Cognitive dispersion as a function of cognitive reserve: influence on cognition and functional connectivity in a healthy middle‐aged population

Background Cognitive dispersion index (CDI), a measure of intraindividual variability across neuropsychological tests, is considered a sensitive marker of prodromic stages of Alzheimer’s Disease (Halliday et al. J. Intell 2018;6(1):12). CDI has been negatively associated with cognitive reserve (CR), another factor that influences the risk of dementia (Stern et al. Alzh&Dement 2020;16(9):1305‐1311). Nevertheless, how these two elements interact with each other and are linked with cognition and brain state (e.g., functional connectivity), remains unclear. Our aim was to investigate whether CDI has a differential influence on cognitive performance (CP) and on resting‐state functional connectivity (rsFC) as a function of CR level, in a healthy middle‐aged sample. Method Five hundred forty‐four healthy volunteers (aged 53.11±7.10 years; 269 female) from the Barcelona Brain Health Initiative cohort (BBHI, https://bbhi.cat/en), were stratified by CR (median = 14) into High CR (HCR, N = 230) and Low CR (LCR, N = 314) using the cognitive reserve questionnaire (CRQ; Rami et al. Rev Neurol 2011;52(4):195‐201). CDI was extracted by intraindividual standard deviation method (Costa et al. Clinic Neuropsychol 2019;33(2):369‐389) and CP was calculated as a global cognition composite score including episodic memory, speed of processing and executive function tests. Functional magnetic resonance imaging (fMRI) was acquired, and measures of rsFC within networks were computed through Shirer atlas (Shirer et al. Cerebral Cortex 2012;22(1):158‐165). Statistical analyses were based on univariate ANCOVA GLMs, and partial correlations adjusted by age and gender. Result Though no significant group interaction was found, a negative correlation between CDI and CP was identified for the HCR group (r = ‐0.170,p = 0.010; Fig.1A). On fMRI there was a significant group difference (HCR vs. LCR) in the association between CDI and rsFC in anterior salience network (aSN: F = 5.005,p = 0.026), where only HCR participants showed a significant negative association between CDI and aSN rsFC values (r = ‐0.187,p = 0.010; Fig.1B&C). Conclusion Our findings suggested that CDI has a clearly measurable effect on HCR participants, both in terms of its effect on CP and on rsFC. These results highlight the value of considering CDI measurements when performing neuropsychological assessments in middle‐aged individuals, particularly in those with high CR estimates.