fMRI Studies of Emotional Brain Circuitry in People With Major Depression: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for fMRI Studies of Emotional Brain Circuitry in People With Major Depression (NCT00749125), a non-randomized na study. The primary outcome is Activations in Different Cortical Regions Caused by Emotionally Evocative Task.

2. Clinical Question

The study asks whether activation in selected brain regions during an emotionally evocative task differs between people with major depression and healthy controls, and whether the activation pattern in the depressed group changes after treatment. The registry identifies the amygdala and dorsolateral prefrontal cortex (DLPFC) as the regions of interest. The task included attending to fearful material and ignoring fearful material.

The registry description also states directional expectations: more positive values for depressed participants than controls during the attend-fearful task at baseline, and more negative values for controls than depressed participants during the ignore-fearful task at baseline. It expected these differences to lessen after treatment. These are stated expectations, not reported findings. The registry record does not report whether the observed data matched them.

3. Trial Design

This is a completed, nonrandomized, parallel-group study with no masking listed. The record lists the phase as NA. Its purpose, as described in the registry, was to examine emotional brain circuitry in people with major depression before and after treatment and to identify brain areas involved in chronic depression.

The two groups were not created by random assignment. One arm comprised depressed participants given Lexapro; the other comprised nondepressed controls who did not receive an intervention for depression. The registry describes 23 depressed participants matched with 18 demographically similar healthy controls for the primary outcome analysis. Because the controls were healthy rather than depressed participants assigned to an alternative treatment or no treatment, group differences cannot be interpreted as the effect of Lexapro alone. The groups differ by depression status as well as by intervention exposure.

The primary outcome description refers to fMRI at baseline and week 8. Depressed participants began Lexapro 10 mg/day immediately following the first scan. Controls did not receive an intervention for depression. Thus, the before-and-after assessment applies to the depressed group, while the control group provides a comparison group for the imaging measure.

The registry lists enrollment as 99, while the primary outcome description says that MRI scans from 41 participants were analyzed. The adverse-event record separately lists 51 participants at risk in the Lexapro arm and 48 in the control arm. These counts refer to different descriptions in the record; the record does not explain their relationship or why the imaging analysis included a smaller group. The study is marked completed, with a start date of 2001-07 and primary completion date of 2008-01.

Allocation, masking, analysis populations, and outcome reporting should be interpreted from these registry descriptions. Further details of recruitment, follow-up, and the conduct of the imaging analysis 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.

Before assignment: 99 participants started the study and were assigned to either the depressed or control group.

Overall Study

Analysis population for the primary outcome, as defined in the registry: 23 patients with major depression were matched with 18 demographically similar healthy controls. Depressed subjects were treated with Lexapro 10 mg/day, initiated immediately following the first fMRI scan.

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 Activations in Different Cortical Regions Caused by Emotionally Evocative Task, measured over baseline and week 8.

The primary outcome is described as activation in different cortical regions caused by an emotionally evocative task, measured in voxels using a mean and standard deviation. The scans used BOLD signal during two task conditions: attending to fearful material and ignoring fearful material. The regions selected in advance were the amygdala and DLPFC.

BOLD is an fMRI signal used to assess changes associated with brain activity. In this record, positive values are defined as activation in a region, while negative values are defined as de-activation. The outcome is therefore not a direct count of thoughts, emotions, or clinical symptoms. It is an imaging measure of regional signal during specified task conditions.

Using defined regions and task conditions provides a focused way to examine the study's question about emotional circuitry. The baseline group comparisons address whether activation patterns differ between depressed participants and controls. The later scan addresses whether the depressed group's measured pattern changes after treatment. These are related but distinct comparisons: a difference between groups at baseline is not the same as a change over time within the depressed group.

The registry record does not report the group means, standard deviations, voxel-level estimates, or results of the comparisons. It therefore does not establish the direction or size of any observed activation difference, nor the precision of such a difference.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a voxel-wise comparison mean?

The registry describes voxel-wise comparisons within selected brain regions. In general, a voxel is a small unit in an imaging volume, and a voxel-wise analysis evaluates signal values across such units rather than reducing the entire scan to one whole-brain value. The record names the amygdala and DLPFC as a priori regions, meaning the description identifies them as regions selected for examination in advance. It does not report the voxel-level estimates or the results for those regions.

The primary measure is a continuous outcome summarized using a mean and standard deviation. In general, a mean describes the central value of measurements, while a standard deviation describes their spread around that mean. These summaries can help characterize data, but they do not by themselves establish whether a group difference is reliable or how precisely it is estimated. The record does not give the summaries for the outcome.

Why use ANOVA?

The registry says that voxel-wise comparisons using ANOVAs were performed to determine group differences in activation. In general, analysis of variance (ANOVA) is a framework for comparing means across groups or conditions and can accommodate more than one explanatory factor. Here, the description identifies group, task condition, region, and assessment time as relevant dimensions. It does not specify the exact ANOVA model, its factors and interactions, any covariates, or how repeated observations within a participant were handled. Those details are not reported in the registry record and should not be inferred from the word ANOVA alone.

Because the same depressed participants were scanned at baseline and after treatment, measurements across time may be related within a participant. In general, analyses of repeated measurements need to account for that dependence; treating related measurements as independent can misstate uncertainty. The registry does not specify how this issue was addressed. It also does not describe missing-scan handling, outlier procedures, model diagnostics, or any correction for multiple comparisons across voxels, regions, task conditions, and time points.

What does the BOLD measure represent?

BOLD is the fMRI signal used in the outcome description. Positive values are defined in the registry as activation, and negative values as de-activation. Those terms describe the direction of the measured signal under the study's convention; they are not direct measures of clinical improvement or worsening. The task contrasts attending to fearful material with ignoring fearful material, so the resulting activation values are tied to the experimental task and its specified conditions.

In general, a measurement's interpretation depends on how it was collected, processed, and modeled. The registry record names the emotional distractor task and selected regions but does not report technical acquisition or preprocessing details, the precise contrast construction, or the units beyond voxels. Readers cannot determine from this record alone how those implementation choices affected the estimates.

How should the planning illustrations be read?

General sample-size planning estimates ask how much information a future study might need to detect a selected effect under stated assumptions. The planning approach described for this study uses a two-sided significance level of 0.05, equal allocation, and a normal approximation; the target effect is expressed as a fraction of the observed effect. These are planning assumptions, not evidence that the original study was designed with those assumptions or that it had a particular probability of detecting an effect.

The planning illustration is explicitly based on the observed effect rather than the trial's own design. No observed primary-outcome effect estimate is reported in the registry record, and no numerical planning estimates appear in the registry record for interpretation. Such estimates should not be described as observed power. In general, planning values based on an observed effect are sensitive to uncertainty in that estimate; if the original estimate is imprecise, planning around it can give a misleading sense of certainty.

The record does not report an original sample-size justification, target effect, or planned analysis. These omissions matter because the observed enrollment and the number contributing to the described imaging analysis are not interchangeable measures of the information available for a particular outcome. Without a reported analysis population and effect estimate, the precision and adequacy of the primary analysis cannot be judged from this registry description.

What can be concluded from the method description?

The record establishes that the study intended to compare imaging activation in selected regions across groups and task conditions and to examine the depressed group after treatment. It does not provide the fitted model, numerical estimates, or inferential results. General explanations of ANOVA, repeated measures, and confidence intervals clarify how such analyses are commonly interpreted, but they cannot substitute for the missing study-specific results.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome description involves two selected brain regions, two emotional-task conditions, group comparisons, and baseline and week 8 assessments for the depressed group. This creates multiple possible comparisons. In general, when many hypotheses are tested, the chance of obtaining at least one apparently notable result can increase. Statistical approaches can address multiplicity, for example by adjusting tests or by defining a limited set of primary contrasts in advance. The registry record does not report whether any multiplicity adjustment was used, which contrasts were treated as primary, or how voxel-level testing was controlled.

The record lists one outcome measure and zero analyses, but the outcome description itself contains several regions, conditions, and assessment times. The single outcome listing does not establish how many statistical tests were conducted. Analysis details and numerical results are not reported in the registry record, so multiplicity handling cannot be evaluated.

The record also does not report missing-data counts or methods. Although it states that scans from 41 participants were analyzed, it does not explain how that analysis group relates to the listed enrollment of 99, describe exclusions, or state whether any participants had incomplete scans or incomplete follow-up. It does not identify an intention-to-treat or other analysis population. The adverse-event risk counts are separate registry entries and do not define the imaging analysis population.

In general, missing data can affect both the composition of the analyzed group and the uncertainty of estimates. The direction of any resulting bias depends on why observations are missing and how the analysis handles them. With no flow description or missing-data method reported, the extent and implications of missing imaging data cannot be determined.

11. 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 adverse-event record lists zero serious adverse events, zero other adverse events, and zero deaths for the Lexapro arm, with 51 participants at risk. For the control arm, it likewise lists zero serious adverse events, zero other adverse events, and zero deaths, with 48 participants at risk. These figures describe what the registry record reports for the listed categories and risk counts.

A zero recorded event count does not establish that no symptoms or other health events occurred. It means that no events are entered in these registry categories for the listed groups. The record does not provide event definitions, surveillance procedures, reporting windows, completeness information, or additional detail about how events were collected. It also does not provide a list of serious adverse-event terms.

The safety risk counts do not match the number of participants described as contributing MRI scans to the primary outcome, and the record does not explain the relationship between these figures. They should not be assumed to represent the imaging analysis group. The adverse-event information cannot answer questions about comparative risk beyond the narrowly reported counts, and the nonrandomized groups are not a like-for-like safety comparison.

These figures should therefore be read as registry reporting, not as proof that the intervention caused no adverse effects or that the groups had equivalent safety outcomes.

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

The registry record gives a research question, a study design, an imaging outcome definition, and a description of the planned comparisons. It does not report the primary outcome values or the statistical analyses' numerical results. As a result, it cannot establish whether activation differed between depressed participants and healthy controls, whether activation changed in the depressed group, or how precise any such estimates were.

The study's stated directional expectations should remain separate from its findings. The registry expected more positive baseline values for depressed participants during the attend-fearful task and more negative baseline values for controls during the ignore-fearful task. It also expected those differences to lessen after treatment. The record does not report whether any of these patterns occurred. Expectations in a study description do not demonstrate an observed effect.

Even if numerical group differences were available, this nonrandomized comparison would require careful interpretation. Depressed participants received Lexapro and healthy controls did not receive an intervention for depression. A difference between these groups could reflect depression status, treatment exposure, or other differences between the groups. A before-and-after change in the depressed group would also not, by itself, isolate the effect of treatment from changes over time or repeated measurement.

The imaging measure is BOLD activation in selected brain regions during an emotional task. It is not a direct measure of symptoms, day-to-day functioning, or a clinical outcome. A change in the imaging signal should not automatically be treated as evidence of a corresponding clinical change.

Accordingly, the record supports conclusions about what the investigators intended to measure and how they described the groups, but not about the direction, magnitude, statistical significance, or clinical meaning of the observed imaging results. Those findings are not reported in the registry record.

13. Why This Trial Matters Statistically

This study illustrates the difference between a clearly described research question and a fully interpretable statistical result. The record specifies an imaging measure, selected brain regions, emotional-task conditions, and a before-and-after assessment for the depressed group. Yet it does not include the numerical outcome summaries or analysis results needed to evaluate the findings. A method description alone cannot show the magnitude or precision of an effect.

It also illustrates why the comparison group matters. The controls were healthy participants, not depressed participants assigned to an alternative intervention or no treatment. This design can describe differences between the groups, but it cannot isolate the effect of Lexapro from differences associated with depression status and other group characteristics. Nonrandomized allocation makes that distinction especially important.

The different participant counts in the record raise a separate reporting issue. The enrollment listing, imaging analysis description, and adverse-event risk counts do not describe the same population in the same way, and their relationship is not explained. Transparent participant flow and analysis-population reporting are necessary for readers to understand who contributed to each result and how much information supports it.

Finally, the study focuses on a brain-imaging measure rather than a direct clinical outcome. Statistical interpretation must keep the measurement scale and the clinical question distinct. Even a precisely estimated change in BOLD signal would need careful interpretation before being connected to symptoms or treatment effects. Here, the registry does not report an estimate or its precision, so the appropriate statistical conclusion is limited to identifying the intended comparison and the reporting gaps.

14. 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 interpreting the registry entry is the absence of primary-outcome results. There are no reported means, standard deviations, voxel-level estimates, group contrasts, confidence intervals, or p-values. The record lists zero analyses. As a result, the stated expected directions cannot be evaluated, and no statistical conclusion about the outcome can be drawn.

The nonrandomized design also constrains causal interpretation. The depressed participants received Lexapro, while the controls were nondepressed and received no intervention for depression. The groups therefore do not provide a direct randomized comparison of treatment versus no treatment among people with depression. Matching is described as demographic, but the record does not detail the matching variables or methods, nor report adjustment for other potential group differences.

The record describes MRI scans from 41 participants for the primary outcome, while listing enrollment as 99 and separate adverse-event risk counts. It does not explain eligibility, participant flow, scan availability, exclusions, or the analysis population. This prevents readers from assessing how representative the imaging analysis group was of the enrolled study population.

Important statistical details are also absent. The record does not specify the exact ANOVA model, how repeated observations were handled, which contrasts were prespecified, how multiple voxel-wise tests were addressed, or how missing data were managed. It does not report a sample-size rationale or the original statistical plan. These omissions limit assessment of model suitability and uncertainty.

Finally, the outcome is a task-related fMRI signal in selected brain regions, not a direct clinical endpoint. The record does not provide enough technical or statistical detail to assess the robustness of the imaging measure, and it does not establish how any imaging pattern relates to clinical change.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: fMRI Studies of Emotional Brain Circuitry in People With Major Depression — NCT00749125.
• Registry record last updated 2018-07-17; status last verified 2018-06.
• Results first posted 2018-07-17.
• 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.

18. Record Summary

The completed study examined emotional-task fMRI activation in people with major depression and healthy controls. It was nonrandomized, parallel-group, and unmasked. Depressed participants received Lexapro 10 mg/day after the baseline scan; healthy controls received no intervention for depression.

The primary outcome was mean activation, measured in voxels using BOLD signal, in the amygdala and DLPFC during attend-fearful and ignore-fearful task conditions. The description refers to baseline and week 8 assessments and reports scans from 23 depressed participants and 18 controls. Enrollment is listed as 99.

The registry record does not report primary outcome values or statistical results. It lists zero analyses. Adverse-event entries report zero serious events, zero other events, and zero deaths in each arm, with risk counts of 51 and 48, respectively. The record does not explain how the different participant counts relate.