The participants (20 men without known health conditions) slept at the laboratory 3 times, each 1 week apart, and consumed different drinks 15 minutes before bedtime as follows:

Familiarization night: water
Test night 1: one of the test drinks
Test night 2: the other test drink

The high-glycemic-index drink contained 50 grams of glucose (GI = 100), and the low-glycemic-index drink contained 50 grams of isomaltulose (GI = 32). Four hours before bedtime, the participants consumed a standardized meal.

Overall, the different drinks did not affect sleep and memory outcomes.

The low-glycemic-index drink increased slow-wave sleep (+5%; medium effect size) and some memory scores compared to the high-glycemic-index drink, but only when data involving the participants who consumed the low--glycemic-index drink on the second test night were analyzed. The authors suggested that a residual first-night effect^[1] may have hindered the detection of these changes.

Are low-glycemic-index drinks better for sleep and memory? Original paper

Are low-glycemic-index drinks better for sleep and memory? Original paper