Clinical Manifestations

Background: Type 2 diabetes (T2D) and older age are well-known risk factors for dementia. Indeed, there is evidence that older adults not diagnosed, but at-risk for T2D can show early signs of cognitive decline, further exacerbated by excessive body weight or high blood glucose levels. Such a finding would have implications for early treatment strategies; however, the evidence is still sparse. We examined the correlation of risk factors for diabetes with cognitive function in older adults at-risk for T2D using a battery of touchscreen tasks translated from their rodent versions, as well as traditional pen-to-paper cognitive tests.

Method: Sixty-five older adults (69.39 ± 10.31 years old, 68% female) at-risk for T2D (BMI ≥ 25 kg/m2, hemoglobin A1c ≥ 6.0%, CANRISK score ≥ 21) completed 3 novel touchscreen tasks: paired associative learning (PAL) (learning and object-in-location memory), progressive ratio (motivation), and trial unique, non-matching to location (TUNL) (spatial pattern separation and working memory). They were also tested on pen-to-paper cognitive tests: trail-making (task switching), Stroop (selective inhibition), and digit span (working memory). A correlation analysis was performed between BMI or HbA1c and cognitive performance. Performance on touchscreen tasks was analyzed using a repeated measures one-way ANOVA.

Result: Higher HbA1c levels were correlated with lower digit span scores (r2 = 0.12, p = 0.011). There was no correlation between breakpoint (motivation level) and BMI or HbA1c during the progressive ratio task (r2 = 0.0, p = 0.38). Interestingly, this population performed at chance level (59.7 ± 5.3% accuracy) on the PAL task, indicating they were unable to learn object-location paired associates. When manipulating the spatial similarity in distance between stimuli during the TUNL task, older adults at risk for diabetes were 10% lower in accuracy when stimuli were close together compared to further apart (p<0001). Participants also responded more slowly to stimuli at choice during the TUNL task during the heaviest working memory load condition (p = 0.003).

Conclusion: Older adults at-risk for T2D exhibit decreased performance on tasks with higher demands on spatial and working memory. Future research will compare performance on all tasks to healthy age-matched controls and reassess performance after a six-month exercise intervention.