Program A · 5 sections
What sets alpha's frequency and shape, how it fluctuates, and whether anything about it drifts over months.
Replicated within archiveAlpha runs faster in every LSD session, 10.3–10.9 Hz against an ordinary median of 9.8 Hz; nothing about it persists across the months, and thalamocortical dysrhythmia is absent.
06 — The pacemaker Retrospective · exploratory
Analysis basis 56-recording cohort; 6 LSD vs 37 ordinary for the headline contrast; Resting baseline 003 part 2 (2.4 clean minutes) kept with placeholder values · dataset release 2026.10.10 · last recomputed in analysis release 4.26 · current
Strip away the 1/f background from any of the 56 cohort recordings and the same structure appears: a single sharp peak near 10 Hz — the posterior alpha oscillation, conventionally attributed to the loop between thalamus and visual cortex (seen under a common average reference it is a front–back pattern of the whole head, §22). Its frequency, sharpness and harmonic content are the through-line of this entire archive.
Median posterior center frequency across the 37 non-psychedelic cohort recordings (corrected labels; the twelve ordinary recordings added in October included), range 8.9–10.25 Hz.
Median across the 19 acute sessions, range 10.00–10.89 Hz. Ten of the nineteen sit above the fastest non-psychedelic recording (10.25 Hz); the other nine, at 10.00–10.23 Hz, overlap the top of the non-psychedelic range.
All six LSD sessions, the newest at 10.89 Hz once its flicker sweeps are masked out. The fastest class in the archive; the lowest LSD value (10.32 Hz) is above the highest non-psychedelic one (10.25 Hz), and above every DMT, ayahuasca and psilocybin session but two.
Of acute beta peaks sit at 2× the alpha frequency — what a non-sinusoidal alpha waveform produces (§11).
A sixth LSD-labeled session lands where the first five did: 1/f-corrected alpha peak 10.89 Hz across the session with its flicker sweeps and final movement episode masked out (10.65 unmasked), at the top of the LSD range, with a periodic alpha share of 0.86. The session mean hides a rise of roughly 1 Hz per hour, so "the LSD alpha frequency" is a time average. The driven resonance says the same thing about the generator that spontaneous alpha does: within each sweep the fitted f0 spans only 0.1–0.7 Hz across the ten to fourteen channels with a reliable fit — consistent with a common or highly coherent generator seen from fourteen positions. The strongest driven response was not always occipital (FC5 and AF3 led one LSD-labeled sweep), which on this montage is as likely reference, volume conduction or forehead muscle as a frontal generator.
The fourteen recordings added in October (§05) fall where the earlier ones did. The two ayahuasca sessions sit at 10.15 and 10.12 Hz and the second psilocybin session at 10.56 Hz, all inside the psychedelic band, the psilocybin session above every ordinary recording; the eleven ordinary additions run from 9.04 to 10.08 Hz with a median of 9.98 Hz, within 0.1 Hz of the ordinary median. The frequency ladder of §17, LSD above the other three classes, is unchanged in its order: the lowest LSD session (10.32 Hz) stays above every DMT, ayahuasca and psilocybin session except two, the new psilocybin session and one DMT session. §18.
07 — Beyond the pacemaker Retrospective · exploratory
Analysis basis 56-recording cohort for per-recording values (19 psychedelic, 37 ordinary, of which 4 resting and 28 workday); the within-session eye-state analysis is the September one on 41 recordings · dataset release 2026.10.10 · last recomputed in analysis release 4.26 · current
Band power is a blunt instrument. Four further families of measure — extracted from every clean channel-window of the cohort recordings — get closer to what the consciousness literature actually argues about.
Spatiotemporal Lempel–Ziv complexity, the measure that increases under LSD, psilocybin and ketamine in the published work, sits 7.4% below the 37 non-psychedelic recordings (permutation p = 0.013), a gap that widened with the twelve ordinary recordings added in October. The comparison the published work actually makes, against non-psychedelic rest, now has four resting recordings behind it (Resting baselines 001 to 003, parts 1 and 2 counted separately), and against them the psychedelic sessions sit 0.4% below (p = 0.92): no difference. Compared session against session, this archive does not reproduce the increase.
Compared within a session it looks different (this is the September analysis on 41 recordings; the window-level statistics below have not been recomputed on the 56). Taking each session’s largest excursion above its own opening fifth, and running the identical peak-picking on the non-psychedelic sessions so that a long noisy recording cannot win by length alone, the psychedelic sessions rise +0.114 against +0.078 (p = 0.017), and LSD alone rises +0.130 (p = 0.003). That is the first version of this test to favor the published claim, and it does not survive the eye-state control: restricted to windows matching each session’s own baseline eye closure, the gap falls to +0.088 against +0.077 (p = 0.45).
The reason is eye-state movement rather than eye-state level. Closing the eyes lowers this measure — eyes-closed windows sit near 0.72 against 0.75–0.77 for eyes-open — so a session that begins with the eyes closed and later opens them shows a large rise above its own baseline for that reason alone. The psychedelic sessions move between the two states about five times as much as an ordinary day does (within-session spread 0.31 against 0.06, p = 0.005), which is exactly what the excursion statistic rewards.
Holding eye state fixed instead of matching it leaves a small positive term: adjusted for the classifier’s p(closed), the psychedelic coefficient is +0.020 (95% +0.003 to +0.036), and among eyes-open windows alone the psychedelic sessions sit 2.6% higher (p = 0.044). Among eyes-closed windows there is nothing (−1.3%, p = 0.94). An increase of two to three percent is the most this archive supports, and it should not be read as a drug effect: the eyes-open comparison sets psychedelic sessions resting against ordinary sessions spent working at a desk, so task, posture, screen and hour differ alongside the compound. The cleaner eyes-closed comparison has only ten ordinary sessions, none of them resting. The comparison this needs is a set of ordinary evening recordings made the way the acute ones are, with deliberate eyes-open and eyes-closed blocks (§16).
Beside it sits a trap. Single-channel LZ is lower under psychedelics than during office work, because low-amplitude desynchronized task EEG binarizes into something that looks maximally complex. Whichever measure you quote changes the answer, which is why both are reported here.
The DFA exponent of the alpha envelope ranks the three states monotonically: work 0.72 < rest 0.73 < psychedelic 0.80. Every value falls between 0.5 (memoryless) and 1.0 (scale-free), meaning alpha amplitude always carries long-range memory — but psychedelics stretch it furthest and desk work compresses it most.
DFA exponents above 0.5 indicate persistent long-range temporal correlations in the alpha envelope. That is compatible with near-critical dynamics, but it is not diagnostic of criticality on its own: slow non-stationarity, changes of eye state within a session and filtering all lengthen correlations too. What the ordering does show is that the envelope's memory is longest in psychedelic-labeled sessions and shortest at a desk.
The work < rest < psychedelic ordering is recomputed on the current cohort: 28 workday recordings, 4 resting recordings (the two resting baselines of the August sessions and the two parts of the September one) and 19 psychedelic sessions. The ordering holds with a smaller rest-to-work gap than the September version, which rested on a four-recording rest tier that turned out to hold three DMT sessions; read the rest median as four recordings, not a group of equals (two of the four are under five minutes).
Neuronal-avalanche statistics — branching parameter, size exponent, κ — were run because self-organized criticality is the natural formalism for "is anything emergent here". The median branching parameter is ~1.30, mildly supercritical, which is typical for scalp EEG measured this crudely. But the recordings with the most extreme values (1.7–2.3) are precisely those with dead or intermittently contacting channels. The avalanche family is artifact-driven at 14 channels and 128 Hz, and is not reported as a finding; it needs clean 32–64 channel data to mean anything. The multistep-regression estimator, built for the case of few sensors seeing a large system, gives a branching parameter of 0.98 in §22 (Result 13): close to critical and below it, the same in every state.
At rest the alpha field reorganizes into a different spatial configuration about forty times an hour. Emergent in the strict sense — a property of the coupled field, not of any electrode.
DFA15 exponents in every state. Persistent long-range temporal correlations in the alpha envelope over the measured scales, in every state, across eight months and four psychedelic classes. Compatible with near-critical dynamics; not diagnostic of them.
The beta peak sits at exactly twice alpha in 83% of acute spectra, consistent with an arch-shaped, non-sinusoidal alpha wave. Strong phase coupling supports a harmonic interpretation rather than an unrelated beta oscillator, although a distinct oscillator entrained 2:1 is not uniquely excluded. (§11)
08 — Channel dynamics Retrospective · exploratory
Analysis basis 55 recordings with window-level band series (19 psychedelic-labeled vs 36 ordinary; Resting baseline 003 part 1 is too short to window) · dataset release 2026.10.10 · last recomputed in analysis release 4.26 · current
Band amplitude per channel in two-minute windows, for the 55 recordings with window-level band series (every recording of the archive long enough to window; the recomputation of October replaced the 36-recording version). The level differences are what the earlier sections describe. This section is about the motion: how much each channel's alpha wanders across a session, whether the fourteen channels move together, and which way the session drifts.
The amplitude no longer wanders more. Withdrawn. On the 36 recordings of September the coefficient of variation of each channel's alpha across a session was 0.68 under psychedelics against 0.31 without (p = 0.014). On the 55 recordings it is 0.74 against 0.74 (recording medians, 19 vs 36, permutation p = 1.00; posterior channels 0.75 vs 0.83, p = 0.58). The difference did not shrink; it was never a psychedelic difference. The ordinary recordings made from mid-September onward (Workday recordings 016 to 028) have a median coefficient of variation of 1.07 against 0.28 for the earlier workday recordings (p = 0.0007), wider than the psychedelic sessions themselves, so the September contrast was mostly the difference between early and late ordinary recordings. Why the later office recordings swing so much more is an open question for §15; the fit of the headset and eyes-closed pauses at a desk are the first candidates.
The channels move together, weakly. Mean pairwise correlation of the fourteen alpha envelopes is 0.89 under psychedelics against 0.83 without (p = 0.09; it was 0.89 against 0.77, p = 0.054, in September). The direction held and the margin did not: the swings are somewhat more global under psychedelics, but the archive cannot yet call it a difference.
The memory does not change. Window-to-window autocorrelation of global alpha is 0.45 under psychedelics and 0.32 without (p = 0.18): the same persistence from one two-minute window to the next, as in September.
What does separate the states is where the alpha sits. The posterior-to-frontal log ratio of alpha power, averaged over the session, is -0.24 under psychedelics against -0.11 without (p = 0.001): alpha is less posterior-dominant in a psychedelic session, the frontal share higher. This is the window-level form of the frontal alpha that §14 and §13 report, and it is the one dynamics measure here that survived the fourteen October recordings; the two ayahuasca sessions and the psilocybin session all sit on the psychedelic side of it.
Non-psychedelic sessions lose alpha as they go — the slope is negative at thirteen of fourteen channels, steepest at F4 and O2, the fatigue signature of an hour at a screen. Psychedelic sessions build it: the slope is positive at twelve of fourteen, strongest at P7, AF3 and FC5. That matches the within-session trajectory of a trip, in which alpha is suppressed early and recovers with a higher frequency late, and it is the reason a psychedelic session's mean tells you so little about any of its minutes. These slopes are not matched on eye state, and most of the rise is the eyes closing later in the session: within one blink-defined eye class, alpha power falls over the eyes-closed time of the LSD sessions and rises in the other psychedelic ones (§22, Result 11).
09 — Long-term impact Retrospective · exploratory
Analysis basis 56 recordings (37 ordinary, 19 acute); stimulation intervals masked in every recording that has them · dataset release 2026.10.10 · last recomputed in analysis release 4.26 · current
Nineteen acute sessions across four classes, interleaved with thirty-seven ordinary recordings over eight months, allow two questions. Does the non-psychedelic brain change as the sessions accumulate? And does the acute state itself change from one session to the next? The two groups are analyzed separately, because they differ in eye state and context, and each measure is fitted with a constant, a straight line and a small set of curved forms — a plateau, a step, a bend, a smooth spline — rather than a straight line alone.
Alpha frequency is a constant over the first five months and steps up by about half a hertz, from a median of 9.2 to 9.8 Hz, with the recordings of August onward (step at day 127; permutation p = 0.007, q = 0.02; 37 recordings). The posterior center frequency, posterior alpha power and alpha bicoherence step with it and gamma share steps down: the whole cluster moves the way an eyes-closed block moves it, and the step arrives when the fixed pre-session block, the resting baselines and the stimulation-masked sessions entered the protocol. The psychedelic recordings, eyes closed throughout, show no step (p = 0.81). Whether the pacemaker of §06 moved or the recordings did cannot be decided from these data.
Whether the dependence is allowed to be a line, a plateau, a step or a smooth curve, the constant wins for 13 of 14 measures (p from 0.14 to 0.97; 37 recordings). There is no washout gradient to find, straight or bent — and the two non-psychedelic sweeps after the dosed session agree (§18).
Global synchrony drops once, in mid-March, from 0.67–0.70 to 0.49–0.60, and stays there: a step beats a straight line by 48 AICc units (p < 0.001, q = 0.005; 37 recordings). The aperiodic exponent steps up at the same point. A second step, in the alpha measures, arrives with the August-to-October recordings; it is in the non-psychedelic cohort only.
What looked like a drift is mostly one early step. A rank correlation cannot tell a slope from a step, and the straight-line reading of earlier versions was hiding one: after mid-March, none of the fourteen measures shows any dependence on time in the later non-psychedelic recordings until the recording regime changes again in late July (below). The same step appears in the psychedelic recordings, where the two sessions before mid-March have the highest synchrony of the sixteen. A change that arrives in both groups in the same three weeks, after two or three sessions, and then stops is what a change in how the recordings were made would look like — headset fit, electrode technique — and it is not the shape a cumulative effect of the acute sessions, accumulating across these months, would take. The data are consistent with, but not sufficient to establish, either reading. Time and cumulative exposure remain the same clock here (rank correlation 0.99).
Wave |k| no longer rises. With nineteen recordings a gradual rise remained possible (ρ = 0.69); with 37 it is a one-bend line (hinge at day 192, p = 0.0045, q = 0.022; ρ = 0.36) whose late segment turns down, and after day 15 it has no support (p = 0.12). The traveling-wave score and state switching show nothing (p = 0.13 and 0.21).
A second step, in late July. Alpha peak frequency is no longer a constant: its best description is a step up at day 127, a median of 9.21 Hz over the fourteen recordings before it (8.77–9.96) and 9.77 Hz over the twenty-three after (9.00–10.25); p = 0.007, q = 0.023; ρ = 0.50; nothing beyond a step (p = 0.36). The posterior 1/f-corrected center frequency steps the same way (9.61 to 10.00 Hz), posterior alpha power above the background doubles, gamma share falls, spectral entropy declines and alpha-to-2-alpha bicoherence steps from 0.08 to 0.30. The twenty-three recordings after the step are the whole August-to-October regime: both August resting baselines (eyes closed by design), the two parts of Resting baseline 003, every recording made under the fixed pre-session block introduced on 4 September, the eleven stimulation-masked workdays and the office sessions of late September and October. Every measure in the cluster moves in the direction a larger eyes-closed share moves it, and the psychedelic recordings, eyes closed throughout, show no step at all (alpha frequency p = 0.81). Time of day did not drift. Whether the pacemaker of §06 moved or the recordings did cannot be decided from these data; the fixed weekly block decides it.
The day-after brain is the ordinary brain. Eighteen of the 37 non-psychedelic recordings fall within a week of a session and nothing separates them from the rest once the recording regime is accounted for. Two measures now differ in the within-a-week dichotomy, state switching (p = 0.046) and wave-state entropy (p = 0.02), with the far group at 9.5 per hour and 0.41 against 3.7 and 0.14; against the continuous clock both are constants (p = 0.52 and 0.47), and the far group is dominated by the September-to-October regime in which the long gaps after the 5 and 26 September sessions fall. It is the late step seen through a dichotomy, not a washout effect. No measure tracks days since the last session. If psychedelics leave a signature in this person's spontaneous EEG a few days on, it is below what fourteen electrodes can see.
Does the acute state change from session to session? Pooled across classes, no: none of 42 tests survives correction and a constant is the best description for 33 of them. The three sessions recorded after the models and claims were published (Psilocybin experiment 017, Ayahuasca experiments 018 and 019) sit inside the earlier range on every measure. Because the mix of classes changes over the months, a pooled trend could be pure composition55, so each trend was re-tested within class; no measure reverses direction, and DMT and LSD sessions turn out to be spread evenly across the archive (median day 100 and 106).
Within the six LSD sessions the strict orderings reported in September are weakened by the stimulation-masked row of LSD experiment 016: wave-state entropy no longer rises in perfect order (ρ = 0.66, not 1.00; the session now at 1.33 rather than 1.52), while the aperiodic exponent still does (ρ = 0.94, p = 0.005) and spectral entropy and synchrony still fall (ρ = −0.77 each). Nothing passes correction within the class (smallest q = 0.13), the fraction of clean data retained in those sessions rose in almost the same order, dose is known for only one of them, and six points cannot separate exposure from technique. Within the seven DMT sessions, where data retention did not trend, nothing changes: a constant is best in all 42 tests. There is no evidence that the acute state has strengthened or weakened.
The lead reported in September, higher alpha wave-state switching after intervals longer than about two weeks (p = 0.013 with fifteen intervals), does not hold with eighteen: a constant is the best description (p = 0.24; ρ = 0.32), and no measure depends on the inter-session interval (p from 0.09 to 0.87). It is withdrawn as a lead. Lasting changes after psychedelics have been reported in resting-state imaging weeks after dosing565758; nothing here contradicts that literature, which used far more sensitive instruments, and nothing here supports it either.
An earlier first-to-last comparison reported a 57% loss of standing-alpha organization and a doubling of gamma share between the archive's first and last recordings, framed as a possible six-month change. Its first endpoint, DMT experiment 001 — then archived as a baseline — is now known to have been a DMT session. The comparison was DMT against office work, not February against August, and it is withdrawn.
Forty-two tests were run in each group — fourteen measures against three clocks — each a permutation test with the search over curved forms repeated inside the null. At a false-discovery rate of 10%, 10 of the fourteen measures pass against time in the 37 non-psychedelic recordings: the mid-March step in synchrony and the aperiodic exponent, and the late-July step in alpha frequency, the posterior center frequency, posterior alpha power, gamma share, spectral entropy, bicoherence and frontal asymmetry, and the one-bend line in wave |k|; only synchrony and the exponent pass the test for curvature. In the psychedelic recordings none passes. Everything here describes what the archive shows; the design that would answer the question — a fixed eyes-closed block at the same hour on a weekly schedule, independent of dosing — is part of the current acquisition protocol (§25).
The question most often asked of long-term psychedelic effects is whether the coupling between frontal and posterior cortex, the scalp-level counterpart of the default-mode network that resting-state imaging reports as changed weeks after dosing, has moved over the months. Fourteen electrodes cannot image a network, but they can measure phase coupling between the front and the back of the head. The proxy used here is the debiased weighted phase-lag index between the anterior-frontal pair (AF3/AF4) and the lateral-posterior pair (P7/P8), in every band, computed in the same windows as the measures above, adjusted for eye state by a linear term in the classifier's closed probability, and followed against calendar time in the non-psychedelic recordings, the psychedelic recordings and all recordings together, with the same constant, line, step and curved-form search as the rest of this section. The frontal F3/F4 pair and the occipital O1/O2 pair are run beside it as controls: a change confined to the occipital pair is visual, not a network.
Nothing drifts; the mid-March step is here too. In delta, theta, beta and gamma the first three or four recordings have coupling of about 0.03 and every later recording about 0.01, in the non-psychedelic recordings, in the psychedelic recordings and in the F3/F4 pair alike: over all 56 recordings the theta step (median 0.036 before day 17, 0.010 after; p = 0.0003, q = 0.004) is the strongest test of the whole analysis, and the non-psychedelic beta step has p = 0.0025 (q = 0.019). After the step the constant model wins in every band (all p > 0.37 after day 20), no single date removes it, and shuffling dates within class leaves it. The same weeks show a fall in 20–45 Hz spectral content in every channel, so the step is most economically the change in recording conditions that §09 already assigned to mid-March, not a change in the brain.
Alpha is the one band without the early step, and the one band with a within-psychedelic decline. Non-psychedelic alpha coupling is a constant across the archive (permutation p = 0.42; ρ = +0.23), and no band depends on days since the previous psychedelic session (p from 0.28 to 0.95), which is the washout test this section asks of every measure. Across the 19 acute sessions the best description is a step down at day 75: the first seven sessions have a median of 0.094, the twelve later ones 0.049 (p = 0.021, q = 0.06; ρ = −0.58). The decline survives date shuffling within class, is present in the eyes-closed stratum and absent in the eyes-open one, holds after day 20, and no single date removes it; it does not survive the stricter Benjamini–Yekutieli correction, and the seven high sessions are the first seven of the archive, recorded before the fixed pre-session block existed and including the two with the poorest data retention. With nineteen points, and time, cumulative exposure, technique and dose all on one axis, it is a lead for the planned fixed-protocol recordings, not evidence of a lasting change.
Acutely, frontal–posterior alpha coupling is higher, and the day after it is ordinary. Against the non-psychedelic recordings the eye-adjusted alpha coupling is higher under LSD (+0.027, p = 0.0001), DMT (+0.021, p = 0.011) and ayahuasca (+0.019, p = 0.08), and beta coupling is higher in every class (+0.011; p from 0.0009 to 0.004); the psilocybin pair of sessions does not differ. In the eyes-closed stratum alone the alpha difference vanishes (class medians 0.12–0.14 against 0.13), so the acute alpha difference is largely the eye state of the sessions, while beta stays elevated. Within LSD experiment 016 the post-ingestion windows show the largest acute change in the data in the occipital pair (+0.129, p = 0.0007), not the frontal–posterior proxy. The two day-after recordings, Daily-life recording 002 and Daily-life recording 005 (the morning after Psilocybin experiment 017), lie inside the range of the ordinary recordings around them in every band.
Washout now has a direct measurement: the driven resonance two and three days after the LSD-labeled session was 9.75 and 9.72 Hz, inside the pre-specified non-psychedelic window and 0.4–1.0 Hz below the one post-dose sweep that can be placed; the session's own pre-dose sweep (9.93 Hz) and the two later days bracket the post-dose value on both sides. The driven shift did not persist to day two. The psilocybin session of September adds a second instance: its driven peak never left the 10 Hz grid point, and the next day's sweep (Daily-life recording 005) sat at 10.24 Hz, inside the non-psychedelic window. The fixed weekly eyes-closed block exists as a protocol and has been run, but not at a fixed hour on a dosing-independent schedule. §18.
10 — Thalamocortical dysrhythmia Retrospective · exploratory
Analysis basis 56-recording cohort · dataset release 2026.10.10 · last recomputed in analysis release 4.26 · current
Two null results in a row, and this is the one with clinical stakes. Thalamocortical dysrhythmia is a proposed common mechanism behind neuropathic pain, tinnitus, Parkinsonian symptoms and some depression: hyperpolarized thalamic relay cells fire calcium bursts at 4–7 Hz, cortex entrains to that theta, and abnormal gamma appears at the edges. It makes five testable spectral predictions. All five were tested.
| Criterion | What the archive shows | Verdict |
|---|---|---|
| Dominant posterior oscillation shifted into theta | 0 of 56 cohort recordings. The posterior center frequency never falls below 8.9 Hz, and that minimum is a three-minute eyes-open fragment (Resting baseline 003 (part 1)); every recording longer than five minutes sits at or above 9.0 Hz. | Absent |
| Posterior theta spectral peak | ≤5% of non-psychedelic channel spectra, and those are frontal-midline task theta, not posterior. | Absent |
| Sustained theta burden | Theta-dominant runs ≥1 s occupy 0.53% of posterior time non-psychedelic, 0.64% acute. | Absent |
| Theta→gamma phase–amplitude coupling | ~1% gamma modulation depth. The comodulogram peaks at 8.5–9.5 Hz — alpha itself — not 4–6 Hz. Large values appear only at F7/F8/T7, the saccade and jaw-muscle contacts. | Negligible |
| Edge-effect gamma tied to theta | Theta–gamma envelope correlation 0.13, below alpha–gamma (0.19). Gamma tracks alpha and muscle, not theta. | Absent |
| Direction of pharmacological perturbation | LSD moves the resonance up by ~0.5 Hz — the mirror image of hyperpolarization-driven slowing. Sharpening is what the model predicts but is not shown here: Q is resolved in only six of fifteen sweeps and shows no ordering by state (§18). | Opposite |
| Thalamic source confirmation | Impossible with 14 scalp electrodes. Would require MEG or high-density EEG with source modeling. | Not assessable |
This is a research screen on consumer EEG, not a clinical test. It cannot rule any disorder in or out. What it can say is that across the 71 hours of clean signal in the 56 cohort recordings, the specific spectral signature the dysrhythmia model predicts never appears — and the one pharmacological lever available pushes the system in the opposite direction from where dysrhythmia lives.
The direction of the pharmacological lever is now measured directly rather than inferred. Driven f0 was 9.7 Hz on two non-psychedelic days and 10.2–10.7 Hz forty minutes into the LSD-labeled session — up, away from theta, as the screen above assumed. No sweep produced a theta-range resonance, and the one spontaneous theta peak in the archive — 5 Hz at the occiput late in LSD experiment 016 — is half of a fast alpha, not a slowed dominant oscillation, and is partly head motion (§11). Sharpening is not shown: the resolved values of Q run from 6 to 17 with no ordering by state, and per-channel widths did not narrow. The psilocybin session of September adds a second compound: its three sweeps (one before the dose, two after) kept the driven peak on the 10 Hz grid point throughout, with the fitted f0 at 10.65, 9.86 and 9.76 Hz, so the upward movement seen under LSD is not a property of every compound, and nothing moved toward theta. A thalamic source remains unassessable with fourteen electrodes. §18.
A single-subject EEG archive: 56 recordings, 100.1 hours, 19 acute sessions, four psychedelic classes, 34 photic resonance sweeps, thirteen light-noise runs, one brain, February to October 2026. All 56 recordings form the analysis cohort.