Overview

Direct Mental Interaction with Living Systems (DMILS) is an experimental paradigm that asks whether one person’s focused intention can produce small, statistically detectable changes in the physiology of another living organism held at a distance, with ordinary sensory and informational channels blocked. The target measure is usually autonomic: electrodermal activity (EDA, the moment-to-moment electrical conductance of the skin), respiration, heart rate, or brain electrical activity. The receiver sits in a shielded room and cannot see, hear, or otherwise sense the distant influencer.

The paradigm was formalized by William G. Braud, PhD at the Mind Science Foundation, who built the EDA-based design that most later studies follow. Before the mid-1980s the same phenomenon was usually called Bio-PK (biological psychokinesis), framing the result as a mind acting upon a living body rather than as information transfer. DMILS and Bio-PK therefore name the same line of inquiry from different theoretical angles, but they are not interchangeable concepts: Bio-PK is the older label for the putative phenomenon, while DMILS is the modern label for the controlled method used to study it.

It is important to separate DMILS from remote healing. DMILS measures a brief autonomic shift in a healthy volunteer under tightly controlled laboratory conditions; clinical distant-healing trials test whether intention improves a patient’s medical outcome and use very different designs.^[1]

Protocol

A standard DMILS session pairs two people who never meet during the run. The receiver, sometimes called the influencee, rests in an isolated chamber while sensors record skin conductance. A computer program randomly schedules alternating periods in which a distant influencer either tries to calm or activate the receiver, or simply rests as a control.^[2] Because the schedule is randomized and known only to the machine, neither person can anticipate when an influence period will occur.

EDA became the favored target because it tracks autonomic arousal closely, responds within seconds, and can be logged continuously without the receiver doing anything. The analysis compares the receiver’s average physiological activity during influence periods against control periods. Related variants have swapped in other autonomic readouts: electrogastrography, which measures stomach activity as a proxy for “gut feelings,” has been used to ask whether a receiver’s body responds to a distant sender’s emotional state.^[3]

Remote staring is an adjacent paradigm rather than a subset of the DMILS protocol, and the quantitative reviews that cover both pool its results separately. Here the influencer covertly watches a live video feed of the receiver during randomly scheduled staring periods, and the question is whether being watched unawares produces measurable autonomic reactivity. The receiver gets no feedback and cannot know when the staring is happening.

History

The roots of the field reach back to the early 1960s, when biological targets began to be tested under the banner of Bio-PK. The decisive methodological turn came in the 1970s and 1980s, when Braud and colleagues at the Mind Science Foundation standardized the EDA paradigm and the randomized influence/control structure that defines DMILS today. Working with Marilyn J. Schlitz, PhD, Braud reported across a series of experiments that a distant influencer’s intention was associated with shifts in a receiver’s electrodermal activity.

Braud and Schlitz summarized the program’s first thirteen years in a 1991 review covering 37 experiments, 655 sessions, 449 influencees, 153 influencers, and 13 experimenters, with target systems that included another person’s electrodermal activity, blood pressure, and muscular activity, the spatial orientation of fish, the locomotor activity of small mammals, and the rate of hemolysis of human red blood cells; combining all systems, they reported a Stouffer z of 7.72 (one-tailed).^[16] Their 1997 assessment of the electrodermal line counted 19 direct-intentionality experimental series involving 434 different people (317 influencees, 105 influencers, and 12 experimenters), of which 7 (37%) were independently significant against the 5% expected by chance; effect sizes ranged from −0.25 to +0.72 with a mean r of +.25, and the overall Stouffer z was 4.82 (p = .0000007). A companion set of 11 remote-observation experiments comprised 230 sessions.^[17]

In the mid-1990s Dean I. Radin, PhD, with Braud, ran a pilot replication that found remote mental influence on EDA at a magnitude in line with earlier work, and reported that geomagnetic field fluctuations appeared to modulate the strength of the effect in a correlational analysis.^[4] Later studies extended the design across large distances and time gaps; two exploratory experiments reported autonomic changes in a distant person whether the sender was separated by thousands of miles or by two months.^[5]

Key findings

Across decades, individual DMILS studies tend to report small effects that nonetheless cluster above chance. A representative laboratory study directed compassionate intention from partners of cancer patients toward those patients and measured skin conductance under double-blind conditions; activation of the patient’s autonomic nervous system was strongest when senders had been trained in cultivating intention and were highly motivated.^[6] Studies of “gut feelings” using stomach-activity recordings likewise reported that a receiver’s physiology shifted in response to a distant sender’s emotions.^[3]

The remote-staring line has its own quantitative summary. A meta-analysis of sixty staring experiments concluded that a genuine ability appears to exist, and that the best-controlled studies, designed to rule out the receiver implicitly learning sensory cues, yielded consistent effects well beyond chance.^[7]

A subset of 10 of these studies, designed to preclude implicit learning of sensory cues, resulted in a homogeneous distribution of effect sizes and a weighted mean effect size substantially beyond chance expectation.^[7]

For the broader paradigm, a meta-analysis of attention-focusing facilitation experiments, drawing together hundreds of sessions with nearly identical designs, found a small but statistically significant effect supporting the idea that benevolent distant intention can help a target person’s attentional performance.^[8] Reviewers integrating the wider psi literature have placed DMILS effects among parapsychological findings whose magnitudes are comparable to accepted results elsewhere in psychology, while stressing that the individual effects are small and replication is uneven.^[9]

Two quantitative summaries anchor the DMILS-EDA literature. Schmidt’s 2003 systematic review located 24 experiments with 636 sessions from four laboratories (San Antonio, Edinburgh, Las Vegas, and Freiburg); 9 of the 24 (37.5%) were significant according to their own hypotheses, against the 5% expected by chance, and the unweighted mean effect size was r = 0.22 (95% CI 0.11 to 0.32), falling to r = 0.16 when weighted by number of sessions, with the largest effect sizes in the first five small experiments.^[21] Schmidt, Schneider, Utts, and Walach’s 2004 meta-analysis, the fullest published enumeration of this literature, reported a small significant quality-weighted mean effect of d = .11 (p = .001) across 36 studies and 1,015 sessions, a data set that passed its homogeneity test; their best-evidence synthesis restricted to the 7 highest-quality studies (188 sessions) yielded a nonsignificant d = .05 (p = .50).^[23]

The remote-staring thread carries its own pooled figures, and the reviews that cover both keep the two data sets apart because the designs differ: in staring studies the receiver is watched over closed-circuit video rather than targeted with calm or activate intention. Schmidt’s 2003 review counted 13 staring experiments with 300 sessions across four laboratories, with an unweighted mean effect size of r = 0.33 (session-weighted r = 0.31).^[21] The 2004 meta-analysis included 15 staring experiments with 379 sessions and reported a mean effect size of d = .13 (p = .01) after correcting a confound with sampling error, down from an uncorrected d = .28.^[23] A later Freiburg staring study with a distraction manipulation found no significant staring effect on electrodermal activity or on conscious guessing (conscious-guessing staring main effect p = .43, 43 participants).^[25] The library holds these reviews in full text together with a minority of the primary reports they pool, among them Radin, Taylor, and Braud’s 1995 pilot replication^[4] and Delanoy and Morris’s 1998 training study^[19]; the Braud-era primary experiments are represented here through the reviews’ own study tables rather than as standalone documents.

Methodological critiques

The most influential methodological episode in the field is the collaboration between Schlitz and the skeptic Richard Wiseman on remote staring. The two researchers ran the same protocol but obtained different outcomes, a pattern consistent with an experimenter effect rather than with any single party’s error. A later interview-based study examined how the two prepared for sessions, interacted with participants, and held their own belief and intention during observation, concluding that such tacit differences in practice may shape results.^[10] Expectancy effects of this kind remain a central interpretive problem: when the experimenter’s own stance appears to influence the data, separating a genuine anomaly from a subtle procedural artifact becomes difficult.

The joint series ran to three experiments, and each paper reports both experimenters’ arms. In the first study (1997), receivers run by Schlitz showed significantly higher electrodermal activity during stare than non-stare trials (Wilcoxon z = −2.02, p = .04; 16 receivers, 512 trials), while receivers run by Wiseman did not differ from chance (Wilcoxon z = −0.44, p = .64; 16 receivers, 512 trials).^[18] In the second (1999), the split repeated at larger scale: Schlitz’s participants were significantly less activated during stare than non-stare periods (z = −1.93, p = .05; effect size −0.33; 35 participants, 1,120 trials), a direction opposite to the first study, while Wiseman’s participants again did not differ from chance (z = −0.39, p = .69; effect size −0.07; 35 participants, 1,120 trials); in neither study did the direct comparison between the two arms reach significance.^[20] In the third (2006), run with Watt and Radin in a shielded chamber with greeter and sender roles crossed, both key conditions returned null results: Schlitz as both greeter and sender gave z = −0.17, p = .87 (25 participants, 1,000 trials), and Wiseman in both roles gave z = −0.35, p = .72 (26 participants, 1,040 trials); no effect of who greeted (F(1, 93) = 0.46, p = .50) or who sent (F(1, 93) = 0.21, p = .64) was found, and coded greeter–participant rapport did not correlate with session outcome (r = −0.028, p = .86).^[24]

Schmidt, Schneider, Utts, and Walach reported that, in their 40-study DMILS set, effect size correlated negatively with overall study quality (r = −.43) and with the number of methodological safeguards (r = −.53), and their best-evidence synthesis of the 7 highest-quality studies was nonsignificant (d = .05, p = .50).^[23] In the staring line, Baker and Stevens’s three EEG experiments found that peak brain-response differences tracked the camera-feed manipulation rather than the physical presence of a starer, with the effect’s direction reversing between experiments, a pattern the authors discuss in terms of possible methodological artifact.^[26]

Replication failures temper the positive meta-analytic picture. Preregistered attempts at related mind-matter paradigms have failed to reproduce earlier positive findings, and some authors argue that certain designs may be inherently resistant to confirmation under standard signal-transfer assumptions.^[11] Critics within parapsychology have also warned against over-reading long-term experimenter and decline effects that do not survive rigorous meta-analytic scrutiny.^[12] Methodological debate over inclusion criteria, heterogeneity, and effect-size metrics in psi meta-analyses continues to be vigorous.^[13]

A separate caution concerns translation to medicine. Even authors sympathetic to DMILS note that laboratory proof-of-principle has not yielded reliable clinical healing, and they attribute the gap in part to a mismatch between conventional trial protocols and the phenomenon being tested.^[1] This reinforces the boundary between DMILS as a laboratory method and distant healing as a clinical claim.

Current practice

Contemporary work continues to refine the autonomic-measure approach and to probe what moderates the effect. The cancer-partner study illustrates the current emphasis on participant selection, training and motivation, as variables that may strengthen distant intention effects, alongside double-blind safeguards.^[6] Researchers also continue to examine experimenter expectancy directly, treating it as a measurable factor rather than a nuisance, an emphasis traceable to the staring collaboration.

Some investigators have extended distant-intention designs to electronic and random-event targets, reporting proof-of-concept influence on random-number-driven devices under preregistration, which blurs the historical line between Bio-PK and micro-PK paradigms.^[14] Reviews that survey the wider field, including DMILS, increasingly frame the results within debates about whether consciousness can be treated as nonlocal rather than as a simple product of brain activity.^[15] Methodologists such as Caroline Watt, PhD have pressed for preregistration and transparent reporting, while statisticians including Jessica M. Utts, PhD have shaped how the cumulative evidence is evaluated. The practical state of DMILS is a paradigm with a long methodological pedigree, modest but persistent meta-analytic support, and unresolved questions about experimenter influence and replication that keep it firmly within the boundaries of contested science.