This effect was quite substantial. It was over 100 minutes greater increase than in the placebo, for both healthy subjects and insomniacs. Total Sleep Time (TST) increased in the placebo group by 1 hour in healthy subjects and 2 hours in insomniacs. Of course, most of that was probably not placebo, but rather getting used to the sleeping conditions. As I said, ashwagandha increased sleep time 100 minutes more than that. In real world conditions, I think it's reasonable to expect an hour increase after 8 weeks of ashwagandha. Indeed, this effect implies it's not just an antioxidant, as antioxidants didn't have an effect that significant on TST.

Edit: actually, the increase in TST appears not to be that great, assuming the graph is to be trusted. Given that TST was reported as such low numbers, I had assumed it must be change in TST from baseline. However, the graphs show a 20 or so minute increase. The reported numbers in the table are all messed up. Sleep onset latency is reported as 85 in the healthy group taking ashwagandha. Time in bed is reported as 20 minutes. None of it makes sense, and none of it matches up with the graphs. That table isn't trustworthy. Best to go by the graphs, which show decent but more realistic benefit.

Benefits accrued throughout the 8 weeks. I recall reading on Longecity forum that ashwagandha takes a month for benefits to kick in. This study demonstrates that benefits continue to increase over two months. I suspect they continue even further than that.

Interestingly, this is pretty similar to the two placebo controlled studies on antioxidants for sleep. Thus, I wonder of the benefits of ashwaganha extract are largely antioxidant capacity. This would be a bit surprising because the ORAC of dried ashwaganda is just slightly above raw pinto beans. Based on the recommended doses, the extract isn't vastly more potent than the whole root. Though, this comment saying that the Withanolide/Withaferin A (edit: withaferin A is purportedly cytotoxic) reside mostly in the leaves has greatly confused me. Either the extract has more antioxidant activity than I realize (directly or indirectly), or the benefits come primarily from the purported mechanisms of ashwagandha (which include cortisol reduction and GABAergic activity). Edit: the full text mentions a 15 to 1 extract ratio, which is enough to put the antioxidant mechanism back on the table. It's probably a partial explanation, but after seeing the full text I think the benefits are too great to be simply from antioxidants.

I see no mention of the time of day of administration. I'm assuming it was in the morning, which contrasts with the near bedtime dosing in the antioxidant studies. If I later find out that antioxidants in the morning don't help with sleep, then that will suggest ashwagandha works by other mechanisms. However, I expect antioxidants at any time of day help with sleep. Nonetheless, I'm not discounting that ashwagandha may work by other mechanisms.

This was most likely morning and bedtime. 300 mg times 15 equals 4.5 grams. In other words, they administered the equivalent of 9 grams ashwagandha root per day.

This is likely the standard. It is a sufficient ratio to give it a high ORAC, meaning the antioxidant mechanism is back on the table. Nonetheless, the benefits appear greater than that of mere antioxidants.