Day 4c: contact check before the first night
In short
- Dry: all four sensors Poor, 0% on target.
- After dampening: AF7 and AF8 (forehead) went to Good, 100%. TP9 and TP10 (ears) stayed Poor, 0%.
- Removing hair from under the ear clips did not fix TP9/TP10 — the hair guess was wrong.
- AF7 dropped from 100% back to 0% within under 10 minutes, replicating Day 1’s ‘contact fades as it dries’ finding a third time.
- AF8 was the one sensor that held at Okay/50-60% across repeated checks — unexplained, worth tracking, not worth guessing at yet.
- Running the first night anyway, on whatever contact is available at lights-out, to see the pattern over 7-8 hours instead of minutes.
On this page
What this was
A contact check right before starting tonight’s first whole-night recording with the night recorder built today.
Dry vs. dampened
| Sensor | Dry | After dampening |
|---|---|---|
| TP9, left ear | Poor · 201 µV · 0% | Poor · 187 µV · 0% |
| AF7, left forehead | Poor · 190 µV · 0% | Good · 21 µV · 100% |
| AF8, right forehead | Poor · 279 µV · 0% | Good · 17 µV · 100% |
| TP10, right ear | Poor · 167 µV · 0% | Poor · 158 µV · 0% |
Dampening fixed both forehead sensors outright. It did nothing measurable for either ear sensor. The app’s own hint — “make sure no hair is under the sensor” — points at the ears specifically, so hair was checked next.
Hair removed, contact re-checked twice more
| Sensor | 20:24 | 20:29 |
|---|---|---|
| TP9, left ear | Poor · 253 µV · 0% | Poor · 245 µV · 0% |
| AF7, left forehead | Poor · 128 µV · 9% | Poor · 178 µV · 0% |
| AF8, right forehead | Okay · 61 µV · 50% | Okay · 51 µV · 60% |
| TP10, right ear | Poor · 273 µV · 0% | Poor · 128 µV · 9% |
Clearing the hair did not fix the ears — TP9/TP10 stayed at 0% or close to it. Hair wasn’t the (sole) cause there; the earlier guess was wrong. AF7, which had been Good at 100% right after dampening, had already dropped back to 9% and then 0% five minutes later — a third replication of Day 1’s “contact fades as it dries,” now timed at under 10 minutes rather than ~15. AF8 is the one outlier: not Good, but holding at Okay/50-60% across both checks, more stable than every other sensor. No explanation for why yet — worth tracking whether that holds, rather than guessing.
This isn’t a new problem
Day 1 already found that a dampened forehead goes back to floating within about 15 minutes, once the water dries. Tonight’s test is a replication of that, not a new finding: dampening gets you a green light for a few minutes, not for a night.
The open question for tonight
A first contact check can only show a snapshot. The real question for a whole-night recording is what happens to AF7/AF8 contact over 7-8 hours of sleep, not 15 minutes awake at a desk: does it fail completely, fail in patches that the 30-second epoch quality label can catch, or drift slowly enough that only part of the night is unusable? Recording tonight anyway, on whatever contact is available at lights-out, specifically to get that answer — a night with some bad epochs marked is more useful than no night at all.
Why this matters beyond this project
This is the same question Dreamento’s goodContact flag and the lab’s ezscore-f/eegFloss tools exist to handle for ZMax recordings: non-expert, self-applied dry/semi-dry EEG contact is not a one-time pass/fail, it’s a moving target across a night. A system that only checks contact once, at the start, misses this entirely.