Fig. 5: Time series associations between physical activity and sensor-glucose outcomes by insulin resistance status.

a Interaction effects plot for step counts and SSPG status (IS/IR) on CGM. Effects of step count and SSPG status on mean glucose values were assessed through linear models at each time window, permuted as in the Pearson correlation analysis. We split the time-series of step counts into seven-time windows of the day. The X axis indicates the standardized step counts of a specific time window, and the Y axis represents the corresponding glucose values up to the next 48 hours. The orange line represents IR, and the green line IS, where interaction effects were considered significant (asterisk) if multiple testing-adjusted q value < 0.01. b–d Shifted correlation analysis plot between step count and CGM in different time windows of the day ((b) Insulin-sensitive, (c) Insulin-resistant; (d) overall cohort). The color gradation represents correlation coefficients ranging from −0.5 (negative correlation) to 0.5 (positive correlation). CGM continuous glucose monitoring. e 2D scatter plot that shows the distribution of CGM as a function of HR/HRmax for all participants over a shared period. Each point represents a data entry. The scatterplot shows a noticeable pattern between HR/HRmax and CGM values for all participants.