Visual delayed recall as a predictor of conversion to Alzheimer's disease in patients with amnestic mild cognitive impairment

Abstract

Background:Amyloid imaging studies have shown early and focal deposition in preclinical PS1 and PS2 carriers among ADAD families. We investigated whether amyloid burden was associated with performance on cognitive tests, includingmeasures putatively sensitive to cognitive functions.Methods:We studied n1⁄411 cognitively normal (CN) ADAD family participants (age 22-44 years) and n1⁄43 with MCI (age 35-48). Participants included 8 PS-1 and 4 APP mutation-carriers, and 2 non-carriers. Participants completed PiB-PET imaging and standard neuropsychological tests, including measures of episodic memory, visuospatial construction, language, attention and executive functions. As well, participants completed a striatal cognitive battery including measures tapping 1) procedural learning (serial reaction time task of implicit learning; Rotary Pursuit task; mirror reading task); 2) risk-reward decision making (Iowa Gambling Task); and 3) psychomotor speed/dexterity (finger tapping; grooved pegboard). PiB-PET imaging data were acquired 50-70minutes after injection. Striatal PiB bindingwas computed as the SUVR (cerebellum reference) of a hand-drawn anterior-ventral striatum region of interest. Precuneus was a comparison ROI.Results: PiB SUVR in the anterior-ventral striatum ranged from 1.01 to 3.41 (mean 1.83; SD 0.78); in the precuneus PiB SUVR range from 1.26 to 3.31 (mean 1.76; SD 0.53). Age was correlatedwith PiB (r1⁄4.67, p<.01) but not in the precuneus. AmongCN participants only, partial correlations (age-controlled) between PiB and cognitive measures were significant (p .05, two-tailed) for finger tapping (r1⁄4.65); mirror reading (r1⁄4.79) and Iowa Gambling Task (r1⁄4.74). Traditional neuropsychological tests also correlated with PiB in CN participants, including MMSE, Boston Naming Test, Block Design and CERAD Word List Learning (partial r’s .63 to .78). In CN and MCI participants combined (n1⁄414), only finger tapping (non-dominant hand) and the IowaGambling Task were correlated with PiB (see Figure). A similar pattern of correlations was observed between precuneus PiB SUVR and cognitive measures, as well as with traditional neuropsychological measures. Corresponding analyses using pons as reference are underway and will be presented. Conclusions: Cognitive measures tapping behavioral functions of the striatum may be useful to explore in larger studies of preclinical AD.