As far as "memory recall is dependent on some chemical in the brain", I think you're right. On that front, the study is interesting as "we found a chemical mech

As far as "memory recall is dependent on some chemical in the brain", I think you're right. On that front, the study is interesting as "we found a chemical mechanism that's wearing out late in the day". There are obviously a host of chemicals involved in memory recall, and a variety of possible time links (e.g. "we forget over time", "we remember things better in the same context we learned them"), so it's interesting to find chemical channels that respond specifically to time of day and not those other factors. As far as "What if its not because of the time of day, just hours awake?", the study doesn't seem to have directly distinguished the two possibilities, which is a bit surprising. The study did show that tasks learned at "night" and tested in the morning performed better than the reverse, so it's not all-around tiredness, but it doesn't seem to distinguish whether rhythm or tiredness affected memory. (The day/night cycle was 12-hour, and the tasks were conducted at 4 and 10 hours into light, so the mice weren't "up late", but it was an analogue of 8-9PM.) I suspect the study avoided differentiating these factors because we're not entirely sure where the differences lie. Broadly, we know at least three components to "tiredness" exist. Adenosine levels rise while awake and fall when sleeping: it's the "awake too long" part of tiredness. Melatonin is produced and broken down rapidly, but looking at (blue) light inhibits its production: that's the "dark out" part of tiredness. But "circadian rhythm" includes more factors we don't totally understand: abruptly adjusting both light and time awake still requires acclimation, and blindfolded people with control over their sleep schedule still respond to light cycles. So for the mice, all of that was piled together to say "late in a normal day, for whatever reason, these chemical factors impede memory recall". Trying to break out the key elements would be fascinating. I'd love to know what happens with different light cycles, or with the abrupt adjustments like waking the mice up early/late on the day of the test, or after giving the mice caffeine or melatonin. And for the graveyard shift, I suspect it's just a mess all around: people in submarines or Antarctica can have wholly-shifted schedules like these mice, but most night-shift workers operate on shifted sleep schedules and an awful mix of imperfectly-adjusted day/night cycles. (Hmm... Now I also wonder what happens to the test if you keep mice awake during 'dark', then let them sleep during 'light'?)