Tryptophan MRI in People With Schizophrenia and Healthy Controls: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Tryptophan MRI in People With Schizophrenia and Healthy Controls (NCT02067975), a randomized phase 2, phase 3 study. The primary outcome is Change in Verbal Memory Scores From Baseline to 4 Hours Post-Treatment.

2. Clinical Question

The key statistical question is a within-person comparison: each participant completed both challenge conditions, so the intended contrast concerns the change after tryptophan relative to the change after placebo. The registry also sought to examine differences between healthy controls and people with schizophrenia-related disorders. Those are distinct questions. A difference between the groups’ average scores is not, by itself, evidence that the treatment-related change differed between groups.

The posted numerical results list mean scores and standard deviations for each group-condition combination. They do not give the paired baseline-to-post-treatment changes, the within-person treatment contrast, or a treatment-by-group interaction estimate. Thus, the displayed scores can describe the recorded groups and conditions, but they do not directly answer the primary within-person question. The registry record does not report numerical MRI or brain-chemistry findings among the posted outcome results.

3. Trial Design

The registry describes a randomized, triple-masked, crossover study with phase listed as Phase 2 and Phase 3. Enrollment was 93, and the study status is completed. The listed start date is 2014-08-14, and primary completion is recorded as 2019-12-31. The sponsor is the University of Maryland, Baltimore.

In a crossover design, participants receive both conditions, rather than remaining in a single treatment group throughout. The registry says that participants received tryptophan and placebo in randomized order, with at least two weeks between challenge days. Each challenge visit is described as lasting seven hours, and the intervention was given at the start of the visit. The placebo is described in the study summary as a liquid drink without tryptophan.

The endpoint description says the Hopkins Verbal Learning Test-Revised was administered 90 minutes before treatment and four hours after treatment. Participants performed the task before and after each challenge condition. That repeated measurement is important: the primary question concerns change within a participant, and a crossover comparison ordinarily uses the relationship between that participant’s responses under both conditions. However, the registry’s posted outcome values are means and standard deviations for each group-condition combination, not paired change summaries.

There are differences between the counts displayed for the outcome and those displayed for safety. The outcome groups have 30 or 31 healthy controls and 37 people with schizophrenia-related disorders, depending on condition. The adverse-event groups list 36 healthy controls and 41 people with schizophrenia-related disorders at risk in each challenge condition. The registry does not explain the differences in these counts or provide a participant flow account sufficient to reconcile them. Randomization details beyond randomized treatment order, the masking procedures, and the original design assumptions 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

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 Change in Verbal Memory Scores From Baseline to 4 Hours Post-Treatment, measured over The order in which participants received either the tryptophan or placebo was randomized. The HVLT was administered 90 minutes prior to treatment and 4 hours post treatment. There were at least two weeks between the challenge days..

The primary endpoint is described as change in verbal-memory scores from baseline to four hours after treatment. The assessment was the Hopkins Verbal Learning Test-Revised (HVLT-R), a task of verbal learning and memory. The registry states that raw total scores range from 0 to 36 and are converted to t-scores for group comparisons, with a stated t-score range of -10 to 80. Higher scores represent better performance.

Measuring the task before and after each challenge allows the analysis to focus on change during each condition, rather than only on a participant’s score at one time. Comparing those changes between tryptophan and placebo is the stated primary question. This approach is useful in a crossover study because each participant serves as their own comparison, which can account for stable differences between people. General statistical explanation: paired comparisons can be more informative than treating repeated observations from the same person as independent, but they require information about within-person variation.

The posted outcome data give means and standard deviations for each group and condition. They do not provide the baseline and post-treatment means separately, the change means, or the standard deviation of the paired changes. Consequently, the numerical results do not show the primary change contrast directly. They also do not establish how the listed score summaries relate to the 90-minute pre-treatment and four-hour post-treatment measurements beyond the endpoint description. The registry record does not report additional numerical outcome measures.

7. Results Posted in the Registry

• Schizophrenia Related Disorders: Placebo: mean 40.49 versus 53.90 on control; difference -13.41 (95% CI -18.04 to -8.78), standardized effect d = -1.35; z-test p = <0.001.
• Healthy Controls: Tryptophan: mean 53.03 versus 53.90 on control; difference -0.87 (95% CI -5.28 to 3.54), standardized effect d = -0.10; z-test p = 0.699.
• Schizophrenia Related Disorders: Tryptophan: mean 38.49 versus 53.90 on control; difference -15.41 (95% CI -19.91 to -10.91), standardized effect d = -1.60; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean describe?

The mean is the arithmetic average of the observed scores in a group. Here, the reported outcome is continuous and expressed as a score on a scale. Means offer a compact description of the group’s average score, while the standard deviation describes how dispersed individual scores are around that average. General statistical explanation: a mean does not reveal the full distribution, and two groups with similar means can still differ in their spread or in other features of their score distributions.

The registry reports means and standard deviations for each group-condition combination. For example, the healthy-control placebo group has a mean of 53.9 and a standard deviation of 8.01. These values describe that posted group summary. They are not, by themselves, a baseline-to-follow-up change or a paired treatment effect.

What is a mean difference?

A mean difference expresses the distance between two group averages, using a specified reference. The recalculated contrasts compare each listed group with healthy controls receiving placebo. A negative difference indicates that the first group named in the contrast has a lower mean than the reference under that subtraction convention. The contrast between schizophrenia-related-disorders participants receiving placebo and healthy controls receiving placebo is a group comparison in the placebo condition. The contrast between healthy controls receiving tryptophan and healthy controls receiving placebo is a condition comparison, but its recalculation does not use the paired nature of the crossover data.

The schizophrenia-related-disorders tryptophan contrast compares one group-condition combination against healthy controls receiving placebo. Because both diagnosis group and condition differ, that contrast cannot distinguish their contributions. A treatment-by-group question would require an interaction contrast that compares treatment-related changes across the two groups. No such estimate is reported.

Why does the crossover structure matter?

General statistical explanation: measurements from one person tend to be related. A paired analysis uses that relationship and estimates each participant’s difference between conditions. Its uncertainty depends not only on the variation in each condition but also on how strongly the two measurements are associated within participants. An independent-groups calculation does not use this pairing and can therefore give a different estimate of uncertainty.

The primary endpoint is explicitly a change from baseline to four hours after treatment, with each participant tested in both conditions. Yet the posted summaries do not include individual paired differences, their mean and standard deviation, or the covariance needed to reconstruct their uncertainty. The recalculated independent mean contrasts are consequently not a substitute for the intended paired analysis. The registry record does not report a paired statistical model or a crossover-specific analysis.

How should confidence intervals and p-values be read?

General statistical explanation: a confidence interval describes a range of estimates compatible with the data and the method used, reflecting sampling uncertainty. It is not a range that contains a fixed proportion of individual participants’ scores, and it does not address bias from missing data, measurement choices, or an unsuitable analysis. The recalculated intervals describe unadjusted mean contrasts, not the unreported paired treatment effect.

A p-value measures how unusual data at least as incompatible with a specified null hypothesis would be under the statistical model, assuming that model and its conditions are appropriate. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The reported p-values are unadjusted and relate to the calculated contrasts. The registry does not report whether any multiplicity adjustment was planned or applied.

What does the standard deviation add?

General statistical explanation: standard deviations describe person-to-person spread, whereas standard errors and confidence intervals describe uncertainty in an estimated average or contrast. A larger standard deviation can reflect more heterogeneous individual scores. Standard deviations for the two conditions do not reveal the within-person correlation in a crossover design. That missing relationship matters for estimating the variability of paired changes.

The displayed standard deviations range from 8.01 to 11.27 across the outcome groups. These values describe the reported score distributions, but the registry does not report a standard deviation for baseline-to-post-treatment changes. The available spread summaries therefore cannot fully characterize the precision of the primary comparison.

What analysis details are not available?

The registry reports no listed analysis results and no secondary posted outcomes. The statistical model, handling of missing observations, analysis populations, adjustment for multiple comparisons, and any prespecified covariates are not reported in the registry record. The original sample-size justification and trial-specific power assumptions are also not reported. These omissions limit assessment of whether the posted summaries align with a prespecified analysis and how robust the conclusions would be to alternative methods.

10. Confidence Intervals and Precision

The recalculated contrasts have confidence intervals of different widths, reflecting different degrees of statistical precision under the unadjusted independent-group calculations. For the schizophrenia-related-disorders placebo group versus healthy controls receiving placebo, the estimated mean difference is -13.41, with a 95% confidence interval from -18.04 to -8.784. For healthy controls receiving tryptophan versus healthy controls receiving placebo, the estimate is -0.87, with an interval from -5.282 to 3.542. For schizophrenia-related disorders receiving tryptophan versus healthy controls receiving placebo, the estimate is -15.41, with an interval from -19.91 to -10.91.

General statistical explanation: an interval’s width reflects the precision of an estimate under the assumptions used to calculate it. A wider interval indicates greater uncertainty about the mean contrast, while a narrower one indicates less sampling uncertainty under that method. An interval that includes zero is compatible with both negative and positive differences at the confidence level used; it does not prove that there is no difference.

The precision in these intervals must be interpreted in relation to what they estimate. They are unadjusted comparisons based on posted group summaries, not confidence intervals for the primary paired change contrast. Since the registry does not report paired changes or within-person covariance, the precision of the primary treatment comparison cannot be assessed from these intervals. The intervals also do not capture possible uncertainty from missing data, analysis choices, or multiplicity.

The registry reports outcome group sizes of 30, 31, and 37, depending on the listed group-condition combination. Those counts are not a complete account of paired observations: the registry does not state how many participants contributed complete data under both conditions or how many had both pre- and post-treatment assessments. Therefore, the displayed sample sizes should not be interpreted as a clear count of complete paired records.

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 planning figures are participants per arm for each stated power target; they are not total event counts. The calculations use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each planning scenario targets the stated fraction of the observed effect. The original trial’s design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the planning number of participants per arm for a design that would have an 80% chance of detecting the specified effect under those assumptions, if that effect were the true one. The 90% power column gives the corresponding planning number for a 90% chance. These are prospective planning illustrations based on observed contrasts, not the trial’s own design calculation.

The scenarios include planning for the full observed effect, planning for 75% of the observed effect, and planning for 50% of the observed effect. Smaller effects are harder to distinguish from random variation, so they require more participants under the same assumptions. This pattern is especially apparent for the small healthy-control tryptophan-versus-placebo contrast: its planning figures are much larger than those based on the larger between-group contrasts. The planning figures should not be used to infer the trial’s original target enrollment or whether it was adequately designed for its primary paired analysis.

This is not observed power. The displayed figures describe hypothetical future planning based on the stated assumptions and selected effect sizes. They do not measure the probability that the completed trial detected an effect, and they do not resolve the missing paired-change information.

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 = -13.410; pooled standard deviation σ = 9.948.
2. n = 2 × (1.96 + 0.842)² × 9.948² / -13.410² = 8.6, rounded up to 9 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 <- -13.41                      # observed difference in means
sigma <- 9.94841100570646                     # 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(2067975)
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 = 9 (80% power), 12 (90% power)
 75% of observed effect: n per arm = 16 (80% power), 21 (90% power)
 50% of observed effect: n per arm = 35 (80% power), 47 (90% power)
power.t.test: 9.7 
Simulated power at n = 9 per arm: 0.766 (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 posted outcome measure and zero analyses, with no secondary posted outcomes. It does not report a prespecified statistical analysis plan, the number of comparisons intended for testing, or whether any multiplicity adjustment was planned or applied. Multiple testing can arise when investigators examine several outcomes, groups, time points, or contrasts. General statistical explanation: when many hypotheses are tested, the chance of at least one small p-value can increase unless the analysis accounts for the testing structure. The available record does not establish how that issue was handled.

Missing-data methods are not reported in the registry record. The posted outcome group sizes differ by condition, and the registry does not explain whether missing observations, withdrawals, or incomplete visits account for those differences. It does not report an intention-to-treat population, a per-protocol population, or rules for including participants with incomplete measurements. Without those details, it is not possible to determine how the posted summaries relate to all enrolled participants or whether missingness could affect the comparisons.

The record also does not report a statistical model for the repeated measurements or the crossover structure. In particular, it does not provide a paired treatment estimate, a treatment-by-group interaction, or information on how treatment order was incorporated in analysis. The randomized order is described, but any modeling of order, period, or carryover is not reported in the registry record. No specific carryover effect should be inferred from the two-week interval.

The registry reports no sponsor analysis results to compare with the recalculated unadjusted contrasts. Those contrasts are not a replacement for a prespecified analysis and cannot establish how the trial investigators handled multiplicity, missingness, or the repeated-measures design. These limits chiefly affect interpretation of the primary question, which calls for a within-participant comparison of change under tryptophan and placebo.

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.

Most frequent non-serious adverse event terms

The registry lists serious adverse events separately for healthy controls and participants with schizophrenia-related disorders, and separately by tryptophan, placebo, and screening. It reports zero serious adverse events in each listed group. The numbers at risk were 36 for each healthy-control challenge group and 41 for each schizophrenia-related-disorders challenge group. Screening groups list 8 at risk in each population.

These are descriptive registry counts. The record does not provide event timing, participant-level detail, or an account of how safety populations were defined. The safety denominators differ from the outcome counts, and the reason is not reported in the registry record. Because the study is a crossover, the same individuals may contribute to more than one condition’s safety count; the listed groups should not automatically be treated as independent populations.

The registry lists no serious adverse-event terms. It does report other adverse events: 13 in healthy controls during tryptophan, 6 in the schizophrenia-related-disorders group during tryptophan, 3 in that group during placebo, and 1 during healthy-control screening. No other adverse events are listed in the healthy-control placebo group or schizophrenia-related-disorders screening group. These counts do not provide details about event severity, attribution, or clinical significance. The absence of reported serious events in these groups is a statement about the posted counts, not a broader safety conclusion.

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

The registry describes a randomized crossover challenge in which healthy controls and people with schizophrenia-related disorders received tryptophan and placebo in randomized order. Its stated primary aim was to compare verbal-memory change from before treatment to four hours afterward between the two conditions. The assessment used the HVLT-R, and higher t-scores indicate better performance.

The posted score summaries show different average scores across the diagnostic groups and conditions. The recalculated unadjusted comparison of schizophrenia-related-disorders participants receiving placebo with healthy controls receiving placebo has a negative mean difference, with a confidence interval that excludes zero. The healthy-control tryptophan-versus-placebo contrast has a confidence interval spanning negative and positive values. The comparison of schizophrenia-related-disorders participants receiving tryptophan with healthy controls receiving placebo combines differences in both group and condition.

Those are not equivalent to the primary treatment question. The record does not report the change in score for each condition or the within-person contrast between changes. Nor does it give the paired variability required to estimate that contrast’s uncertainty. Therefore, the posted summaries do not establish whether change differed between tryptophan and placebo, or whether such a difference varied between diagnostic groups.

The registry summary describes the study’s rationale for manipulating kynurenic acid and using cognitive tests and MRI to study related measures. But numerical results for MRI, brain activity, or brain chemistry are not reported among the posted outcomes. The record also does not provide a clinical interpretation or a detailed account of missing data and analysis methods. The most defensible reading is limited to the reported score summaries, the specified unadjusted recalculations, and the gaps in the direct evidence for the primary paired comparison.

16. Why This Trial Matters Statistically

This trial is statistically notable because its stated primary question and its posted numerical summaries do not have the same form. The endpoint is a within-person comparison of change under two conditions. The posted values, however, are condition-specific group means and standard deviations, followed by unadjusted contrasts against healthy controls receiving placebo. Recognizing that difference prevents a between-group score contrast from being mistaken for a treatment effect.

The crossover design makes pairing central to the analysis. Each participant experienced both conditions, and the registry says their order was randomized. A paired comparison can use each participant’s responses under both conditions, but its uncertainty depends on how those responses are related. Since the registry does not report paired changes or within-person covariance, the primary contrast’s estimate and precision cannot be reconstructed from the posted summaries.

The record also illustrates why reference groups matter. A comparison that changes both diagnostic group and treatment condition cannot isolate either factor. The posted contrast for the schizophrenia-related-disorders tryptophan group versus healthy controls receiving placebo does both. Separate treatment comparisons within each diagnostic group, and a comparison of treatment effects across groups, would answer different questions; those paired estimates are not reported.

Finally, the registry reports numerical safety counts and outcome summaries but does not supply analysis details, missing-data methods, multiplicity handling, or the original sample-size rationale. Those omissions do not erase the information that is posted, but they constrain how far statistical interpretation can go. The central lesson is to match each estimate to its actual comparison and avoid treating descriptive group means as a direct test of a repeated-measures primary endpoint.

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 most important limitation for interpretation is that the primary outcome is framed as change from baseline to four hours after each challenge, whereas the registry posts group-condition means and standard deviations without the paired change estimates. No treatment-versus-placebo comparison of within-person changes is reported. The within-person correlation, needed to calculate uncertainty for a paired contrast, is also not reported.

The three recalculated contrasts do not each answer a clean treatment question. Two compare schizophrenia-related-disorders participants with healthy controls under a particular condition; the third compares schizophrenia-related-disorders participants receiving tryptophan with healthy controls receiving placebo, changing both group and condition. The registry does not report a treatment-by-group interaction estimate.

The outcome counts differ across conditions, and the adverse-event denominators are larger than the outcome group counts. The reasons for these differences, completeness of paired measurements, and handling of missing observations are not reported in the registry record. The statistical model, analysis population, any adjustment for treatment order or period, multiplicity strategy, and original sample-size justification are also not reported.

The record lists MRI and brain-chemistry procedures in the study rationale, but does not post numerical results for those measures. The safety record gives aggregate counts without event terms for serious adverse events or detailed context for other events. These limits mean the posted results support description of the recorded summaries, but not a complete assessment of the primary crossover effect or the broader study aims.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Tryptophan MRI in People With Schizophrenia and Healthy Controls — NCT02067975.
• Registry record last updated 2024-07-30; status last verified 2024-07.
• Results first posted 2021-08-26.
• 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

NCT02067975 is a completed randomized, triple-masked crossover study sponsored by the University of Maryland, Baltimore. It enrolled 93 participants and lists Phase 2 and Phase 3. Healthy controls and people with schizophrenia-related disorders received tryptophan and placebo in randomized order, with at least two weeks between challenge days.

The primary endpoint is verbal-memory score change from baseline to four hours after treatment, assessed with the HVLT-R. The registry posts means and standard deviations by group and condition, but not paired change estimates or a direct treatment-versus-placebo analysis of change. The unadjusted recalculated contrasts therefore do not directly answer the stated primary question.

The registry reports zero serious adverse events in the listed groups and does not report numerical MRI or brain-chemistry outcomes. Analysis details, missing-data methods, multiplicity handling, and the original sample-size justification are not reported in the registry record.