Spinal Plasticity to Enhance Motor Retraining After Stroke: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Spinal Plasticity to Enhance Motor Retraining After Stroke (NCT03645122), a randomized na study. The primary outcome is Motor Control & Learning.

2. Clinical Question

The study asks whether paired corticospinal-motoneuronal stimulation (PCMS), delivered under different hand and digit conditions, is associated with a change in motor control and learning after stroke. The primary outcome is based on a force-tracking task: variability around the mean force required to follow a target is measured in Newtons. Lower or higher values describe different amounts of variability, but the registry record does not define a threshold for a meaningful difference or state which direction would be clinically important.

The study is listed as randomized and crossover, so the conditions are intended to be compared within a crossover design rather than as conventional parallel groups. However, the registry record does not report the order of conditions, whether the order was randomized, or how the analysis accounted for repeated measurements within participants. Those details matter for interpreting comparisons, because observations from the same person may be related.

3. Trial Design

The registry describes a randomized, unmasked crossover study with an enrollment of 23. The study is marked completed, and its phase is listed as not applicable. The sponsor is the VA Office of Research and Development. The record lists the study start as 2019-02-04 and primary completion as 2023-09-29.

Three conditions are reported: rest PCMS, sham PCMS, and active PCMS. In the rest PCMS condition, stimulation is triggered while the hand and digits are at rest. In the active PCMS condition, stimulation is triggered when a criterion level of force stability is achieved. The sham condition is also triggered when that criterion level is achieved. The primary outcome is measured under each condition, with approximately 2 months required for completion of all conditions.

The registry lists 23 participants for each of the three conditions. In a crossover study, the same participants may contribute measurements under multiple conditions; these condition counts should not be interpreted as independent groups of different people. The record does not specify participant flow through each condition, whether every participant completed every condition, or how missing condition measurements were handled.

Masking is listed as none. The registry does not describe whether outcome assessors or analysts were masked, and it does not report the sequence of conditions, any washout interval, or the interval between conditions. The trial's original sample-size assumptions, planned statistical model, and planned analysis population are not reported in the registry record. The posted comparisons therefore do not provide a complete account of the intended crossover analysis.

The study objective concerns an emerging noninvasive neuromodulation strategy in people with stroke-related movement disability. The registry's brief summary proposes that muscle weakness after stroke may involve abnormal interaction between brain cells that send movement commands and spinal cord cells that activate muscles. The outcome is a laboratory force-tracking measure; the record does not report a direct measure of daily function or a clinical threshold for interpreting its size.

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

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.

6. Primary Endpoint

The registry lists the primary outcome as Motor Control & Learning, measured over Measured under each condition, with approximately 2 months required for completion of all conditions..

The primary outcome is titled Motor Control & Learning. It measures variability around the mean force required to track a target in a force-tracking task, with results expressed in Newtons. The registry says this measure is obtained under each condition. The condition means are reported as 0.339 for rest PCMS, 0.359 for sham PCMS, and 0.301 for active PCMS. The accompanying spread is identified as standard error.

Force-tracking variability is used as a continuous measure of motor performance: it describes how much the applied force varies around the mean force needed for the task. A continuous measure can retain gradations in performance rather than reducing observations to a binary category. In general, this can help characterize differences in task performance across experimental conditions. The record does not specify a clinically meaningful difference, a normal reference range, or a threshold that would distinguish meaningful from unimportant variation.

The outcome can reflect performance on the specified task, but the registry does not establish that it measures all aspects of motor control or learning in everyday settings. It also does not describe how many tracking trials were taken, how variability was calculated beyond its relation to mean force, or whether repeated task measurements were averaged before analysis. Those measurement details are not reported in the registry record.

The timeframe is described as measurement under each condition, with approximately 2 months required to complete all conditions. The record does not state a separate fixed assessment visit or specify the period between condition measurements. As a result, the outcome's timing is defined by the condition-based measurement process rather than by a single follow-up time stated in the record.

7. Results Posted in the Registry

• Rest PCMS: mean 0.34 versus 0.36 on control; difference -0.02 (95% CI -0.41 to 0.37), standardized effect d = -0.03; z-test p = 0.920.
• Active PCMS: mean 0.30 versus 0.36 on control; difference -0.06 (95% CI -0.41 to 0.30), standardized effect d = -0.09; z-test p = 0.748.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a mean force-tracking variability measure expressed in Newtons. A mean summarizes the average outcome across observations, while the standard error describes uncertainty in an estimated mean. Standard error is not the same as the spread of individual measurements. The registry reports standard errors of 0.141 for rest PCMS, 0.139 for sham PCMS, and 0.115 for active PCMS. Without individual-level results or a standard deviation for each condition, these summaries do not reveal the full variability among participants.

What does a mean difference represent?

A mean difference compares the reported average for one condition with the average for sham PCMS. The rest PCMS comparison is reported as -0.02, and the active PCMS comparison as -0.058. A negative estimate means the condition mean is numerically lower than the sham mean under the comparison's stated direction. This description is arithmetic interpretation of the posted estimates; it does not establish that the difference is meaningful or that it represents a causal effect in the intended crossover analysis.

A confidence interval describes a range of values compatible with the estimate and its statistical uncertainty under the method used to calculate it. The intervals reported for these comparisons extend from -0.408 to 0.368 for rest PCMS and from -0.412 to 0.296 for active PCMS. Both include zero, the value corresponding to no mean difference. The confidence level and the method used to calculate these intervals are not reported in the registry record, so the intervals should be interpreted as posted without assuming an unreported confidence level or model.

Why does crossover structure matter?

In a crossover study, the same participant may receive or experience more than one condition. Measurements from that participant can be correlated because they share characteristics of the same person. A paired analysis uses this within-person relationship when estimating condition differences. If the relationship is not accounted for, uncertainty may be estimated differently from a properly paired analysis. The registry identifies the design as crossover but does not report paired outcome data, the within-person covariance, or an analysis model that uses the pairing.

The posted comparisons include pooled standard deviations and standardized effect estimates, but the registry record does not explain the analysis procedure that generated them. The displayed confidence intervals and p-values should therefore be understood as summary-based, unadjusted comparisons rather than a complete report of a prespecified crossover model. The study's original analysis plan is not reported in the registry record.

How should the p-values be read?

A p-value describes how compatible the observed data are with a specified statistical null model, assuming that model and its assumptions are appropriate. The reported unadjusted p-values are 0.92 for rest PCMS versus sham PCMS and 0.748 for active PCMS versus sham PCMS. These values do not state the probability that a treatment effect is true, and they do not measure the size or clinical importance of a difference. Their interpretation depends on the statistical model, which is not described in sufficient detail in the record.

Because the design is crossover and the posted calculations are unadjusted summaries, the p-values should not be treated as definitive evidence from a reported paired analysis. The registry does not report whether any prespecified adjustment, covariate, period effect, sequence effect, or other crossover feature was included. None should be assumed from the condition summaries alone.

What does over-dispersion mean in general?

In general, over-dispersion occurs when observed variation is greater than a statistical model expects. It can make standard errors too small if the model underestimates the true variability, resulting in intervals that are too narrow and tests that are too optimistic. That issue is particularly relevant to count outcomes and models that make restrictive assumptions about variation. The primary measure in this study is continuous, and the registry does not report a model assessment or an over-dispersion analysis. It is therefore not possible to say that over-dispersion was present or that it affected these results.

More generally, a statistical model should reflect the outcome scale, the way observations are collected, and the dependence among observations. For repeated measures, the within-participant structure is important. The registry record does not describe model diagnostics, assumptions checks, or sensitivity analyses for the primary outcome.

What can be concluded from the posted summaries?

The condition means, standard errors, comparison estimates, intervals, and p-values provide a limited statistical description of the force-tracking outcome. They show the direction and imprecision of the reported summary contrasts. They do not reveal individual trajectories, establish the appropriateness of an analysis for the crossover design, or support conclusions about outcomes that were not posted. The registry reports no analyses, so the stated comparisons should not be mistaken for a complete registry account of the trial's statistical analysis.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data locate an estimated difference. The posted comparison for rest PCMS versus sham PCMS is -0.02, with an interval from -0.408 to 0.368. For active PCMS versus sham PCMS, the estimate is -0.058, with an interval from -0.412 to 0.296. Both intervals cover zero and include estimates in both directions. Thus the point estimates alone give an incomplete picture: the intervals show that the posted comparisons are uncertain about both the direction and magnitude of the difference.

The intervals are wide relative to the point estimates. This means the posted results do not pinpoint a narrow range of plausible differences. The record does not report the confidence level or the interval calculation method. It also does not provide the paired covariance needed to assess how crossover dependence was handled. Those omissions limit how specifically the interval widths can be interpreted.

The standard errors for the three condition means are 0.141 for rest PCMS, 0.139 for sham PCMS, and 0.115 for active PCMS. These describe uncertainty around the reported condition means, not the variability of all participant-level observations. Nor do they, by themselves, supply the uncertainty for a paired difference. The comparison intervals are the relevant posted summaries for contrast precision, but their analytic basis is not fully described.

The registry record does not report a clinically meaningful difference for this force-tracking measure. Statistical precision and clinical importance are distinct: even a precisely estimated difference needs an interpretable threshold, while an imprecise estimate may leave both small and larger differences plausible. Here, the record does not give a threshold with which to judge clinical relevance.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the trial's original design assumptions. They are stated as participants per arm for 80% power and participants per arm for 90% power. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect. These calculations are planning scenarios, not observed power.

For rest PCMS versus sham PCMS, the table gives participants per arm for detecting 100% of the observed effect as 17,693 for 80% power and 23,685 for 90% power. For planning for 75% of the observed effect, the figures are 31,453 and 42,107; for planning for 50% of the observed effect, they are 70,770 and 94,740. For active PCMS versus sham PCMS, the corresponding figures are 1,747 and 2,339 for the full observed effect, 3,105 and 4,157 for 75% of the observed effect, and 6,987 and 9,353 for 50% of the observed effect.

The 80% power column describes the planned probability of detecting the specified effect under the calculation assumptions if that effect were the true effect. The 90% column uses a higher target probability of detection, which requires more participants in these illustrations. Smaller effects require more participants because they are harder to distinguish from random variation. The plotted sample-size curve should be read in the same planning context: it illustrates how the calculated requirement changes with the effect size being targeted. It is not a record of actual enrollment or actual power.

The crossover design and its within-person dependence are not described in these planning assumptions. The original protocol's sample-size justification and its assumptions are not reported in the registry record, so these illustrations should not be treated as a reconstruction of the trial's intended design calculation.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.020; pooled standard deviation σ = 0.671.
2. n = 2 × (1.96 + 0.842)² × 0.671² / -0.020² = 17,692.3, rounded up to 17,693 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.02                      # observed difference in means
sigma <- 0.671433541015043                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(3645122)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 17,693 (80% power), 23,685 (90% power)
 75% of observed effect: n per arm = 31,453 (80% power), 42,107 (90% power)
 50% of observed effect: n per arm = 70,770 (80% power), 94,740 (90% power)
power.t.test: 17693.2 
Simulated power at n = 17,693 per arm: 0.798 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and two posted comparisons, each contrasting a PCMS condition with sham PCMS. It does not report a multiplicity adjustment or a prespecified strategy for controlling the chance of false-positive findings across comparisons. The reported p-values are identified as unadjusted. Without an analysis plan, it is not possible to determine how the comparisons were intended to be interpreted together or whether any adjustment had been planned.

The record lists no secondary outcomes and no posted analyses. This narrows the scope of the results available in the registry, but it does not establish what outcomes or analyses may have been considered in the original protocol or elsewhere. The registry record does not report those details.

Missing-data handling is not reported. The record does not state whether all participants completed every condition, how incomplete force-tracking measurements were treated, or whether any imputation or other method was used. In a crossover study, missing data for a condition can affect the available within-person comparisons, so participant flow by condition is important context.

Analysis populations are also not reported. The registry does not identify whether results were based on all randomized participants, participants who completed each condition, or another defined group. Although 23 participants are listed for each condition, those counts do not establish the analysis population or show that the same set of participants contributed to every comparison.

The record does not report a statistical analysis plan, covariate adjustment, sensitivity analysis, or model for repeated measurements. It also does not describe whether period or sequence effects were assessed. Those omissions prevent a full assessment of multiplicity, missingness, and the consistency of the reported comparisons with the intended crossover analysis.

14. 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.

The registry reports zero serious adverse events among 23 participants at risk in each condition: rest PCMS, sham PCMS, and active PCMS. It also reports zero deaths in each condition. These figures describe the events recorded in the registry for the stated groups and at-risk counts.

A count of zero reported events is not proof that an event could never occur, and it does not establish that no participant experienced any symptom or other harm. The registry record does not provide additional detail about event definitions, event collection procedures, duration of safety follow-up, or whether events were assessed separately for each condition in a crossover sequence. The listed at-risk counts should not be added as if they necessarily represented separate groups of people, because the study is described as crossover.

The record lists no serious adverse event terms. The safety figures should therefore be read as registry-reported counts for this study, not as a complete characterization of safety in other populations or settings.

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

The posted results describe force-tracking variability under three conditions in a randomized crossover study with an enrollment of 23. The condition mean is 0.339 for rest PCMS, 0.359 for sham PCMS, and 0.301 for active PCMS. The unadjusted summary-based comparisons with sham are -0.02 for rest PCMS and -0.058 for active PCMS. Their reported intervals include zero, and the estimates are imprecise.

These results do not establish that either PCMS condition differs from sham, but they also do not establish that the conditions are equivalent or that no difference exists. The intervals allow a range of possible differences, and the record does not state a threshold for clinical importance. A high p-value is not evidence that the null hypothesis is certainly true, nor does it show that the study had adequate precision for every effect that might matter.

The outcome concerns performance on a force-tracking task. It is not, by itself, a measure of all aspects of arm or hand function, independence, quality of life, or long-term recovery. The registry does not report those outcomes or a follow-up assessment of durability. The brief summary provides a rationale involving brain-to-spinal-cord interactions, but the posted primary outcome does not directly establish a change in those biological processes.

Interpretation is further constrained by the absence of a reported crossover analysis model, paired data, and missing-data methods. Because the design is crossover, the relation between measurements from the same participant matters for estimating uncertainty. The posted comparisons are summary-based and unadjusted, so they do not substitute for a detailed analysis that explains how repeated measurements were handled.

The most defensible reading is limited: the registry reports condition averages and uncertain summary comparisons for the specified task. It does not provide enough statistical detail to make broader claims about treatment effects, clinical importance, or applicability beyond the measured outcome and study population.

16. Why This Trial Matters Statistically

This trial is statistically notable because its crossover design can make within-person comparisons informative, while also requiring analysis that accounts for repeated measurements. Each participant may contribute outcomes under more than one condition, so the precision of a condition contrast depends on how strongly measurements within the same person are related. The registry identifies the crossover structure but does not report a paired analysis or the covariance information needed to understand that aspect of precision.

The posted summaries illustrate why point estimates and uncertainty should be considered together. The two comparison estimates are close to zero, but the reported intervals are broad and include values on both sides of zero. A reader should not treat the point estimates as precise descriptions of the underlying difference, nor infer equivalence from p-values that do not indicate a clear contrast.

The record also shows the distinction between a trial's observed enrollment and a planning calculation. The sample-size illustrations imply large participant counts under the stated assumptions, particularly when planning to detect a smaller fraction of the observed effect. Those figures are not the original trial design justification, which is not reported, and they are not observed power. They demonstrate how sensitive planning requirements can be to the effect size selected.

Finally, the primary measure is a continuous task outcome with no reported clinically meaningful threshold. Statistical differences on a task measure require context before they can be connected to clinical importance. The record's limitations on analysis details, outcome scope, and missing-data reporting are therefore central to careful interpretation rather than minor technicalities.

17. 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 registry record gives limited information about the crossover procedures. It does not report condition sequence, whether order was randomized, washout periods, timing between conditions, or whether period and sequence effects were considered. These details can affect interpretation when outcomes may vary over time or carry over between conditions.

The results include condition means, standard errors, and summary-based unadjusted comparisons, but no individual-level data or paired difference summaries. The record does not describe the method used to calculate the intervals and p-values, and does not identify a model accounting for within-participant dependence. The reported comparisons therefore cannot be verified as a full analysis of the crossover design from the registry record alone.

Participant flow and missing data are not described. Although 23 participants are listed for each condition, the record does not establish whether those are the same participants or whether all completed all conditions. The analysis population and handling of incomplete measurements are not reported.

The endpoint is a force-tracking variability measure. No threshold for a clinically meaningful difference, participant-level distribution, or additional functional outcome is reported. There is also no secondary outcome analysis posted. The registry does not report the original sample-size justification, the planned statistical analysis, multiplicity handling, model diagnostics, or sensitivity analyses.

Finally, the study is listed as unmasked, and the registry does not specify whether assessors or analysts were masked. The safety record lists no serious adverse events or deaths, but does not give further event-collection detail. These limits constrain conclusions to the posted summaries and the outcome described in the record.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Spinal Plasticity to Enhance Motor Retraining After Stroke — NCT03645122.
• Registry record last updated 2024-09-19; status last verified 2024-09.
• Results first posted 2024-09-19.
• 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.

21. Record Summary

The registry record describes NCT03645122, Spinal Plasticity to Enhance Motor Retraining After Stroke, as a completed randomized, unmasked crossover study sponsored by the VA Office of Research and Development. Enrollment is 23, and the phase is listed as not applicable. The reported conditions are rest PCMS, sham PCMS, and active PCMS.

The primary outcome is force-tracking variability in Newtons, measured under each condition. The registry reports condition means and standard errors, along with two unadjusted summary-based comparisons against sham. The reported intervals include zero, and the comparisons are imprecise. The record reports zero serious adverse events and zero deaths in each condition.

No registry analyses are posted. The original sample-size justification, analysis model, crossover-specific methods, missing-data handling, analysis population, and multiplicity strategy are not reported in the registry record. Interpretation is therefore limited to the posted condition summaries and comparisons.