Cognitive Training and Neuroplasticity in Mild Cognitive Impairment: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Cognitive Training and Neuroplasticity in Mild Cognitive Impairment (NCT03205709), a randomized na study. The primary outcome is Change in the Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog).

2. Clinical Question

3. Trial Design

The registry record describes a randomized, double-masked, parallel-group trial. Its listed enrollment is 107, and its status is COMPLETED. The sponsor is the New York State Psychiatric Institute. The record lists the phase as NA. The two groups were assigned to Computerized Cognitive Training or crossword puzzles. Because both groups received an activity, the comparison estimates the difference between these study conditions rather than the difference between an activity and no intervention.

The brief description frames the study around cognitive performance, everyday cognitive and functional status, and longer-term changes in brain networks. It specifies an 18-month period for evaluating those effects. The primary endpoint, however, is described as change in modified ADAS-Cog score over 78 weeks. The registry record does not explain the relationship between that timeframe and the broader 18-month description.

The primary outcome population is described as participants assessed at baseline and week 78. The posted group entries for this endpoint list 44 participants in Training Group 1 and 51 in Training Group 2. Other outcome entries and safety entries use different participant counts. The record does not explain these differences, so they should not be interpreted as evidence about why participants were or were not included in a particular analysis.

The registry record does not report the original sample-size rationale, details of the masking procedures, the randomization method, or the trial's full statistical analysis plan. It also does not report the planned handling of missing outcome data. Those details are important to understanding how the design was intended to support inference, but should not be inferred from the posted summary statistics.

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

Analysis population for the primary outcome, as defined in the registry: Participants assessed at baseline and week 78, and the change over 78 weeks is the outcome.

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: 107 participants with MCI

6. Primary Endpoint

The registry lists the primary outcome as Change in the Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog), measured over [Time Frame: 78 weeks].

The primary endpoint is change in the modified Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog) over 78 weeks. The registry describes the measure as a cognitive battery assessing learning, memory, language production and comprehension, constructional praxis, ideational praxis, and orientation. Its value range is 0-70, and a higher score indicates worse cognition. The endpoint therefore uses change in a cognitive score to compare outcomes over the trial period.

The record identifies the endpoint measure as a least-squares mean change, with standard error as the dispersion measure. Least-squares means are model-based group means; the registry record does not describe the model used to produce them. The reported population includes participants assessed at baseline and week 78, with change over 78 weeks as the outcome.

In general, a change endpoint can describe how a measured characteristic shifts over a stated interval, while comparing change between groups can account for the direction of movement on the scale. Here, a positive change corresponds to a higher score, which the record identifies as worse cognition; a negative change corresponds to a lower score. The endpoint is a measured cognitive outcome, not by itself a complete account of everyday functioning or all aspects of the condition.

7. Results Posted in the Registry

• Training Group 2: mean 0.55 versus -0.89 on control; difference 1.44 (95% CI 0.07 to 2.81), standardized effect d = 0.42; z-test p = 0.040.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change Over Time in Neuropsychological Testing Composite Score

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Training Group 2: mean -0.27 versus -0.24 on control; difference -0.03 (95% CI -0.24 to 0.18), standardized effect d = -0.05; z-test p = 0.778.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Training Group 2 and the control. The interval gives the range of effects that remain compatible with the data.

Change Over Time in Functional Activities Questionnaire (FAQ)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Training Group 2: mean -0.92 versus -2.00 on control; difference 1.08 (95% CI 0.18 to 1.98), standardized effect d = 0.46; z-test p = 0.019.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Training Group 2 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change Over Time in UCSD Performance-Based Skills Assessment (UPSA)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Training Group 2: mean -0.99 versus 0.55 on control; difference -1.54 (95% CI -5.25 to 2.17), standardized effect d = -0.16; z-test p = 0.416.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Training Group 2 and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover three related but distinct domains: a neuropsychological testing composite, functional activities measured with the FAQ, and performance-based skills assessed with the UPSA. Together they broaden the outcome picture beyond the primary ADAS-Cog cognitive battery. Their results do not all point to the same statistical conclusion: the unadjusted FAQ comparison has an interval excluding zero, while the intervals for the neuropsychological composite and UPSA include zero.

This pattern offers limited alignment with the primary result rather than a uniform confirmation across measures. The FAQ result is in a functional domain, while the primary endpoint assesses cognitive performance. The other secondary comparisons do not provide similarly precise evidence of a nonzero difference on their respective measures. These are separate outcomes, and their scores should not be treated as interchangeable or combined into a single conclusion without an analysis plan that supports doing so.

The unadjusted secondary results carry less weight because the registry record does not report a multiplicity adjustment or a complete statistical analysis plan. In general, testing several outcomes without accounting for the number of tests can increase the chance that at least one result appears notable even when the underlying differences are limited. The registry record lists three secondary outcomes and does not provide reported analysis entries for them. That absence does not mean no analyses were conducted; it means their analysis details are not reported in the registry record.

The secondary outcomes are most useful as complementary measures that help describe the breadth and consistency of the findings. They do not independently settle the primary question, and a single secondary result should not outweigh the overall pattern, the uncertainty in each estimate, and the lack of reported multiplicity and missing-data methods.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change Over Time in Neuropsychological Testing Composite Score
• Change Over Time in Functional Activities Questionnaire (FAQ)
• Change Over Time in UCSD Performance-Based Skills Assessment (UPSA)

9. Statistical Interpretation

10. Statistical Methods Explained

For a continuous outcome, a group mean describes the average score or average change in that group. The primary outcome entries are least-squares means, which are generally model-based estimates rather than necessarily raw arithmetic averages. The registry record does not specify the model, its covariates, or other adjustment details. The group values are changes in ADAS-Cog score, and the comparison is the mean difference for Training Group 2 relative to Training Group 1.

What does the mean difference represent?

In general, a mean difference expresses the distance between two group means on the outcome's original scale. Here, the posted difference is 1.44 score points, with the comparison oriented as Training Group 2 relative to Training Group 1. Because higher ADAS-Cog scores indicate worse cognition, the sign must be interpreted in light of that scale direction. A positive difference means the reported change in Training Group 2 is higher than the reported change in Training Group 1. This is an interpretation of the score contrast, not a claim about the clinical importance of that contrast.

What does the standard error describe?

In general, a standard error describes the estimated sampling uncertainty of a statistic, such as a group mean. It is not the same as the spread of individual participants' scores. The primary group entries report standard errors of 0.51 and 0.48. Smaller standard errors, all else equal, indicate more precise estimates; they do not show how varied individual outcomes were. The registry record does not separately report the standard deviation of observed ADAS-Cog changes.

How should the confidence interval be read?

In general, a confidence interval gives a range of values compatible with the estimate under the statistical procedure used. The primary mean-difference interval excludes zero, so the interval is inconsistent with no mean difference under that procedure at the corresponding confidence level. Its width is also relevant: estimates across the interval represent materially different possible magnitudes. The interval is not a guarantee that the true value lies within its particular bounds, nor does it describe the range of individual participant outcomes.

What does the p-value add?

In general, a p-value describes how unusual a result at least as incompatible with a null hypothesis would be if that hypothesis and the analysis assumptions held. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of an effect. The posted primary unadjusted p-value is 0.04. The word unadjusted matters because the record does not state that this value accounts for testing multiple outcomes or comparisons.

Why does over-dispersion matter?

In general, over-dispersion arises when outcome variability is greater than a model assumes. A model that understates variability can produce standard errors that are too small and intervals that are too narrow. This concept is especially relevant in some count and binary-data models. The primary endpoint is continuous, and the registry record does not report an over-dispersion assessment or a model specification. It would therefore be inappropriate to claim that over-dispersion was present or handled in the primary analysis.

How are the posted planning calculations different from trial results?

In general, sample-size calculations estimate how many participants a future study would need under stated assumptions to have a chosen chance of detecting a specified effect at a specified significance level. The posted planning illustrations use the observed effect as their starting point and consider stated fractions of that effect. They do not describe the trial's original design justification, which is not reported in the registry record. The calculated standardized pooled spread is approximate, based on the posted summary data. Planning figures should be read as illustrations conditional on their assumptions, not as evidence about the statistical power achieved by this completed trial.

The registry record also lists secondary outcome estimates and unadjusted comparisons. In general, examining more outcomes increases the chance of obtaining at least one apparently notable result by chance if no adjustment is made. The record has no reported analysis entries for the secondary outcomes and does not identify a multiplicity strategy. These limits affect how much weight can be placed on isolated secondary findings.

11. Confidence Intervals and Precision

The primary mean difference is 1.44, and its reported 95% confidence interval excludes zero. That interval gives a range for the estimated difference under the analysis procedure, while also showing that the exact magnitude remains uncertain. The registry record does not report a clinical threshold for interpreting the score difference, so statistical precision cannot by itself establish whether the estimated magnitude is important in practice.

The standard errors for the two primary group estimates are 0.51 and 0.48. Standard errors describe uncertainty in the estimated means, not the spread of individual scores. The record does not give a standard deviation for the observed primary changes or a detailed model description, so the sources of uncertainty and the model's assumptions cannot be fully evaluated from these summaries alone.

In general, an interval that excludes zero is evidence of a nonzero contrast under the specified analysis, but it does not make the estimate exact. The width of the interval reflects the remaining uncertainty about the size of the contrast. An interval that includes zero, as occurs for some secondary outcomes, is compatible with both a difference and no difference under the corresponding procedure; it does not prove that the groups are identical.

The confidence interval and p-value are reported without an indicated multiplicity adjustment. If several outcomes are considered together, unadjusted intervals and tests do not account for the increased opportunity for chance findings. In addition, the registry record does not report missing-data methods or analysis-population definitions. These omissions limit a full appraisal of precision and of whether the reported uncertainty captures all relevant sources of uncertainty.

12. 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, not total participants and not total events. The registry record states that the illustrations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is a stated fraction of the observed effect. The original trial's own sample-size assumptions and justification are not reported in the registry record.

For the primary comparison, planning for 100% of the observed effect gives 88 participants per arm for 80% power and 118 participants per arm for 90% power. Planning for 75% of the observed effect gives 157 participants per arm for 80% power and 210 participants per arm for 90% power. Planning for 50% of the observed effect gives 352 participants per arm for 80% power and 471 participants per arm for 90% power. The plotted curve presents this same relationship across the planning scenarios.

In general, power is the probability that a study will detect a specified effect under the calculation's assumptions if that effect is present. The 80% and 90% columns therefore describe different target probabilities for a hypothetical design: the higher target generally requires a larger planned sample. Smaller effects are harder to distinguish from random variation, so the planning illustrations require more participants per arm as the effect to detect becomes smaller.

These figures are planning illustrations based on the observed effect, not observed power for the completed trial. They should not be read as proof that the trial was adequately or inadequately sized, since its original design assumptions are not reported and observed effects can be uncertain.

13. 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 = 1.440; pooled standard deviation σ = 3.407.
2. n = 2 × (1.96 + 0.842)² × 3.407² / 1.440² = 87.9, rounded up to 88 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 <- 1.44                      # observed difference in means
sigma <- 3.40718595457193                     # 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(3205709)
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 = 88 (80% power), 118 (90% power)
 75% of observed effect: n per arm = 157 (80% power), 210 (90% power)
 50% of observed effect: n per arm = 352 (80% power), 471 (90% power)
power.t.test: 88.9 
Simulated power at n = 88 per arm: 0.792 (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.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome and three secondary outcomes, but does not report a statistical plan for controlling multiplicity. It also does not provide reported analysis entries for the secondary outcomes. Without a stated testing hierarchy, correction method, or distinction between confirmatory and exploratory outcomes, the unadjusted p-values should not be interpreted as if they collectively maintained a prespecified error rate.

In general, multiplicity arises when several hypotheses are tested. The more tests considered, the greater the opportunity for at least one result to appear statistically notable by chance. Adjustment procedures can address this issue, though they may change the interpretation and precision of individual tests. The registry record does not say whether any such procedure was used. The primary result is therefore best described using its posted unadjusted p-value, with the limitation that any relationship to other tested outcomes is not documented.

The record describes the primary endpoint population as participants assessed at baseline and week 78. Group counts for that endpoint are 44 and 51, while counts in the secondary and safety entries are 51 and 56. The record does not explain these differences, and they should not be attributed to a particular cause. They do not, by themselves, establish how many participants lacked data or why the analysis counts differ.

The registry record does not report whether the analyses followed an intention-to-treat population, a per-protocol population, or another defined analysis set. It does not specify how missing outcome data were handled, whether imputation was used, or whether sensitivity analyses assessed the effect of missingness. These are important details because the results may depend on who contributed outcome measurements and how unavailable values were addressed. Without this information, the risk of bias from missing data and the generalizability of the estimates cannot be fully assessed.

15. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: vascular disorders (5.9%), respiratory disorders (5.9%), falls (3.9%), musculoskeletal and connective tissue disorder (3.9%), cardiac disorders (3.6%).

The registry safety entries list serious adverse events as affecting 12 participants in Training Group 1, with 51 at risk, and 9 participants in Training Group 2, with 56 at risk. Other adverse events are listed for 19 participants in Training Group 1 and 23 in Training Group 2, using the same group-specific at-risk counts. The record lists 0 deaths in each group. These are counts as recorded, not a causal comparison between activities.

The listed serious-event terms include vascular disorders and respiratory disorders, each reported as 5.9%; falls and musculoskeletal and connective tissue disorder, each as 3.9%; and cardiac disorders and urinary disorder, each as 3.6%. The registry record does not clarify in this summary how these term percentages relate to participants or events, or provide group-specific breakdowns for each term. They should therefore be read as descriptive entries, not as evidence that an activity caused a particular event.

In general, safety comparisons require context about event definitions, ascertainment, follow-up, and the number of participants at risk. The registry record does not provide enough detail to evaluate those elements or to establish a comparative safety profile.

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

The primary result describes a difference in mean change on a cognitive score over 78 weeks between the computerized training and crossword puzzle groups. The reported estimate is 1.44 points for Training Group 2 relative to Training Group 1, and its confidence interval excludes zero. The scale direction matters: higher ADAS-Cog scores indicate worse cognition. The result is a comparison of changes in the recorded outcome, not a complete measure of cognition, daily life, or a participant's overall experience.

The estimate should not be translated into a claim that one activity works or fails. The registry record does not report a threshold for a clinically meaningful difference, and statistical evidence of a nonzero difference does not alone establish practical importance. The posted secondary results also do not form a uniform pattern across cognitive, functional, and performance-based outcomes.

The comparison is between two active study activities. It cannot, on its own, establish how either activity compares with no cognitive activity or with usual care. The brief description also discusses long-term brain-network changes, but the posted outcome details do not report a brain-network result. No conclusion about that objective can be drawn from the cognitive-score comparison.

The registry record does not report the full analysis plan, missing-data approach, or multiplicity procedure. Differences in group counts across endpoint entries are not explained. These gaps limit assessment of how robust the estimate is to analysis choices and incomplete outcome data. The posted result is best understood as a quantitative comparison with uncertainty, within the specific population, measures, time frame, and analysis details that the registry record describes.

17. Why This Trial Matters Statistically

This trial illustrates why a clinical result is more than a p-value. The primary estimate has a direction and magnitude on a defined scale, an interval that conveys uncertainty, and an unadjusted p-value. Reading these elements together prevents a threshold-based interpretation from replacing attention to the size and precision of the estimated difference.

The study also compares two active cognitive activities. That design makes the contrast specific: it concerns the difference between computerized cognitive training and crossword puzzles, rather than the effect of either activity relative to no activity. The secondary measures extend the assessment to functional activities and performance-based skills, but their mixed pattern does not provide a single consistent signal across every domain.

At the same time, the registry record illustrates the limits of public summary results. It does not report the original sample-size justification, detailed analysis plan, multiplicity strategy, missing-data methods, or analysis-population definitions. Group counts also vary among outcome entries without an explanation. These omissions matter because they shape how readers assess precision, bias, and the interpretation of multiple outcomes.

Finally, planning calculations based on an observed effect are not evidence of the power achieved in a completed study. They are conditional illustrations useful for understanding how a future design might change with the target effect and power level. Keeping design planning, observed estimates, uncertainty, and clinical meaning distinct is central to responsible interpretation.

18. 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.

Several limitations specific to interpretation of this trial arise from the registry record. It does not provide the detailed randomization procedure or explain how double masking was implemented for activities that differ in format. This does not establish that masking failed; it means the implementation cannot be evaluated from the recorded details.

The primary outcome has a defined 78-week assessment point, while the brief description refers to an 18-month period. The record does not explain the relationship between these timeframes. The primary population is described as participants assessed at baseline and week 78, but the record does not state how participants without both assessments were handled or whether the analysis included all randomized participants.

The endpoint counts differ from the counts listed for secondary outcomes and safety. The reasons are not reported, so it is not possible to determine from these entries whether the differences reflect endpoint availability, analysis criteria, or another process. The registry record also omits the full statistical model, covariate adjustments, missing-data methods, sensitivity analyses, and multiplicity plan.

The primary result is a summary change on a cognitive measure. The posted details do not provide a clinical-importance threshold, participant-level outcome distributions, or a reported brain-network result despite that objective appearing in the study description. Safety entries are aggregate and do not provide event-level timing or group-specific breakdowns by serious-event term. These limitations restrict the depth of inference but do not, by themselves, establish that the reported results are invalid.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Cognitive Training and Neuroplasticity in Mild Cognitive Impairment — NCT03205709.
• Registry record last updated 2024-09-19; status last verified 2024-04.
• 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.

22. Record Summary

The registry record describes a completed, randomized, double-masked, parallel-group trial sponsored by the New York State Psychiatric Institute. It lists enrollment of 107 and compares Computerized Cognitive Training with crossword puzzles in participants with mild cognitive impairment. The primary outcome is change in modified ADAS-Cog score over 78 weeks among participants assessed at baseline and week 78.

The reported primary mean difference is 1.44 for Training Group 2 relative to Training Group 1, with a confidence interval excluding zero and an unadjusted p-value of 0.04. The secondary outcome pattern is mixed. The record does not report the original sample-size rationale, detailed analysis plan, multiplicity handling, missing-data methods, or reasons for differing group counts across outcomes. The posted findings support a measured comparison with uncertainty, not a broad conclusion about either activity.