The von Neumann–Wigner interpretation holds that the quantum measurement process is not complete until a conscious observer registers the result, at which point the wavefunction collapses from a superposition to a definite state. In a double-slit system, a photon or electron passing through both slits simultaneously produces an interference pattern; any measurement that reveals which slit the particle passed through destroys the interference. Radin’s hypothesis is that focused human attention, directed toward the slits, acts as a weak measurement, partially collapsing the wavefunction and thereby reducing fringe visibility. The researcher’s preferred interpretation is that this constitutes evidence for a consciousness-matter interaction at the quantum level; alternative interpretations include thermal or mechanical perturbation of the apparatus, subtle electromagnetic effects from the participant’s body, and statistical artifacts from optional stopping. Radin and collaborators have addressed the electromagnetic alternative by shielding apparatus and using remote online participants.³ The quantum-biology framing is discussed in Kauffman and Radin’s hypothesis paper.⁸ Radin has also responded to critics who argue that invoking quantum mechanics for psychic phenomena constitutes misuse of the framework, arguing that quantum mechanics provides evidence of physical-world properties compatible with psi, rather than a causal explanation.⁹

The six experiments reported in Radin et al. (2012) used a double-slit optical system in which a laser beam passed through two slits and the resulting interference pattern was recorded by a line camera at 4 Hz. The ratio of double-slit spectral power to single-slit spectral power served as the primary dependent variable. Participants were instructed to focus attention toward or away from the slits in counterbalanced blocks. The combined result across all six experiments was in the predicted direction. The researcher’s preferred interpretation is that focused attention acted as a weak quantum measurement, partially collapsing the wavefunction. Alternative non-psi explanations considered include: (1) thermal expansion of the apparatus from participants’ proximity, partially addressed by remote conditions in later studies; (2) electromagnetic interference from participants’ bodies, partially addressed by shielding; (3) optional stopping, addressed by pre-specifying trial counts per session. The study did not include a fully automated no-participant control run in all six experiments, leaving experimenter-equipment interaction partially unresolved.¹ A follow-up 2013 series of three additional experiments with 50 half-hour sessions further investigated spectral magnitude and phase associated with the double-slit component.¹⁰

The 2021 retrospective report (Radin, Wahbeh, Michel, and Delorme) described data collected 2012–2013 from N=25 participants, each contributing 10 half-hour sessions. The primary analysis compared interference pattern metrics during attention-toward versus attention-away conditions. The paper describes the result as exploratory evidence of a causal influence, distinguishing it from the earlier correlational framing. The specific artifact addressed was proximity-based thermal perturbation: the apparatus was located 3 meters from the participant, partially mitigating direct body-heat effects but not eliminating them. Experimenter cueing was partially addressed by computer-controlled condition assignment. The delayed reporting of these data (collected 2012–2013, published 2021) raises a file-drawer concern that the authors acknowledge; they note the data were not selectively reported but were held pending additional replication attempts.¹¹

The 2016 online experiment (Radin, Michel, and Delorme) used a line camera recording the interference pattern at 4 Hz; fringe visibility was computed for the central 20 fringes per image. Participants were recruited worldwide and received a web-browser-based feedback display linked in real time to the fringe visibility metric. The counterbalanced attention-toward and attention-away conditions were assigned by the server. The physical separation of participants from the apparatus eliminated proximity-based thermal and electromagnetic artifacts, eliminating direct body-proximity cueing but not fully addressing whether the internet feedback link itself introduced timing artifacts. The single-photon apparatus (Radin, Michel, Pierce, and Delorme, 2015) used a photomultiplier tube to detect individual photons arriving at an interference fringe minimum; participants were asked to direct attention toward or away from the optical system while receiving feedback about photon arrival rate. Six experiments with this apparatus found evidence supportive of an interaction that appeared to steer the wavefunction to either reduce or sharpen interference fringes. Effect sizes were small and the studies were characterized as exploratory.²³

The 2022 paper (Radin and Delorme) described a two-year online experiment in which the primary dependent variable was the double-slit spectral component of the interference pattern. The simulated-observer control condition, in which the web server periodically accessed the feedback signal without a human present, addressed the artifact of feedback-link timing effects by providing a matched non-human baseline. The independent reanalysis examined whether the effect was better characterized as a change in variance rather than a directional mean shift, motivated by the hypothesis that fluctuating attention produces bidirectional perturbations. The paper is explicitly labeled exploratory. The multiple-comparisons artifact was partially addressed by pre-specifying the primary dependent variable (double-slit spectral power ratio) before secondary variance analyses were conducted.¹²

The five experiments (Radin, Bancel, and Delorme, 2021) used entanglement correlation strength measured in real time as the dependent variable, presented to participants via a graph or dynamic images on a computer monitor or web browser. Participants were tasked with mentally influencing the entanglement metric. Across the four laboratory studies, the combined result was p < 0.02; three studies at IONS produced p < 0.0002. The online experiment produced modestly significant results (p < 0.05). The researcher’s preferred interpretation is that mental intention influenced the degree of quantum entanglement. Alternative non-psi explanations considered include: (1) electromagnetic interference from participants affecting the detection electronics, partially addressed by shielding; (2) demand characteristics, partially addressed by the online condition where no experimenter was present; (3) multiple comparisons across five experiments, partially addressed by reporting combined statistics. The studies are explicitly labeled exploratory and the authors note that independent replication is required before conclusions can be drawn.⁴ A preprint version of this work was also deposited at the Center for Open Science.¹³

The 2008 pilot study (Radin) placed a Michelson interferometer inside a light-tight, double-steel-walled, shielded chamber. Participants sat outside the chamber with eyes closed and were asked to imagine perceiving the laser beam inside. Light patterns were recorded by a cooled digital camera once per second. The shielded chamber addressed electromagnetic interference from participants’ bodies, eliminating direct electromagnetic artifact, but the study was underpowered as a pilot and the stopping rule was not pre-specified. The result was described as exploratory evidence. This study preceded the double-slit series and informed the design of subsequent experiments by establishing that remote participants could be used without proximity artifacts.¹⁴

The 2025 study (Radin) tested the von Neumann–Wigner consciousness-collapse theory by distributing custom optical interferometers to 47 participants worldwide. The preregistered design specified that illumination in the observed portion of the interference pattern would be compared to the simultaneously recorded but never-observed control portion. This within-apparatus simultaneous control addressed the artifact of apparatus drift and environmental vibration, eliminating common-mode mechanical and thermal fluctuations that would affect both channels equally, but not fully addressing participant-specific handling of the apparatus (e.g., vibration from touching the device). Environmental sensors monitored temperature, humidity, and electromagnetic fields at each participant’s location. Real-time encryption of data addressed data-integrity concerns. The participant selection criterion (experience in focused attention) introduces a sampling-bias artifact: results may not generalize to unselected populations, and the selection process itself was not fully blinded. The preprint version of this study was deposited at the Center for Open Science prior to journal publication.⁵¹⁵