Three sessions, read · 3 of 3
Session 03: the recording that only measured the headset
One minute, one dead electrode, and a spectrum dominated by movement. The session that produced no reading about me, and one correction worth recording.
Limegreen horse · 1:00 · 15,424 samples · 256.48 Hz
The reading
One electrode is dead and the rest are drowning in movement. F5 has an RMS of 15,146 against 209 to 629 for its neighbours, twenty-six times noisier, swinging 600,000 units peak to peak. That is a floating or disconnected electrode, not a brain region. Exclude it entirely.
The other seven pass their electrical checks, but the recording as a whole is dominated by low-frequency movement. Posterior alpha comes to 1.6% across PO3 and PO4, against 7.3% in the eyes-closed session and a peak of 29.7% at PO4 in the meditation. PO3 in particular is 94.6% delta: almost all of its power sits below 4 Hz, which is drift, not rhythm.
This was recorded on the same day as the mentalMath training, where I noted that I had moved from a static position partway through and that a spike should be treated as an outlier. The data agrees with the note. The movement is visible in the spectrum, and the instinct was right.
A correction worth recording: PO3 sits at a constant offset near minus 28,700 and never crosses zero, which initially looked like a railed amplifier. It is not. Only 0.07% of its samples sit at the extreme, it takes 13,496 distinct values, and the cleanest channel in the eyes-closed session carries an even larger offset. Every electrode on this device has a DC bias and the high-pass removes it. The offset alone is not a fault.
Electrode quality
Band power by electrode
What each band means is on Console Decoded, which explains the five bands and the eight electrode positions in one place rather than repeating them per session.
Table view, every figure above
| Electrode | RMS | Delta | Theta | Alpha | Beta | Gamma |
|---|---|---|---|---|---|---|
| CP3 | 629 | 78.9 | 16.6 | 3.5 | 0.8 | 0.2 |
| C3 | 1,419 | 95.0 | 3.9 | 1.0 | 0.1 | 0.0 |
| F5 | 15,146 | 94.1 | 5.2 | 0.5 | 0.2 | 0.0 |
| PO3 | 235 | 94.6 | 4.4 | 0.8 | 0.3 | 0.0 |
| PO4 | 209 | 83.1 | 12.7 | 2.4 | 1.3 | 0.4 |
| F6 | 582 | 85.4 | 12.0 | 2.1 | 0.5 | 0.1 |
| C4 | 584 | 91.4 | 6.5 | 1.1 | 0.5 | 0.4 |
| CP4 | 533 | 90.8 | 7.1 | 1.1 | 0.5 | 0.3 |
What I set out to record
A short check, on the same day as a long training session, without a particular state in mind. One minute, sitting at a desk, mid-afternoon. It is the least deliberate of the three recordings and I include it for exactly that reason: it is what a session looks like when nothing has been controlled.
What the traces showed
Nothing about me. The reading above gives the figures; what they add up to is that seven electrically sound channels were still unusable, because almost all of their power sits below 4 Hz. Set posterior alpha at 1.6% beside 7.3% in the eyes-closed session and the 29.7% peak in the meditation, and the honest description is not “low alpha”, it is “no measurement”.
That distinction is the whole value of the session. A reading of “low alpha” would have been a statement about me, and this recording cannot support one. It is a measurement of the headset and the room.
What I concluded
The value of this recording is negative and it is still worth having. It is the session that proves the quality check earns its place: without it, this file produces a confident-looking band table, a plausible spectrum, and a completely fictional statement about a brain.
It also confirms my own note from that day. I had written that I moved partway through and that a spike should be treated as an outlier. I could not have proved that from the console, which shows a recording and no way to interrogate it. The spectrum agrees with the note, which is the clearest case on this project of a note I wrote and the data agreeing.
Explore more
This site covers what the documentation doesn't: the things I wish someone had handed me first.