About the project
About the project
An immersive audiovisual feedback platform built on a quantum random number generator: a 32 by 32 grid of lights and a chime chandelier, run in sessions whose schedules are fixed before they start.
The platform
Ionosphere is a room-scale installation powered by physical entropy from an avalanche-noise hardware RNG (ubld.it TrueRNG v3) passing through a deterministic 20:8 Galois-field firmware whitener.
The bitstream directly modulates a 192-rod chime chandelier and a 1,000+ addressable LED grid. Rather than staring at a computer monitor or peripheral "gray box" (the primary failure mode diagnosed in PortREG 3-lab consortium trials), the operator sits inside an immersive environment where order and coherence emerge naturally from noise.
A hardware random number generator is sampled continuously. The bits drive a pattern engine, and the engine drives a 32 by 32 grid of LEDs. The canvas on the live page is that grid, pixel for pixel, as it is being displayed in the room.
The pattern engine page runs the same patterns in the browser. Pseudorandom, not hardware: the viewer draws from the browser's generator so it runs anywhere; the chandelier draws from a TrueRNG.
How a session runs
Glitch is how far the current sample sits from what chance would give, mapped into the range the engine uses to distort the pattern. Z is the same departure in standard deviations, plotted over the session so far, with the running high and low marked.
Focus window reports whether a scheduled attention window is open at this moment. The schedule is generated from a seed and hash-committed before the session begins, so it cannot be revised afterwards; the operator is signalled at each window onset rather than being told the times in advance. When it reads unknown, no schedule was armed for this run, which is deliberately shown differently from a resting period.
Viewers can watch a session as it streams, or take part. Taking part means pressing a button when you feel a focus window is about to start, and marking the beginning and end of any stretch where you notice something about your own state, then writing a line about it. A participant sees the light pattern and whether the stream is connected, not whether a focus window is open and not the running statistics from the device. The participation notice says what is recorded and how it is used.
What was fixed before data collection
Each session's window schedule is generated from a single integer seed. Before the first window opens, the scheduler writes a SHA-256 of the generated schedule into the run folder, alongside the seed and the parameters. The schedule itself stays hidden from the operator during the run.
Because the schedule regenerates deterministically from its seed, the hash pins it exactly. A schedule revised after the fact, to move a window onto or off an interesting stretch of data, produces a different hash and fails the check. Publishing the hash and withholding the seed until analysis lock is an ordinary commit-and-reveal: the commitment is public before the outcome is known, and becomes verifiable afterwards.
Seeds are deliberately not published here during active collection to prevent breaking operator blindness.
The pre-commitment record lists every schedule hash. It is a self-hosted commit-and-reveal record, not an external registry entry, and it says so.
Confirmatory, interim and exploratory materials
Confirmatory design
From the study's locked design document. The confirmatory family is evaluated against the Radin and Nelson 1989 anchor. The primary continuous endpoint is mean |z| dispersion evaluated via 20,000 circular shifts per run (Harris 2024).
Definition
A burst is a 60-second period in which the bit-count deviation reaches |z| ≥ 1.96.
Interim pooled results
79 valid runs analysed across the four strata below (2026-06-27 to 2026-10-05):
| Stratum / Arm | Runs | In-Win | Base | Enrichment Δ | Shift p |
|---|---|---|---|---|---|
| Human Solo (Full Confirmatory Arm) | 36 | 4,923 | 5,911 | +0.0314 | p = 0.0062 |
| Human Solo (Calibration Gate-Passing Subset) | 25 | 3,353 | 4,014 | +0.0124 | p = 0.2278 |
| Matched PRNG Scheduled Controls | 2 | 300 | 408 | +0.1413 | p = 0.0224 |
| Exploratory LLM (claude-fable-5, 5 overnight runs) | 5 | 785 | 725 | -0.0351 | p = 0.7994 |
Caution
Why we could not stop there: Post-hoc timing and uncounted multiplicity mean exploratory p-values cannot establish an effect. The only scientifically defensible path was to seal a follow-up: locking window schedules via SHA-256 hashes prior to data collection.
Exploratory arms
The desk-arrival and window-onset arms are described on the record with the date each was written down. The live-stream audience data are built to be published as a labelled exploratory arm without touching the confirmatory family.
Nothing on the live page is an experimental result. It is the instrument, running.
Data, analysis and public materials
Verify a run
Once a run's seed has been released, regenerate the schedule with the study's own scheduler at that run's parameters and seed, hash the output, and compare it with the value in the table. The parameters used throughout this series are ten windows of fifteen minutes across a six-hour span, with a minimum fifteen-minute gap.
The values to compare against are in the pre-commitment record.
TSC 2026 materials
The Science of Consciousness 2026 · Session THU-2: Mental Causation and the Function of Consciousness · Thursday 15 October 2026 · Dockside Room.
- Conference hub: Bursts, not mean shifts
- Conference talk deck (17 slides)
- Master reference deck (32 slides)
- Corpus analysis report (79 runs)
- Ionosphere overview on masonborchard.com, the writeup that predates this site
- Interactive guide, a walk through the pattern types in the browser
Collaborate
Write if you want to review the commitment record or request an export. Replications, an unattended control rung, and talks on the platform, the commitment design or the analysis pipeline are welcome. So are questions about the runs, the hardware, the analysis or the live sessions.
People
Mason Borchard runs Ionosphere Laboratories.