The Silence Experiment: A Multi-State Protocol for Detecting Consciousness-Dependent Decoherence in Shielded Environments

Abstract

This paper specifies a controlled experimental protocol for testing the von Neumann-Wigner interpretation of quantum mechanics, the hypothesis that consciousness plays a role in quantum state reduction, against the prevailing environmental decoherence framework. It measures whether the cognitive state of a nearby conscious observer measurably affects the decoherence rate of a prepared quantum system. The platform places an experienced meditator in EEG-verified states inside a five-layer shielded chamber. A single nitrogen-vacancy centre in diamond measures coherence times across seven consciousness conditions spanning the cognitive-excitation axis, from active cognition to objectless awareness, against an empty-chamber instrumental baseline and a thermal-dummy control cell that reproduces the metabolic heat load of a seated occupant with no observer present. All seven conditions run in one randomised session, so chamber drift is common to the chain and differences out of the within-participant contrasts. Three primary tests are pre-registered at a family-wise threshold of 0.005 one-sided, Bonferroni corrected to approximately 0.00167 per test: an ordered test across the excitation axis, and two orthogonal planned contrasts specified as explicit weight vectors that sum to zero. A fourth, secondary test is the only one that can separate the protective prediction from the suppression pattern, and its non-orthogonality to the first contrast is disclosed rather than corrected away. Both the EEG classifier and the coherence-time estimation pipeline are frozen and lodged before collection, so that neither the independent nor the dependent variable carries unspecified analytic freedom. A pre-registered positive control establishes that the measurement chain is sensitive, not merely quiet, and it is a gate: without it, a null carries no information about effect size. Phase 1 runs on the same chamber with an isotopically purified sample and carries three pre-registered hypotheses of its own, including a coupling-model discriminator and a spectral-density sweep that converts the standard objection to dynamical decoupling into the measurement itself. It is not conditional on a positive Phase 0 result. The protocol is outcome-blinded and is not described as blinded in any stronger sense. Participants and the instructing experimenter cannot be blinded, because the independent variable is an instructed cognitive state, and this is stated rather than glossed. The expected outcome is a null, which is pre-registered as publishable. No calendar timeline is stated anywhere. Session burden is given as a design parameter from which a host institution computes its own schedule. The protocol is pre-registered at OSF and is not peer reviewed; no data have been collected. Changes in this version. Secondary test D is corrected from weights that summed to one rather than zero. The prior weights were not centred under the null and would have been passed by any positive coherence time, and the correction is filed as a timestamped pre-data amendment with its rationale, not as a silent erratum. The transmon Phase 1 is withdrawn and replaced, with the sensitivity comparison retained in the text so the error is visible. Sessions are chained rather than one condition per session, reducing participant burden by roughly a factor of seven. The separate non-meditator resting group is removed, because it placed a between-group boundary inside the ordered test. Suppression is defined by an interval criterion rather than by an indistinguishability claim. Budget figures corrected. No repository identifiers appear in the document.