Electrodermal activity (EDA), the skin’s electrical conductance, driven by sweat-gland activity under sympathetic nervous system control, was the primary dependent variable in the Braud–Schlitz program because it is a sensitive, continuous, and involuntary index of autonomic arousal. The receiver cannot consciously suppress or amplify EDA in response to demand characteristics, partially addressing the response-bias artifact. EDA also responds rapidly to arousal changes, making it well-suited to the 30–60 second epoch structure of a DMILS session. Braud and Schlitz reported on the use of EDA as a bio-PK target in their 1983 work¹ and elaborated the rationale in their 1989 methodology paper.⁴ The systematic review of the DMILS literature confirms EDA as the dominant physiological measure across subsequent replications.⁹

The primary artifact concern in DMILS is sensory leakage, the possibility that the receiver detects the agent’s intention through conventional sensory channels (sound, vibration, electromagnetic signals) rather than through any anomalous process. The protocol addresses this by physically separating sender and receiver in different rooms, with the receiver housed in an acoustically shielded chamber. Where electromagnetic shielding was also employed, this further mitigated the possibility of low-frequency electromagnetic transmission. The systematic review of the DMILS literature notes that the receiver must be housed in an acoustically and ideally electromagnetically shielded chamber as a design requirement.² This arrangement eliminates direct auditory and visual cueing. It does not fully address the possibility of experimenter-mediated cueing during session setup or data handling, which remains a partially unresolved artifact in the paradigm.

A recurring methodological challenge in DMILS research is specifying what the agent is actually doing during intention epochs. Braud and Schlitz’s 1989 methodology paper in the Journal of Scientific Exploration⁴ addressed this by proposing a structured approach to transpersonal imagery: the agent forms a vivid mental image of the receiver in the desired physiological state (calm or activated), sustains that image during the epoch, and uses the real-time EDA feedback displayed on the monitor to guide the imagery. This feedback loop was designed to keep the agent engaged and to provide a functional analog to the feedback that healers report using in clinical settings. The approach does not resolve the question of mechanism, whether any observed effect is mediated by imagery per se, by attention, by intention, or by some other process, but it standardizes the agent’s task in a way that permits cross-study comparison.

The 1983 paper by Braud and Schlitz in the Journal of Parapsychology¹ reported psychokinetic influence on electrodermal activity using the DMILS epoch-comparison design. Receivers were physically separated from agents and blind to epoch type. The study examined whether agents’ calm and activate intentions corresponded with differential EDA levels in receivers. Results were reported as significant, with EDA levels tracking the intended direction across the session. Specific effect sizes and exact sample sizes are not available from the pool entry for this paper; the result is characterized as exploratory given the pre-registration norms of the period. The paper is among the foundational empirical reports in the DMILS literature and is cited in subsequent systematic reviews as part of the core dataset.

The intuitive data sorting hypothesis holds that a receiver with anomalous information-access ability might detect the epoch sequence and adjust their own EDA accordingly, producing the observed covariation without any genuine distant influence from the agent. Braud and Schlitz’s 1989 paper³ examined this by testing conditions designed to dissociate agent-directed influence from receiver-initiated sorting. The paper’s findings are characterized in the pool as addressing the possible role of intuitive data sorting in electrodermal bio-PK. This represents a methodologically important step: rather than dismissing the alternative, the researchers designed a test of it. The extent to which the test fully eliminates the sorting hypothesis versus partially addressing it is not determinable from the pool entry alone; the artifact is classified as partially addressed rather than eliminated.

The 1985 PSI-Research paper by Schlitz and Braud⁵ combined ethnographic observation of Reiki-plus natural healing practice with experimental measurement of physiological outcomes in recipients. This hybrid design was unusual for the period and reflects Schlitz’s anthropological training alongside Braud’s psychophysiological expertise. The study examined whether the intention-based practices of Reiki practitioners produced EDA or other physiological changes in recipients under conditions that controlled for direct physical contact and conventional sensory communication. The ethnographic component documented the practitioners’ own accounts of their methods, providing ecological validity context for the experimental conditions. Specific effect sizes are not available from the pool entry.

Cells marked “not-disclosed” reflect genuine absence from the source paper (typical of pre-2015 parapsychology corpus where preregistration + stopping rules were not standard methodological practice). Per the ESP-Nexus honesty mandate, the gap is surfaced rather than filled with inference.

The meta-analysis of distant-intention experiments⁹ was conducted by Stefan Schmidt at the University Hospital Freiburg and examined the DMILS paradigm from a healing energy research (HER) perspective. The review outlined the standard DMILS procedure, two spatially separated participants, physiological recording from the receiver, randomized epoch sequencing, agent feedback via real-time physiological display, and then evaluated the accumulated literature against HER-relevant questions. The review confirmed that EDA was the dominant physiological measure across studies, that physical separation and shielding were standard design features, and that the paradigm had been adopted by independent laboratories. The review’s conclusions regarding overall effect size and publication-bias assessment are not fully specified in the pool entry; the provenance of the review’s findings is classified as unclear given the absence of preregistration documentation for the review itself. No funnel-plot or trim-and-fill publication-bias analysis has been published on the 30-experiment Schlitz–Braud retrospective dataset; a bias-corrected effect-size estimate for the accumulated DMILS literature is therefore not available, and the reported Stouffer Z = 6.17 / mean r = 0.25 should be read against that limitation. As an indirect signal: Schmidt’s 2012 independent meta-analysis of 11 distant-intention experiments (576 sessions, random-effects model) reported a substantially smaller combined effect size of d = 0.11 (p = 0.03)⁹. The difference between r = 0.25 (Schlitz–Braud retrospective, equivalent d ≈ 0.51) and d = 0.11 (Schmidt independent random-effects) is itself consistent with the attenuation pattern typically observed when publication-bias correction is applied to literatures with selective reporting, even though neither meta directly carried out a formal trim-and-fill correction on the same study set. The general class of meta-analytic divergence — same paradigm, opposite conclusions is treated on the measurement-methodology replication spoke.

Braud and Schlitz’s 1991 paper⁶ proposed that DMILS results reflect genuine consciousness interactions with remote biological systems, a claim that goes beyond the data to assert a specific type of causal mechanism. The data themselves (EDA covariation during intention versus control epochs, above chance) are separable from this mechanistic interpretation. Alternative non-psi explanations that have been considered in the literature include: (1) publication bias, positive results are more likely to be submitted and accepted, inflating apparent effect sizes across the literature; (2) optional stopping, experimenters who observe promising trends may terminate sessions early, inflating significance rates; (3) subtle sensory leakage, acoustic or electromagnetic signals not fully blocked by shielding; and (4) intuitive data sorting by the receiver (directly tested by Braud and Schlitz³). Of these, intuitive data sorting was directly tested and partially addressed; publication bias and optional stopping remain unresolved artifacts across the accumulated literature; sensory leakage was mitigated by shielding but not fully eliminated in all studies. No consensus mechanistic interpretation has emerged.