 group. The present results suggest that explicit learning difficulties when binding information could be one of the earliest signs of the future emergence of episodic memory difficulties on the Alzheimer’s disease continuum. Our findings indicate that the AFE-T is a sensitive test, capable of detecting subtle memory difficulties in PreAD-1.