The Cross-Substrate Access Assay: What an Indicator Test Must Declare to Travel from Brain to Language Model
P. V. van Rooyen
September 14, 2026
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper |
|---|---|
| Sample size | 20 |
| Population | Human electroencephalogram recordings and synthetic datasets |
| Key points | The Cross-Substrate Access Assay proposes a procedure for transferring consciousness-related measurements from brains to artificial systems, scoring models by cross-entropy in nats per trial. A re-implementation of a published global neuronal workspace test on twenty people's EEG partly reproduced the original results, matching first crossing and broad temporal ordering but not window-by-window agreement. On synthetic data, the original models falsely reported two states in 989 of 12,000 single-state datasets, while expanded families reported none; the expanded procedure detected two states in 599 of 600 mixture datasets. Nominal 95% intervals undercovered at six of twelve graded settings. The authors make no claim about experience. |
Abstract
Testing an artificial system for a property linked to consciousness means applying a measurement developed on brains to a system that is not one. Such a transfer must re-examine five parts of the procedure: the competing statistical models, how they are fitted, the unit the inference generalizes over, the quantity the uncertainty interval is about, and the rule that turns a result into a verdict. The Cross-Substrate Access Assay declares all five. Because brain and model signals share no physical scale, every model is scored by the cross-entropy it assigns to held-out data, in nats per trial. The test case is the global neuronal workspace theory, which predicts that near threshold a stimulus either enters a capacity-limited workspace or does not, so that single-trial responses form a mixture of two states. A published test of this prediction on twenty people's electroencephalograms partly reproduces in a re-implementation: the first crossing and the broad ordering over time match, the window-by-window agreement does not. On 12,000 synthetic datasets generated with a single graded state, all of them members of the families the procedure fits and none within 0.0067 nat per trial of the decision boundary, the two models of that test carried over unchanged reported two states in 989 and the expanded families in none; on 600 datasets carrying a mixture the expanded procedure reported two states in 599. Its nominal 95% interval contained the procedure's mean result less often than the required 90% at six of twelve graded settings. No claim about experience is made.