In a comprehensive study of data from 386 patients with major depressive disorder (MDD), researchers have characterized changes in functional connectivity in the brain during the first 2 weeks after beginning antidepressant treatment.

This has yielded distinct neural signatures that correspond with response to individual SSRI antidepressants, as well as the signature of a placebo effect that appears to be nearly universal in all patients who take antidepressants, and which the research suggests may help reduce depression symptoms in some patients even when they don’t respond to a prescribed medicine.

The study, led by Drs. Xiaoyu Tong and Yu Zhang of the Stanford University School of Medicine, addresses longstanding ambiguity about the mechanism(s) of action of popular antidepressant medications, including the SSRI drugs taken by millions of Americans. SSRIs target the brain’s serotonin neurotransmitter system, but it has been very difficult to explain how these effects translate into changes in patient symptoms—or, in the case of many patients, limited or no changes. At least one patient in three, and possibly as many as half, don’t respond clinically to a first SSRI prescription, or have a response that does not endure. This leads to the time-consuming practice of prescribing antidepressants on a trial-and-error basis, trying different drugs in succession until one proves suitable.

The team reporting the results of the new study, which appeared in Nature Mental Health, included Gregory A. Fonzo, Ph.D., a 2019 BBRF Young Investigator; Charles B. Nemeroff, M.D, Ph.D., a BBRF Scientific Council member, 2003 and 1996 BBRF Distinguished Investigator, and 1997 Selo Prize winner; and Desmond J. Oathes, Ph.D., a 2016 BBRF Young Investigator.

Past research has raised the possibility that antidepressant medicines may not alleviate depression symptoms simply by reversing neural circuit abnormalities seen in patients’ brain scans. It “implies the involvement of intermediary neural circuits that translate treatment-induced brain changes into symptom relief,” the researchers note. They began with the hypothesis that functional connectivity, a network-level measure, might help capture coordinated network reorganization during antidepressant treatment.

It remains unclear if early functional connectivity changes seen in people who respond to antidepressants are in fact induced by the drugs themselves, or are a function of the placebo effect or to factors unrelated to treatment such as biological variability in individual patients.

The team used machine learning and related advanced processing techniques to overcome some limitations in functional connectivity analyses in past studies involving fMRI scanning. This includes “noise” in the results that make it difficult to untangle changes in connectivity that are the direct result of treatment versus those resulting from placebo effects or other sources of variation.

They used large, independent fMRI datasets from two completed clinical trials involving 386 adult patients with major depressive disorder (two-thirds female), each treated with one of two SSRI medicines: sertraline and escitalopram (Zoloft and Lexapro), or placebo. 123 were treated with sertraline, 138 with escitalopram, and 125 with placebo.

With the help of their machine-learning framework, the team was able to systematically characterize changes in functional connectivity attributable to medicine and placebo over the first 2 weeks of treatment. This revealed functional connectivity profiles for what they called a “universal placebo effect,” as well as: connectivity changes due to placebo effect that predicted symptom improvement; and drug-induced connectivity effects in participants who received either of the two SSRI medicines.

Among the most interesting findings was the identification of a subgroup of sertraline-treated patients whose early brain responses to treatment resembled the response of those in the trials who received only placebo. This is indirect evidence of what other data indicated was a “universal placebo effect.” In these data sets, placebo-treated patients showed a highly consistent, characteristic pattern of functional connectivity changes; some experienced moderate symptom improvement, others did not.

Overall, sertraline has proven modestly more effective than placebo in clinical trials, and results in the current study suggest that this may be driven by greater symptom improvement in patients who do demonstrate an early change in functional connectivity directly related to sertraline, compared with those who do not display those patterns despite receiving the same medicine.

Connectivity changes related to placebo may help explain why patients who are less responsive or unresponsive to a prescribed medication may still show some benefit—something that happens frequently in clinical practice. Among sertraline-treated patients, those lacking drug-specific brain changes showed more placebo-like response patterns and, on average, less symptom improvement than those who exhibited sertraline-specific changes.

Importantly, this pattern was also seen in a separate cohort of patients in the current study who were treated with escitalopram, rather than sertraline. To the team, this is evidence that in those who respond to SSRIs, a pattern of “treatment-induced functional connectivity changes are generalizable across the SSRI class of medications and across patient cohorts.”

Prior to these results, it was unclear if non-responders—to active medication or placebo—showed no functional connectivity changes, or, alternatively, did have connectivity changes which “simply did not translate into symptom improvement.” Based on their results, the team said the latter appeared to be true: “Treatment [with SSRIs or placebo] induces widespread functional connectivity changes, even in the absence of symptom improvement.”

The data in this study specifically identified how the placebo effect manifests in terms of functional connectivity: it increases such connectivity in brain regions with “the densest local connections and highest level of concurrent metabolic needs and activity.” This suggests a link between placebo-related changes in functional connectivity and cerebral metabolism (processes in which nutrients are converted to energy to support neural activity and cognition). And this, in turn, is consistent with theories suggesting a relation between antidepressant response and metabolic pathways.

When treatment with placebo resulted in symptom improvement, the team noted decreases in functional connectivity involving the striatum and attention networks. These align with prior findings linking connectivity in the striatum with antidepressant treatment responsiveness. Other data indicated specific effects on functional connectivity associated with sertraline treatment.

Taken together, these and other findings from the study “reinforce the biological plausibility” of the importance of changes in functional connectivity in the first weeks of antidepressant therapy, linking signatures to brain processes involved in metabolism and neurotransmission.

“We found that functional connectivity changes [identified in the study] emerge in nearly all medicated patients, regardless of medication type or clinical response—yet only a subset of these changes contribute to symptom improvement,” the team concluded. “While early functional connectivity changes were largely dominated by the placebo effect, we also identified drug-specific alterations that were associated with treatment outcomes. Patients lacking drug-specific changes responded similarly to placebo-treated patients.”

In practical terms, the findings have implications for precision psychiatry. In the future, pretreatment functional connectivity features may help to identify patients more likely to show a drug-specific benefit, while functional connectivity changes measured soon after treatment begins could potentially help guide adaptive treatment decisions.

The team said follow-up studies should ask whether increasing antidepressant medication dosage—frequently done in cases of initial non-response—can induce additional functional connectivity changes. Further research is also needed to elucidate neural circuits that mediate the translation of early functional connectivity changes into symptom improvement.