There is a group they label "control", but they simply get fed flicker at frequencies they predict to be non-optimal. They did measure differences, so the expe

There is a group they label "control", but they simply get fed flicker at frequencies they predict to be non-optimal. They did measure differences, so the experiment does have power (unlike an experiment with no control group at all). It's a valid experiment, on its own terms. But without a 0-flicker control group we can't absolutely say whether it improves learning over the 0-flicker baseline we all experience in daily life, only that some flicker frequencies/phases are better than others (and by a wide margin). This is still an interesting result! There may be other results I'm not aware of that argue against my discombobulation theory. I'm sure the experimenters have a deeper grasp of the field than I do.