The implication was that the brain uses glucose in its operation, and requires a certain level of glucose to function at full capacity. At least one cause of the brain fog we get after a full day's focused work is that we're low on fuel, which happens to be easy to fix. Unfortunately, it has other effects too that we might not want so much. I've been meaning to experiment on myself a bit with some glucose tabs after a long day of code, but haven't gotten around to it yet.
(*) I think that was in "Thinking, Fast and Slow" but I'm not certain now, too many similar books in between.
As I wrote that, I did remember a bit of evidence that the glucose was actually beneficial - namely that they measured performance on the problems over time. At the end of the first set, both groups were showing approximately equal degraded performance. Upon resuming on the second set, the group given glucose was back up to nearly the starting level of proficiency, while the group given no glucose was at the same level they stopped at on the first group.
Anecdotal feedback - they work great for short bursts of activity. I use them when playing multiplayer games at night. Given my opponents are usually young and fuelled on caffeinated sodas I need some kind of edge after getting home from work. When I stomp a lobby on team deathmatch it's usually after popping a couple of glucose tabs. A good low-GI meal does much the same, but of course isn’t as convenient.
Consider the ''eugeroic'' ("Wakefulness Enhancer") class, described as "unproven primary mechanisms but proven efficacy". Bit scary, not something to have a habit around, but widely available through the internets.