OBJECTIVES: To evaluate the predictive utility of proactive semantic interference (PSI) and failure to recover from proactive semantic interference (frPSI) deficits on the longitudinal everyday functional decline on the Clinical Dementia Rating Sum of Boxes (CDR-SOB) among older adults with amnestic mild cognitive impairment (aMCI). DESIGN: Longitudinal prospective cohort study. SETTING: 1Florida Alzheimer's Disease Research Center (1FLADRC). PARTICIPANTS: 97 older adults aged 54 to 98 years who were diagnosed with aMCI following a baseline evaluation. The average age was 71.9 years, 51 % male, average education 15.7 years, and 56 % Hispanic. The mean MMSE score was 28.0. Participants were followed annually for 3 to 4 visits with a mean follow-up time of 38.9 months (range 22.7 to 70.3 months). MEASUREMENTS: CDR-SOB were obtained at each visit and the latent growth curve trajectory of CDR-SOB was estimated. The associations between PSI and frPSI and the growth curve trajectory of CDR-SOB were examined. RESULTS: The growth curve model that best fits the CDR-SOB trajectory was a linear form and included the fixed and random effect of intercept and slope of time. After adjusting for age, sex, education, Hispanic background, Hopkins Verbal Learning Test (HVLT) immediate and delayed recall, and amyloid positivity, frPSI (β = -0.134, se=0.04, p <
0.01) remained statistically significant in predicting a steeper slope on the trajectory of decline in CDR-SOB. CONCLUSIONS: frPSI at baseline predicted the rate of everyday functional decline over time among participants with aMCI regardless of amyloid status and demonstrated its utility of longitudinal prediction of change in CDR-SOB.