Developing a Spatial Navigation Screening Tool Sensitive to the Preclinical Alzheimer Disease Continuum

Abstract

Objective: There remains a need for a non-invasive and cost-effective screening measure that could be administered prior to the provision of a lumbar puncture or positron emission tomography scan for the detection of preclinical Alzheimer disease (AD). Previous findings suggest that a hippocampally-based spatial navigation task may be effective for screening individuals for the preclinical AD continuum (i.e., low cerebrospinal fluid (CSF) Aβ42). Unfortunately, this task took 1.5-2 hours to administer, which would be time-prohibitive in a clinical setting. Therefore, the goal of this study was to compare psychometric properties of six spatial navigation-related tasks in order to take the next steps in developing a clinically appropriate screening measure.

Methods: Psychometric properties (i.e., reliability, diagnostic accuracy, validity) of a modified version of the cognitive mapping task, two binding tasks, a visual perspective taking task, and self- and informant report versions of a questionnaire were examined in a sample of 91 clinically normal (CN) individuals. CSF Aβ42 and ptau181 were available for 30 individuals.

Results: The learning phase of the cognitive mapping task and the self-report questionnaire were sensitive to identifying individuals in the preclinical AD continuum (93% and 87% sensitivity, 60% and 67% specificity, respectively). These two measures also demonstrated good test-retest stability (intraclass correlation coefficients = .719 and .838, respectively) and internal consistency (Cronbach's αs = .825 and .965, respectively).

Conclusions: These findings suggest that a self-report questionnaire and aspects of a cognitive mapping task may be particularly appropriate for development as screening tools for identifying individuals in the preclinical AD continuum.

Keywords: amyloid; cognition; hippocampus; ptau; reliability; validity.

Fig. 1.

A) Cognitive mapping task: Learning phase; B) Cognitive mapping task: Retrieval phase; C) Cognitive mapping task: Landmark identification; D) Self-report questionnaire. Data represent estimated marginal means (controlling for covariates; see text for details).

Fig. 2.

Confirmatory factor analysis for one-factor model. Values at unidirectional arrows represent factor loadings with the 95% confidence intervals in parentheses. Bidirectional arrow between the learning and retrieval phases represents correlated measurement error. CM = cognitive mapping; SE = standard error.

Fig. 3.

Confirmatory factor analysis for three-factor model. Values at unidirectional arrows represent factor loadings with the 95% confidence intervals in parentheses. Bidirectional arrow between the learning and retrieval phases represents correlated measurement error. Values between factors represent the correlations between factors. L-N = Letter-Number; CM = Cognitive Mapping; SE = Standard Error.