Cortical Connectivity Differentiates Antidepressant From Placebo Response

By Will Boggs MD

January 08, 2020

NEW YORK (Reuters Health) - Cortical connectivity can be used to predict placebo and antidepressant outcomes, according to a secondary analysis of a randomized clinical trial.

"Our findings indicate that, while as an average across the population, antidepressants seem only slightly better than placebo, this is in fact due to major neurobiological differences between patients rather than the drugs not being effective," Dr. Amit Etkin from Stanford University in California told Reuters Health by email. "In other words, antidepressants are effective, but only for a subpopulation of depressed patients, who our research suggests can be identified using sophisticated EEG analyses."

Several EEG studies have shown that regional differences in alpha and theta power and global signal entropy are associated with better antidepressant treatment responses, but such studies have lacked statistical power.

Dr. Etkin and colleagues used data from the EMBARC randomized trial of sertraline versus placebo to investigate how individual differences in EEG power envelope connectivity (PEC)-estimated cortical networks differentially predicted outcomes in 221 patients with major depressive disorder (MDD).

Analyses revealed moderation of treatment effects by connections within and between widespread cortical regions for both the antidepressant and placebo groups, according to the online report in JAMA Psychiatry.

Moderating connections within the resting with eyes closed (REC) condition from the alpha and gamma carrier frequencies significantly predicted outcomes primarily in the placebo group, whereas only a few connections predicted outcomes in the sertraline group.

In contrast, in the resting with eyes open (REO) condition, connectivity moderators reflected effects in both the placebo and sertraline groups.

In a separate visualization of the REC connectivity results, greater alpha-band PEC within parietal, temporal, and visual regions predicted better treatment outcomes with placebo and worse treatment outcomes with sertraline. Reduced gamma-band PEC within the frontal, visual, somatomotor, parietal, and temporal regions also predicted better treatment outcomes with placebo and worse treatment outcomes with sertraline.

In a similar analysis of REO connectivity, greater alpha-band PEC within the visual and parietal regions; reduced gamma-band PEC within the frontal, temporal, parietal, anterior cingulate, and visual and somatosensory regions; and reduced beta-band PEC within the frontal and parietal regions predicted better outcomes in the placebo group and worse outcomes in the sertraline group.

Differences in connectivity were also associated with differences in depressive symptom severity.

"What surprised us most was the central importance of connections involving the parietal cortex," Dr. Etkin said. "This region, located towards the back of the brain, is involved in a range of functions (such as cognition) but is not typically a central point in how we describe the brain circuitry of depression. Nonetheless, because we used a very rigorous statistical approach, the strength of our findings leads to key new questions about why it is that the parietal cortex plays such an important role."

"A smaller role was identified for the prefrontal cortex, a region more closely associated with depression, and it is unknown whether other ways to quantify EEG connectivity may reveal different brain circuitry by virtue of examining different aspects of physiology," he said.

"Depression and response to antidepressants should no longer be thought of as an imprecise diagnosis involving a process of trial-and-error prescriptions," Dr. Etkin said. "Rather, our findings indicate that more precise diagnosis, in a treatment-relevant way, is possible through objective measurements such as EEG, and such tools may be less far away from clinical use than we think given the ease and cost-efficiency of EEG."

Dr. Alik Widge from University of Minnesota, Minneapolis, who co-authored an editorial related to this report, told Reuters Health by email, "The most surprising thing, to me, was how much the classifiers depended on visual and other sensory areas and how little they relied on areas we classically think about in depression, such as the prefrontal cortex. We don't know why that would be true, and if it holds up in other datasets, it would suggest a need to really broaden our thinking about depression."

"Clinically, if validated, this would bring us much closer to an era of precision psychiatry," he said. "If a reliable diagnostic test could be developed that could predict response to multiple different medications, we could reduce the amount of trial-and-error in psychiatric practice."

"Right now, these results aren't ready for clinical practice," Dr. Widge concluded. "They are really exciting in terms of neuroscience questions they open up, especially because this is the most rigorous study of its kind that's ever been reported. What they tell us is that EEG biomarkers remain a hot area to watch."

Dr. Ramiro Salas from Baylor College of Medicine, Houston, Texas, who has researched brain connectivity in MDD and bipolar disorder, told Reuters Health in an email, "A major part of this is that EEG is very inexpensive compared to MRI. If this work grows into something clinically actionable, it has the potential to reach many more people as physicians could be able to decrease the 'hit and miss' testing for antidepressants for each patient."

"The manuscript is really interesting, novel, and relevant," he concluded. "However, we must remember that the path from finding results that are statistically significant and thus likely reproducible to patients benefiting from those findings is very long and uncertain."

SOURCE: and JAMA Psychiatry, online January 2, 2020.