The 2025 COGITATE study directly tested contrasting predictions of two prominent consciousness theories. Its value for SMM/DTT is methodological: broad ideas became specific commitments about neural content, timing and connectivity. This essay explains the verified result, its limits and how a source-based consciousness proposal could become comparably concrete without claiming support by default.
The COGITATE Consortium’s Nature paper was published online on 30 April 2025 and appeared in volume 642, pages 133–142. It compared integrated information theory and global neuronal workspace theory through a preregistered adversarial design. Across 256 participants, it used fMRI, MEG and intracranial EEG while people viewed clearly visible stimuli for different durations. The authors reported some results consistent with each theory and substantial challenges to key predictions of both. This was not a test of SMM/DTT. [1]
SMM begins with one nonspatial generative source. DTT distinguishes its inner production from the outer order in which completed results become accessible. The six inner levels are nested functional dependencies within that production; the seventh denotes outer completion. They are not six independent clocks, six spatial directions, or six simultaneously occupied temporal axes. This distinction makes the framework more specific than a general claim that everything changes: it asks how a world of persistent relations can be generated before its spatial representation is taken for granted.
The experienced present is usually richer than an isolated instantaneous sensation. A spoken word, a rhythm or a movement is apprehended through relations among successive phases. SMM/DTT’s production/completion distinction offers a language for asking how an integrated event becomes available, but it does not make the experience of a present identical to a fundamental cosmic completion. A conscious observer is a complex physical and interpretive subsystem with its own memory, expectations and report capacities.
Three questions should be distinguished. Which neural or functional processes allow information to be integrated? Which processes make that information available for report and action? Why is the process accompanied by experience at all? A theory may make progress on one without settling the others. DTT can contribute a general account of record formation and nested dependencies. To become a theory of consciousness, it would need a specific bridge linking those structures to a subject’s experience. The number of inner levels alone does not select a duration of awareness or a neural mechanism. [2]
The study identified a lack of the predicted sustained posterior synchronization as a challenge to IIT. For GNWT, it highlighted the general absence of the expected ignition at stimulus offset and limited representation of some content dimensions in prefrontal cortex. These conclusions concern specified neural predictions and the study’s methods. They do not establish that consciousness is nonphysical, that a cosmic source is conscious or that either theory’s every philosophical commitment has been disproved. The evidential lesson is that a precise prediction can fail while the broader programme responds through explicit revision. [1]
DTT distinguishes inner production from outer accessible completion, but those terms alone do not predict where conscious content is represented or how long a neural response lasts. A consciousness extension needs a bridge from its source architecture to an observer model. That bridge should identify a physical state, an accessible record and a criterion for the conscious content under discussion. The six nested levels cannot simply be mapped onto six anatomical regions after results are known. Their functional roles and the corresponding measurable consequences need independent definitions.
Consider two illustrative observer models. Model S predicts that a measurable completion marker remains elevated throughout a presented interval. Model P predicts only brief markers at onset and offset. For a one-second stimulus, their predicted temporal profiles differ even if both predict an accurate final report. A measurement can therefore discriminate the profiles without resolving every question about experience. A DTT model would become informative by deriving one profile, its tolerances and its dependence on stimulus duration from the same source-memory law. Merely allowing either profile after observation would remove the comparison’s force. These are proposed model structures, not claims about a DTT experiment already performed.
A participant’s later report depends on perception, memory, decision and motor output. A neural marker associated with report can therefore differ from a marker of the experienced content itself. A source-based theory should model these dependencies rather than compress them into a single completion event. One useful architecture distinguishes availability for immediate action, retention for later report and the proposed phenomenal state. The same source can generate all three while their timings differ. This is a constructive use of dual-time language: it identifies which completion is being measured and prevents the observer’s entire cognitive history from being treated as one undifferentiated output.
The appropriate next claim is a restricted observer model with a specified law, readout and prediction. Its comparison should state which outcomes would favor it, which would favor a conventional alternative and which would remain inconclusive. Parameters used to calibrate a timing curve should not be counted again as successful predictions of that curve. New data should probe a consequence not already used to construct the model. This approach gives SMM/DTT a route toward evidence while preserving its philosophical motivation. Failure of competing theories is an invitation to make a better-defined proposal, not evidence that an untested source ontology is already confirmed.
Hear the three notes C, E and G in sequence. At the third note, the first two are no longer physically arriving in the same way, yet their temporal relation contributes to hearing an arpeggio. If only the current acoustic sample existed for the listener with no retention, the musical relation would be lost. If the notes were all treated as simultaneous, their order would disappear. The experienced present therefore includes a differentiated relation to what has just occurred. This example motivates a memory-sensitive model of temporal experience, while leaving open whether the relevant unity is explained by neural integration, phenomenological structure or a deeper metaphysical account.
THE SAME QUESTION, DIFFERENT STARTING POINTS
| Question | SMM/DTT | the 2025 adversarial theory test |
|---|---|---|
| Testable extension | A consciousness extension of DTT needs a specified observer model, readout and quantitative prediction beyond its general source ontology. | The 2025 COGITATE study tested preregistered predictions associated with integrated information and global neuronal workspace theories. |
| Interpretation of results | SMM/DTT would earn evidence through successful independent predictions, not by inheriting support from a competitor’s difficulty. | The study reported both supported and challenged predictions and did not establish an alternative source theory of consciousness. |
KEEP THE CONNECTION
Connects this comparison with the author’s discussion of experience.
Connects this comparison with the author’s discussion of experience.
Connects this comparison with the author’s discussion of experience.
Connects this comparison with the author’s discussion of experience.
Connects this comparison with the author’s discussion of experience.
References support the comparator and the specific mathematical or interpretive claims indicated. SMM/DTT connections are developed in the analysis above.
How these comparisons distinguish premises, results and physical claims
SMONAD
Share this page, or keep a clear reference to the work you are reading.
Use this reference for the explanation on this page. “n.d.” means its first publication date is not specified; the update date is recorded separately.
