Advancing the Modified Face Name Associative Memory Exam in Cognitive Aging Research: Insights into Connectomic Correlates and Task Reliability

Yilin LiuMark H SundmanDalen HinderakerAllison Yu-Chin ChenJacob M GreenLisbeth G HaaheimHannah M SiuCatherine JezercKaitlyn LaiCarol Y ChenParker GussYing-hui Chou^*

• Brain Imaging and TMS Laboratory, University of Arizona, Tucson, United States

The ongoing shift in research efforts towards earlier stages of Alzheimer’s disease (AD) continuum underscores the need for more sensitive neuropsychological assessments to detect subtle cognitive deficits. Traditional cognitive outcome assessments (COAs), tailored for assessing deficits in AD or Mild Cognitive Impairment (MCI), may lack sensitivity to capture preclinical cognitive dysfunction associated with AD pathology. The Face-Name Associative Memory Exam (FNAME), a cross-modal task that integrates verbal and non-verbal memory, offers significant advantages in this regard. The FNAME demonstrates high sensitivity, and task performance is reportedly correlated with amyloid-β burden across the AD continuum, even in asymptomatic older adults. This manuscript reports on two experiments, broadening insights into this promising COA. Experiment 1 (descriptive observational, repeated-measures design) (N=85) evaluates the alternate form reliability of a modified FNAME (mFNAME) by serially administering eight distinct versions of the task, revealing good reliability for mFNAME metrics and the absence of significant practice effects. Experiment 2 (cross-sectional observational design) (N=32) examines structural and functional network topology to investigate neural correlates of mFNAME performance in non-demented older adults. Graph theoretical analyses revealed significant associations between mFNAME performance and network properties like global efficiency, local efficiency, and system segregation in the default mode network (DMN) and medial temporal network (MTN). Subsequent analyses into more granular elements of the MTN and DMN revealed latent variables accounting for up to 44% of the covariance in mFNAME performance. These findings deepen the understanding of the FNAME’s psychometric properties and the neural correlates underlying task performance, providing insights into its utility as a sensitive COA early in the continuum of AD and related dementias.