Filing, not resting
During the day you accumulate experience in a form that is fast to write and unreliable to keep. Overnight, a good deal of that gets reprocessed and moved into storage that lasts. Researchers call it consolidation, and the useful thing to understand about it is that it is an active process with a scheduled slot.
Which has a consequence most people have never been told. A bad night does not simply make tomorrow harder. It degrades your hold on the day that came before it. The material was there; the shift that was meant to file it got cancelled.
Two stages, two jobs
Deep slow-wave sleep, concentrated in the first half of the night, appears to do most of the work on facts and events — names, conversations, where you put things. REM sleep, which dominates the second half, is more associated with procedural learning and with the emotional weight attached to a memory.
That split explains a common pattern. Somebody who goes to bed at the usual hour but wakes at four has lost disproportionately more REM than deep sleep. Somebody who goes to bed three hours late and wakes at the usual time has lost the other end. Both are short on sleep; they are short on different things.
What actually interferes with it
- Alcohol in the evening. It shortens the time it takes to fall asleep and then suppresses REM for much of the night. People who use it as a sleep aid are trading the half of the night they notice for the half they do not.
- Untreated apnoea. Repeated interruptions fragment exactly the deep sleep that consolidation depends on, which is part of why it shows up in cognitive research at all.
- An inconsistent schedule. The system runs on a clock. Going to bed at wildly different times week to week is closer to permanent mild jet lag than to flexibility.
- Late caffeine. Its half-life runs around five hours, which makes a four o'clock coffee a meaningful presence at nine.
The thing worth testing on yourself
If you suspect sleep is behind a run of forgetfulness, the experiment is cheap. For two weeks, hold the wake time fixed — the same time every day, weekends included — and let bedtime fall where it wants. A fixed wake time is a stronger lever on the body clock than a fixed bedtime, and it is easier to keep.
Then compare against whatever you wrote down before you started. If you wrote nothing down before you started, you will not be able to tell, which is the single most common reason these experiments produce no useful answer.