Dynamic associations between glucose and ecological momentary cognition in Type 1 Diabetes

Abstract

Type 1 diabetes (T1D) is a chronic condition characterized by glucose variability due to autoimmune destruction of insulin-secreting pancreatic beta-cells. Prior research (linking greater glucose variability to cognitive impairment and adverse health outcomes) and recent technological innovations (continuous glucose monitoring [CGM], cognitive ecological momentary assessment [EMA]) laid a foundation for the current study, which aimed to advance understanding of between- and within-person influences on physiological and cognitive processes relevant to health in T1D. We obtained intensive longitudinal, naturalistic measurements of cognition (EMA) and glucose (CGM) in N=199 adults with T1D. Given reduced cognitive performance during periods of low (hypo) and high (hyper) glycemia, our analyses focused on the strength of within-person quadratic associations between glucose and cognition—operationalized as the rate of change in the slope of the glucose-cognition relationship, hereafter, glucose acceleration. First, we used multilevel modeling to obtain group (i.e., aggregated across participants) and individual (i.e., specific to each participant) estimates of glucose acceleration. Next, we used machine learning to identify between-person predictors of individual estimates of glucose acceleration. Finally, we built personalized models to contextualize individual estimates of glucose acceleration in relation to time-varying processes. Group estimates of glucose acceleration were significant for processing speed but not sustained attention. Specifically, processing speed was fastest at glucose concentrations slightly above individuals’ means, regardless of absolute levels of the means. Between individuals, estimates of glucose acceleration varied in relation to several clinical and demographic variables, including age, microvascular complications, and percent time in hypoglycemia (< 70 mg/dL). Within individuals, fluctuations in glucose and processing speed were related to prior night’s sleep duration, self-reported assessment context (e.g., noise), and self-reported stress. Results generate testable insights about shared (across individuals) factors that impact cognitive fluctuations. They also highlight opportunities to support personalized medicine by increasing understanding about unique (person-specific) influences on cognitive fluctuations in T1D.

Assessing dynamic cognitive function in the daily lives of youth with and without Type 1 Diabetes (T1D): a pilot and feasibility study (Preprint)

“…Next steps include using EMA in a fully powered study to determine the relationship between short-term glycemic control and cognition in youth with T1D. approaches in adults with T1D, significant glucose fluctuations were shown to be related to slower processing speed in the moment [14]. However, whether this relationship or others are seen in youth with T1D is unknown.…”