Regulating Homeostatic Plasticity and the Physiological Response to rTMS: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Regulating Homeostatic Plasticity and the Physiological Response to rTMS (NCT03309696), a randomized na study. The primary outcome is Log Transformed P100 Amplitude of TEPs From the Global Mean Field Analysis..

2. Clinical Question

3. Trial Design

The registry classifies the study as randomized, double-masked, and parallel, with a planned enrollment of 10. Its title is “Regulating Homeostatic Plasticity and the Physiological Response to rTMS.” The study is identified as a device study and lists University of Arkansas as sponsor. Its status is terminated. The recorded study start was 2017-11-16, and the primary completion date was 2019-10-01.

The main rationale was that the physiological effects of non-invasive neuromodulation may be difficult to interpret when the initial activation state of the targeted brain circuit is not measured or controlled. The study proposed using tDCS before rTMS to alter that starting state, then examining physiological responses with EEG. The registry describes homeostatic plasticity as relevant to whether rTMS induces long term potentiation or long term depression; it does not report a clinical outcome intended to establish treatment benefit.

Arm descriptions cover tDCS over the temporal cortex or dorsolateral frontal cortex before 1 Hz or 10 Hz rTMS over the temporal cortex. They also describe sham and active stimulation conditions. In the listed descriptions, active tDCS is 2mA, and active rTMS involves 900 rTMS pulses at 110% motor threshold. The intervention list names three combinations, while the arm descriptions set out several stimulation sequences; the record does not explain this difference in detail.

The primary outcome timeframe is up to 8 weeks. The primary outcome population reported in the registry is 5 subjects assigned to the temporal cortex tDCS and 1 Hz rTMS arm. This is an analysis-population statement, not an explanation of why the study was terminated or why that population was used. A sample-size justification, allocation ratio, and detailed protocol for the study’s conduct are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The analyses population is 5 subjects who were assigned to the arm "tDCS and 1 Hz rTMS over the temporal cortex". The group titles reflect sequences of sham and active tDCS and rTMS conditions used to create contrasts for data analysis. Outcome measures are not reported for the DLPF arm because, due to insufficient resources and the required personnel needed to perform the extensive data cleaning and pipeline analysis, no P100 amplitude data can be reported.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Baseline characteristics are reported for the two arms to which subjects were assigned.

6. Primary Endpoint

The registry lists the primary outcome as Log Transformed P100 Amplitude of TEPs From the Global Mean Field Analysis., measured over Up to 8 weeks.

The primary endpoint is the log-transformed P100 amplitude of TMS-evoked EEG potentials, abbreviated as TEPs, and expressed in log µV. A TEP is an EEG response associated with TMS. The registry describes P100 amplitude as one way to assess cortical excitability and notes that it has been shown to be a reliable metric in studies of healthy subjects.

The endpoint is used to assess the excitation state of two regions of interest: the temporal cortex and the DLPFC. Measurements are described at baseline, after tDCS, after rTMS, and after a 20 minute delay. The stated timeframe for the primary outcome is up to 8 weeks. The registry identifies the outcome as a continuous measure and reports group means and standard deviations.

These measurements are physiological outcomes. They are not themselves measures of tinnitus symptoms or clinical improvement. The registry says that an exploratory aim concerning physiological changes in patients with tinnitus could not be completed because of funding limitations; it was not part of the pilot physiological investigation. Outcome measures for the DLPFC arm are not reported in the registry record. The remaining posted endpoint results therefore describe a limited analysis population and should be interpreted as physiological findings for the reported stimulation conditions, not as evidence about a patient-centered treatment effect.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports group means and standard deviations for a continuous outcome and gives four mixed-model comparisons. A mean describes the central value of a set of measurements. A standard deviation describes how dispersed those measurements are around their mean; it is not a confidence interval and does not directly show the precision of the estimated mean. The reported outcome is log-transformed P100 amplitude, so its stated unit is log µV.

What does a mixed model do?

In general, a mixed model can represent more than one source of variation in data, including correlation among repeated measurements from the same participant. Such models can be useful when outcomes are measured at multiple times or under multiple conditions. The registry names mixed models as the method for the reported comparisons, but the model specification is not reported in the registry record. Details such as the fixed and random effects, covariance structure, adjustment variables, treatment of incomplete observations, and exact contrast definitions are not reported. The method label alone does not show which of these features the fitted analyses used.

How should a mean difference be read?

A mean difference describes the contrast between two conditions on the outcome scale defined for the analysis. The record labels the estimates “Mean Difference (Net)” and reports the values alongside p values. The posted results do not include confidence intervals for these estimates. In general, a confidence interval expresses the uncertainty around an estimated effect and indicates which values are reasonably compatible with the data and model. Without one, the range of plausible effects cannot be read from these results. The p value does not supply that range and does not measure the probability that a hypothesis is true.

What does an equivalence analysis require?

In general, an equivalence analysis asks whether a difference is sufficiently small to fall within a prespecified range considered practically negligible. That range is called the equivalence margin. The analysis must assess the estimated difference and its uncertainty in relation to that margin. Although the registry classifies all four posted analyses as equivalence analyses, it does not report the margin or confidence intervals. It is therefore not possible to verify the equivalence conclusion from the reported figures. The reported p values should not be treated as a substitute for those missing elements.

Why can repeated measurements and dispersion matter?

Measurements taken from the same participant are generally not independent observations. In general, a model that accounts for this dependence can estimate contrasts more appropriately than an analysis that treats every measurement as unrelated. The endpoint description includes multiple recording periods, and the study compares sequences of stimulation conditions. The record does not report how the model represented those features. Standard deviations describe observed dispersion, but they do not resolve uncertainty about model assumptions or the precision of a contrast. If outcomes are more variable than a model assumes, standard errors and tests may be poorly calibrated; this is a general reason to examine model fit and variance assumptions, not a claim that over-dispersion occurred in this study.

What do planning calculations mean?

In general, a sample-size illustration based on an observed effect is a planning exercise, not an estimate of the power achieved by the completed study. The stated assumptions describe a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and effects set as a stated fraction of the observed effect. The illustration is not the trial’s own design justification. It should not be read as observed power or as evidence that the study was adequately sized. The registry does not report the study’s original sample-size justification.

The record reports four analyses, but does not describe whether or how multiplicity was handled. In general, examining multiple contrasts can increase the chance of a false-positive finding if no suitable approach is used. The absence of a reported adjustment does not establish that no adjustment was used; the analysis plan is simply not reported in the registry record. The same caution applies to interim analyses, which are not reported. These missing methodological details limit how fully the mixed-model results can be evaluated or reproduced.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Sham tDCS Preconditioning", "Active tDCS Preconditioning", "Sham tDCS Preconditioning of Sham rTMS", "Sham tDCS Preconditioning of Active rTMS", "Active tDCS Preconditioning of Active rTMS")
n   <- c(5, 5, 5, 5, 5)
m   <- c(1.85, 1.26, 1.59, 1.3, 1.11)
sp  <- c(0.71, 0.33, 0.82, 0.58, 0.48)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Sham tDCS Preconditioning  5  1.850  0.710  0.968  2.732
CARM  Active tDCS Preconditioning  5  1.260  0.330  0.850  1.670
CARM  Sham tDCS Preconditioning of Sham rTMS  5  1.590  0.820  0.572  2.608
CARM  Sham tDCS Preconditioning of Active rTMS  5  1.300  0.580  0.580  2.020
CARM  Active tDCS Preconditioning of Active rTMS  5  1.110  0.480  0.514  1.706

11. Analyses Reported by the Sponsor

The registry’s posted comparisons are identified as mixed-model analyses and report mean differences (net) with p values. They are the sponsor-reported analyses in the record; no separate recalculated estimate is presented in the reported analysis results. There is no discrepancy flag indicating disagreement between a sponsor result and a recalculation. The appropriate interpretation is therefore to describe the model results as recorded, without replacing them with a newly derived contrast.

That does not remove the limitations of the reported analysis. Confidence intervals and the equivalence margin are not reported in the registry record, and the model specification is not described. As a result, the uncertainty around the estimates and the basis for the equivalence labels cannot be assessed from the posted figures alone. The descriptive group means and standard deviations provide context for the outcome values, but they do not independently establish the mixed-model result or its assumptions.

The registry identifies the analysis population as 5 subjects assigned to the tDCS and 1 Hz rTMS arm over the temporal cortex. It does not report outcome measures for the DLPFC arm, and it does not explain the population selection in sufficient detail to evaluate how broadly the estimates apply. The sponsor’s reported analyses should thus be read as results for the listed contrasts and population, not as a complete account of every planned arm or the exploratory tinnitus aim.

12. Multiplicity, Missing Data and Analysis Populations

The registry reports four mixed-model comparisons for the primary continuous outcome. It does not report a multiplicity strategy, such as an adjustment to account for testing several contrasts or a hierarchy that identifies which comparison should be treated as primary. In general, the more comparisons examined, the more important it is to understand whether the analysis plan accounted for the opportunity to observe a small p value by chance. The absence of a reported strategy limits interpretation of the individual p values, particularly when considered together. It does not establish that no strategy was used; details are not reported in the registry record.

The primary outcome population is stated as 5 subjects assigned to the arm using tDCS and 1 Hz rTMS over the temporal cortex. The posted group summaries list five subjects for each named group. The record does not explain whether those group entries represent repeated conditions within the same participants or how they relate to the stated analysis population. Outcome measures for the DLPFC arm are not reported, and the reason text for that omission is incomplete. The number and pattern of missing outcome observations, reasons for missingness, and any imputation or other missing-data method are not reported in the registry record.

The record does not specify whether the analysis used an intention-to-treat population, a per-protocol population, or another population definition. It does not provide enough detail to determine how participants were included in each model or how deviations from the assigned conditions were handled. The protocol’s interim-analysis plan and the statistical analysis plan are also not reported in the registry record. These omissions matter because they affect the interpretation of who contributed to each estimate, how the comparisons were selected, and whether the reported uncertainty reflects all relevant sources of variation.

The exploratory aim concerning patients with tinnitus was not completed because of funding limitations and was not part of the pilot physiological investigation. Accordingly, the posted primary outcome analyses do not provide results for that exploratory aim. No secondary outcome results are posted. The registry record therefore offers a narrow analysis of a physiological endpoint, with limited information about missing data, analysis populations, and multiplicity.

13. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports cumulative adverse-event figures for a group with 10 at risk. It records 0 serious adverse events and 1 other adverse event, with 0 deaths. No serious adverse-event terms are listed. These are the posted counts; the record does not provide additional event descriptions in this safety summary.

Serious and other adverse events are recorded separately, so the other adverse event count should not be combined with the serious-event count or interpreted as a serious event. The record does not provide details about the nature, timing, severity, or relationship of the other event to a study intervention. Nor does it provide enough information to judge whether event collection was complete across all stimulation conditions or follow-up periods.

These counts describe what was reported for this study population. With the limited information in the registry, they do not establish the overall safety profile of tDCS or rTMS, and they should not be generalized beyond the study. The registry’s terminated status does not, by itself, explain the adverse-event figures or indicate a safety-related reason for termination.

14. What the Result Does, and Does Not, Mean

The posted results describe a small physiological investigation of P100 amplitude under selected tDCS and rTMS conditions. They give group means, standard deviations, and mixed-model mean differences for four comparisons. The results can be read as an initial account of how the measured EEG endpoint varied across the conditions represented in the reported analyses.

They do not establish that a stimulation condition is clinically effective or ineffective. The endpoint is a physiological measure of cortical excitability, not a reported measure of tinnitus severity, daily function, or another patient-centered outcome. The exploratory aim in patients with tinnitus was not completed. Therefore, the results do not answer whether the approach changes tinnitus or benefits patients.

The equivalence labels also require caution. The registry does not provide the equivalence margin or confidence intervals, so the data cannot be checked against a defined threshold for equivalence or judged for precision in the usual way. A p value alone cannot establish equivalence, and the reported p values should not be treated as direct measures of effect size or clinical importance.

The results are further constrained by the stated analysis population, the absence of DLPFC arm outcome measures, and missing details about the model, missing observations, and multiplicity. The study’s terminated status and the funding limitation affecting the exploratory aim add context, but do not explain the statistical findings beyond what is explicitly reported. The most defensible reading is that the registry presents limited, condition-specific physiological results whose uncertainty and generalizability cannot be fully evaluated from the recorded information.

15. Why This Trial Matters Statistically

This trial raises a statistical question that is important in physiological research: how should an intervention’s response be understood when the starting state of the targeted circuit may affect the observed response? The study’s rationale was to use tDCS preconditioning before rTMS and measure EEG responses. That design connects a mechanistic hypothesis with a measurable continuous outcome, while also making the analysis dependent on careful handling of repeated observations and condition contrasts.

The registry illustrates why naming a statistical method is not enough to make a result fully interpretable. A mixed-model label indicates a broad class of methods, but readers also need the model specification, the definition of each contrast, and an account of how repeated measurements and incomplete data were handled. The absence of confidence intervals and an equivalence margin is especially consequential because the analyses are labelled as equivalence analyses. Without those details, the conclusion cannot be checked against a stated criterion.

The record also shows the value of distinguishing physiological aims from clinical aims. P100 amplitude is described as a measure of cortical excitability, whereas the exploratory tinnitus aim was not completed. Keeping those aims separate prevents an EEG finding from being mistaken for evidence about patient symptoms or treatment benefit.

Finally, a small pilot analysis can help identify questions for future study, but it cannot by itself settle them. Clear reporting of populations, missing data, multiplicity, model assumptions, and uncertainty would allow readers to understand what the estimates support and what remains unresolved. Those reporting needs are central to evaluating mechanistic experiments and to planning later research.

16. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The analysis population reported for the primary outcome is 5 subjects assigned to the temporal cortex tDCS and 1 Hz rTMS arm. The registry does not explain how that population relates to every listed condition or how participants contributed to each comparison. Although the arm summaries each list five subjects, the record does not clarify the structure of those observations. This limits evaluation of the effective information available for each contrast.

The statistical reporting is incomplete for assessing precision and equivalence. Confidence intervals are absent, and the equivalence margin is not reported. The model specification, handling of repeated measurements, treatment of missing data, and definition of the analysis population are also not reported in the registry record. No multiplicity plan is described. These are specific gaps for interpreting the four posted comparisons, not evidence that any particular modeling or adjustment choice was made.

Outcome measures for the DLPFC arm are not reported, and the record’s explanation is unfinished. The exploratory aim in patients with tinnitus could not be completed because of funding limitations and was separate from the pilot physiological investigation. Thus, the posted results do not cover the full set of research questions described in the summary.

The study is identified as terminated, but the registry does not state why the study was terminated. The funding limitation is reported as the reason the exploratory aim could not be completed; it should not be assumed to explain the termination itself. The posted safety summary contains limited event detail, and no serious adverse-event terms are listed. These constraints, together with the narrow physiological endpoint and limited statistical detail, make the findings difficult to generalize beyond the reported analyses.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Regulating Homeostatic Plasticity and the Physiological Response to rTMS — NCT03309696.
• Registry record last updated 2020-11-17; status last verified 2020-11.
• Results first posted 2020-11-06.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

20. Record Summary

NCT03309696 is a randomized, double-masked, parallel device study sponsored by University of Arkansas and listed as terminated. It investigated tDCS preconditioning before rTMS, using TMS-evoked EEG P100 amplitude as the primary continuous outcome. The registry reports four mixed-model comparisons for the stated analysis population of 5 subjects assigned to the temporal cortex tDCS and 1 Hz rTMS arm.

The posted results include group means and standard deviations, mean differences (net), and p values. The analyses are labelled as equivalence analyses, but confidence intervals and an equivalence margin are not reported. DLPFC arm outcome measures are not reported. The exploratory aim in patients with tinnitus was not completed because of funding limitations and was not part of the pilot physiological investigation. The results should be interpreted as limited physiological findings, not as evidence about clinical benefit.