Computational Efficiency η and Neural Dissipation ||Δ|| During Propofol Sedation: EEG Evidence for the RDRT Consciousness Gradient

A reanalysis of EEG data from 21 people receiving propofol sedation shows a sharp drop in a measure of computational efficiency (η) from 0.893 during wakefulness to 0.330 during sedation, alongside a large increase in a dissipation norm (||Δ||) from 1.96 μV to 30.05 μV. Pattern stability collapsed across all frequency bands during sedation, while amplitude rose, especially in delta and gamma bands. These results place propofol sedation in a dissipative transition zone where the brain expends substantial energy without producing organized spatial patterns, distinct from both wakefulness and deep unconsciousness. The findings support the prediction that η decreases along the consciousness gradient and establish sedation as an intermediate state in the η-||Δ|| phenomenal state space.

Study at a glance

Abstract

Abstract We apply the Effective Power framework (Eff = Amp x Sgamma; Waterman 2026a) and the integrated scalar metrics η (computational efficiency) and ||Δ|| (dissipation norm) to a propofol sedation EEG dataset (ds005620; Bajwa et al. 2024, N=21, 65 channels). Recordings during wakefulness (eyes-closed resting) are compared with two levels of propofol sedation. The primary result is a massive reduction in η during sedation relative to wakefulness (awake: 0.893; sedated: 0.330; d=-2.151, p<0.001), accompanied by a dramatic increase in ||Δ|| (awake: 1.96 uV; sedated: 30.05 uV; d=+0.846, p<0.001). Pattern stability Sgamma collapses across all frequency bands during sedation (awake: 0.79-0.86; sedated: 0.06-0.38), while amplitude rises — particularly in delta and gamma bands. These results place propofol sedation in the dissipative transition zone of the η-||Δ|| phenomenal state space: low η combined with high ||Δ|| indicates that the sedated brain is expending substantial energy without producing organised spatial patterns. This is distinct from deep unconsciousness (where both η and ||Δ|| would approach zero) and from wakefulness (high η, low ||Δ||). The findings confirm the RDRT prediction (P4) that η monotonically decreases along the consciousness gradient, and establish sedation as an intermediate state in the phenomenal state space defined by η and ||Δ||. Combined with results from three previously analysed datasets (reversal learning, insight, meditation), these results establish η as a universal marker spanning both intra-conscious phenomenal transitions and the full consciousness-level gradient.