Suppose you wanted to find out — actually find out, not gesture at — whether a given hands-on or intention-based practice produces a measurable physiological change. What would the study look like?
This is not a rhetorical question. The methodological requirements are well established, uncontroversial, and largely absent from the existing literature in this area. Setting them out plainly clarifies what a serious research program here would have to commit to, and makes visible how much of the current evidence base falls short of it.
1. State a claim that could be false
The first requirement is a hypothesis specific enough to fail. “Practice X promotes wellbeing” cannot fail; every result is compatible with it. “Practice X produces a change in metric M, measured immediately post-session, that differs from the change produced by control condition C” can fail, and therefore means something.
This requires committing to a single primary endpoint in advance. One — not a panel of twenty metrics from which the most cooperative will be selected afterward. With twenty uncorrelated metrics at conventional thresholds, you should expect one apparent finding by chance alone, and the literature in contested fields is full of exactly that finding, reported as though it had been the plan.
2. Pre-register
The pre-registration — on ClinicalTrials.gov, the Open Science Framework, or an equivalent public registry — should be timestamped before the first participant is enrolled, and should specify the primary endpoint, the sample size and the power calculation behind it, the randomization scheme, the analysis method, the handling of missing data, and the stopping rule.
The purpose is not bureaucratic. A public, dated record of the plan removes the researcher’s ability to make dozens of small, individually defensible analytic choices after seeing the data. Those choices are how well-intentioned people generate false positives. In a field where prior probability is contested and scrutiny will be intense, pre-registration is what separates a result from an anecdote with statistics attached.
3. Control the right things
Randomization to condition, with allocation concealed from whoever enrolls participants. That much is standard. The harder question in this domain is what the control condition should be.
The control has to be matched to the intervention on everything except the intervention itself. Posture, room, lighting, privacy, duration, and whether another person is present all move physiological measures on their own. Any difference between arms that is not the thing under test becomes a rival explanation for whatever you find — and if that difference happens to push in the same direction as your hypothesis, no result can distinguish them.
The intuitive control — a sham practitioner who mimics the movements without the intention — handles the participant’s experience but not the practitioner’s. And the practitioner cannot be blinded to their own intention; that is a structural feature of the question, not a design flaw to be engineered away. The burden of blinding therefore falls on the participant and the analyst, and a study that does not say so is not being honest about its limitations.
4. Use an objective endpoint
This is the piece the series argues has been missing, and it is worth being precise about the role it plays. An objective physiological endpoint does not make a study rigorous by itself. It removes one specific failure mode — the contamination of self-report by expectation — and it makes blinded analysis possible, because a number produced by an instrument can be handed to an analyst who does not know which arm it came from.
It also imposes an obligation that is easy to skip: characterizing the instrument first. Before any interventional data is collected, the study should establish the endpoint’s test-retest variability under the exact conditions of the protocol — same room, same interval, same time of day, no intervention. Without that number, a post-session change score cannot be interpreted, because there is nothing to compare it against.
This is where the Aurorus platform belongs in a study design of this kind: as one component, the endpoint, in a study designed and run by someone else. Not as the study. The distinction matters commercially as well as scientifically — we license a measurement, and the credibility of the measurement depends on our not owning the conclusion.
5. Power the study honestly
Effect sizes in this domain, if they exist at all, are likely to be small. Small effects require large samples. A pilot of twenty participants cannot detect a small effect, and a pilot of twenty participants reporting a significant finding should prompt caution rather than excitement — the most common explanation for a large effect in a small sample is chance, followed closely by an analytic choice made after the fact.
The appropriate framing for early work is therefore explicit: a pilot or feasibility study estimates variance, tests logistics, and generates a hypothesis for a properly powered confirmatory trial. It is not designed to establish anything, and reporting it as though it were is the most common way this literature loses credibility.
6. Analyze as planned, then replicate
The pre-specified analysis is the result. Everything else is exploratory and should be labelled as such, without embarrassment — exploratory analysis is legitimate and valuable, and the only thing that makes it illegitimate is presenting it as confirmatory.
One point of technique worth naming, because it is common in this field and easy to make in good faith: the comparison that answers the question is between arms, not within them. Testing before against after inside each arm separately tells you whether something changed during each session. It does not tell you whether one changed more than the other. That requires comparing the change scores directly.
Then the result has to survive independent replication, ideally by a group with no stake in the outcome and no relationship to the instrument manufacturer. A single positive study in a contested field is a reason to run another study. It is not a finding.
What our own pilots suggest is worth considering
We have run two pilots in this area. They are useful here less as results than as a source of concrete lessons, and we mention them briefly for that reason.
The first compared Reiki delivered by an experienced practitioner against a sham condition, with participants blinded to which they received. It returned no significant difference. The second compared an intention-specific Reiki session against a relaxation session of the same duration, with the practitioners’ intention fixed in advance on the heart and chest region; the largest responses appeared in the cardiovascular and respiratory measures.[1]
Neither establishes anything on its own. Both are pilots, which is exactly what pilots are for. What they offer is a short list of items worth considering for anyone designing in this space:
- Match the control on everything but the intervention. Posture and undisturbed quiet move cardiorespiratory measures by themselves, so a control that differs in those respects cannot cleanly answer a question about cardiorespiratory response.
- Compare the arms directly, as above, rather than testing each arm against itself.
- Pre-specify the target. Naming a region in advance and predicting the response will concentrate there is a far stronger test than predicting that something, somewhere, will shift.
- Report the nulls. The sham-controlled pilot found nothing, and that result belongs on the record alongside the other one.
The last two points together suggest a design worth building. Hold everything constant between arms — same room, same posture, same duration, same practitioner present — and randomize only the target of the intention. The prediction then is not that something changes, but that the change appears where it was aimed and not elsewhere. That is a hard prediction to satisfy by accident, and a clean one to fail. A full reanalysis of the second pilot along these lines is in preparation.
Why lay this out
Two reasons. The first is practical: any group planning research in this area now has a checklist, and any reader evaluating a study in this area now has one too.
The second is that a field’s seriousness is measured by the standards it volunteers for. It is easy to demand rigor of one’s critics. The useful move is to specify, in advance and in public, the conditions under which one’s own preferred hypothesis would be abandoned.
[1] Rizzo NR. Intention-Specific Reiki vs. Simple Relaxation: Quantifying the Effects on the Physical and Nervous System Biofields. Presented at: International Reiki Research Conference; 2023. Randomized two-round crossover pilot, n=20.
Regulatory note: Aurorus is a non-invasive electrophysiological measurement platform. It was granted FDA De Novo classification (DEN150004) as a Class I non-invasive electrophysiological measurement device. It is not cleared for diagnostic purposes. Nothing in this article claims that any therapy, modality, or practice diagnoses, treats, cures, or prevents any disease.
End · 7 min
Nancy Rizzo Prior to developing the science and technology for Aurorus, Nancy spent 13 years at Mayo Clinic Scottsdale as a board certified radiological therapeutic physicist and cochaired the Complementary and Alternative Medicine program. Nancy is board certified as a Radiological Therapeutic Physicist by the American Board of Radiology (DABR) and holds a Ph D. in Energy Medicine with an emphasis on Biophysics from Holos University.
Follow →