Energetic Efficiency of Two-Stage Wave–Commitment Decision Architecture: A Meta-Analytic Comparison with Drift-Diffusion Models

The Coherent Resonance Network (CRN) framework proposes a two-stage decision architecture: a fast, low-energy wave reconnaissance stage (Stage-I, modeled by GKSL/Lindblad dynamics) that extracts a selectivity prior R₀, followed by a standard diffusion-to-bound commitment stage (Stage-II). We derive an analytical energy comparison between this two-stage architecture and three baselines: standard fixed-bound DDM, urgency-gating DDM with collapsing bounds, and a naive wave-only architecture. A PRISMA-guided meta-analysis of 25 task conditions from 11 publications demonstrates that 92% of published perceptual decision tasks fall within the regime where the wave-stage investment pays for itself. Two prior-manipulation conditions from Mulder et al. (2012) provide direct empirical validation with measured energy savings of +8.6% and +4.8% over the urgency-gating baseline. Companion paper: "Noise-Assisted Transport Windows in Biological Connectome Subgraphs: A Coherent Resonance Network Study" (doi:10.5281/zenodo.18249249).