Computational Efficiency η and Dissipation Norm ||Δ|| as Universal Markers of Phenomenal Transition in EEG: Evidence from Three Independent Datasets
Zenodo (CERN European Organization for Nuclear Research) March 19, 2026 DOI: 10.5281/zenodo.19119020 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Secondary analysis of three independent EEG datasets Peer reviewed |
|---|---|
| Population | Participants in three EEG paradigms: probabilistic reversal learning (N=22), anagram insight (N=30), and meditation mind-wandering (N=24) |
| Key points | Across three EEG paradigms, the proposed computational efficiency metric η was significantly reduced and the dissipation norm ||Δ|| significantly elevated at cognitive transition events, with effect sizes ranging from small to large. The authors argue these metrics operationalize the efficiency of the current phenomenal state and quantify energy expenditure that fails to produce organized output, marking transitions toward a thermodynamic halt boundary. |
Abstract
Abstract We introduce two integrated scalar metrics derived from the Effective Power framework (Waterman 2026): computational efficiency η = ||Ψ|| / ||A||, where ||Ψ|| is the norm of the five-band Phenomenal Vector and ||A|| is the norm of total oscillatory power; and dissipation norm ||Δ|| = ||A - Ψ||, the norm of the Dissipative Vector. These metrics aggregate across all five frequency bands to provide scalar summaries of the organised-to-dissipative ratio of neural computation at any moment. We test whether η and ||Δ|| constitute universal markers of phenomenal transition across three independent EEG paradigms: probabilistic reversal learning (ds004295, N=22), anagram insight (Oh et al. 2020, N=30), and meditation mind-wandering (ds001787, N=24). η is significantly reduced in all three paradigms at cognitive transition events (reversal: d=−2.583, p<0.001; insight: d=−0.492, p=0.011; meditation MW: d=−0.862, p<0.001). ||Δ|| is significantly elevated in all three paradigms (reversal: d=+1.619, p<0.001; insight: d=+0.577, p=0.003; meditation MW: d=+0.410, p<0.001). The frequency centroid Ψ_c = Σ(f_i × Eff_i) / Σ(Eff_i) is significantly elevated during mind-wandering (d=+0.415, p=0.025) but not at reversal or insight events. We interpret these results within RDRT: η operationalises the computational efficiency of the current phenomenal state, and its reduction marks the transition toward the thermodynamic halt boundary at which P_noncalc is generated. ||Δ|| directly quantifies the energy expenditure that fails to produce organised output — the operational EEG signature of P_noncalc generation. These two metrics, together with band-specific Eff profiles, provide a complete multi-level characterisation of phenomenal transitions in EEG data.