Open problems · P10
Recovery near the glycaemic edge
The mountainside test in MIDUS used heart rate after a thinking task, a weak proxy for the edge that diabetes stands for. The system-matched test is glucose: after the same meal or glucose load, does glucose take longer to return to its own baseline in prediabetes and type 2 diabetes than in normoglycaemia, and does the return slow more steeply than the peak rises as HbA1c approaches the diagnostic threshold?
Why it matters
Glucose regulation has a well-defined edge, measured routinely, and continuous glucose monitors record the whole return after every meal. If recovery time lengthens near that edge, it is the clearest test of critical slowing in people; if only the peak changes, the loss-of-gain reading on the mountainside page gains support.
What exists
- CGMacros (PhysioNet): healthy, prediabetes and type 2 diabetes groups with timestamped, partly standardised meals and two monitors; CC BY-NC-SA licence.
- Hall and colleagues (PLoS Biology 2018): standardised meals with continuous monitoring, mostly normoglycaemic, with windows ending a few hours after the meal.
- Metwally and colleagues (Nature Biomedical Engineering 2025): oral glucose tolerance tests with continuous monitoring, groups by HbA1c.
- NHANES records glucose only fasting and two hours after a glucose load, which gives a level, not a return time, and it excludes people on diabetes drugs.
A first step
Using CGMacros, align each standardised breakfast and lunch, fit each person's return to their own pre-meal level, and compare the time to return across the three glycaemic groups with the same response-robust approach as the mountainside page (return at equal peak). State the smallest slowing of interest before looking.