![[Adobe Stock]](https://sp-ao.shortpixel.ai/client/to_webp,q_glossy,ret_img,w_750,h_375/https://www.psypost.org/wp-content/uploads/2025/05/neurons-3-750x375.jpg)
A recent study published in Nature Mental Health suggests that using biological and behavioral markers can help predict how well a person with depression will respond to common medications. By evaluating specific brain patterns and cognitive traits before treatment, scientists found evidence that certain patients are much more likely to experience symptom relief from antidepressants. These findings provide a foundation for developing personalized treatment plans that could reduce the time and guesswork typically involved in finding the right medication for depression.
Major depressive disorder affects millions of people, yet finding an effective treatment often relies on a trial and error process. A patient might take a medication for a month or more before knowing if it works, and a large portion of individuals do not experience adequate relief from the first drug they try.
“Treating depression still involves too much trial and error: only about 30-50% of patients respond to the first antidepressant they receive, and it can take several weeks to determine whether a medication is working,” said Diego A. Pizzagalli, founding director of UC Irvine’s Noel Drury, M.D. Institute for Translational Depression Discoveries and distinguished professor of psychiatry and human behavior, neurobiology and behavior, and biomedical engineering.
To reduce this waiting period and improve outcomes, scientists are searching for objective signs that might predict a person’s response to specific drugs before a prescription is written. These objective signs are known as biomarkers, which are measurable indicators of a biological state or condition. For depression, a biomarker might be a specific pattern of brain activity, a cognitive trait, or even a demographic factor like a person’s employment status. By identifying these markers, clinicians might eventually match patients with the exact medication most likely to help them.
In this context, the study aimed to predict responses to two widely prescribed antidepressants: sertraline and bupropion. Pizzagalli explained the reasoning behind looking at these specific options. “We wanted to test whether information collected before treatment, including brain connectivity, reward learning, cognitive performance, and clinical characteristics, could help identify who is most likely to benefit from two commonly prescribed antidepressants, sertraline and bupropion,” he said.
Sertraline belongs to a class of drugs that increases levels of the chemical serotonin in the brain to help regulate mood. Bupropion operates differently, targeting other brain chemicals called dopamine and norepinephrine, which are associated with motivation and reward processing. Because these drugs work through different chemical pathways, researchers suspected that distinct biological profiles might indicate which medication tends to be most effective for an individual patient.
The authors first developed a predictive tool using data from a previous large clinical trial known as the EMBARC study. “An important feature of this work is that we did not simply examine existing treatment data after the fact,” Pizzagalli said. “We first developed the prediction models in an earlier, independent study and then tested them prospectively in a double-blind clinical trial, making this an unusually rigorous early test of biomarker-guided antidepressant treatment.”
To build this predictive algorithm, the researchers examined various potential biomarkers, including demographic details like employment status and clinical factors such as depression severity and personality traits like neuroticism. They also included specific cognitive tests and brain imaging data to round out the predictive tool. By analyzing these past records, they built a mathematical model to identify the key traits of people who responded well to each medication.
One specific measure in this model involved functional magnetic resonance imaging, or fMRI. This technology measures brain activity by detecting changes in blood flow over time, allowing scientists to see how different regions communicate. The researchers looked at the resting connection between two brain areas involved in processing rewards and regulating emotions. These two regions are the nucleus accumbens, located deep within the brain, and the rostral anterior cingulate cortex, situated near the front of the brain.
Along with brain scans, the predictive tool incorporated specific behavioral tests to measure cognitive functions. Participants completed a computer-based assessment called the Probabilistic Reward Task to measure reward sensitivity, which evaluates how strongly a person responds to positive feedback. They also completed a cognitive control assessment known as the Flanker task, which measures a person’s ability to focus on specific targets while actively ignoring distracting information.
By combining these diverse data points, the algorithm assigned a positive or negative marker for both sertraline and bupropion. To test this algorithm in real time, the researchers enrolled a new group of 48 unmedicated adults experiencing major depression. Each participant underwent the baseline MRI scans, cognitive assessments, and clinical interviews. Researchers used standardized clinical questionnaires, such as the Montgomery-Åsberg Depression Rating Scale, to quantify the exact severity of each participant’s depressive symptoms before starting any treatment.
Within a few days of these initial tests, the algorithm analyzed the new participants’ data to determine their biological markers. Out of the original group, 47 participants completed at least one week of treatment and were included in the main analysis. Following their biomarker assessment, participants were randomly assigned to receive either sertraline or bupropion.
Half of the group received a medication that was consistent with their biological marker, meaning the algorithm predicted they would respond well to it based on their specific biology. The other half received a medication that was inconsistent with their marker profile. To prevent any bias, the trial was double-blind, meaning neither the patients nor the doctors evaluating them knew which specific medication had been prescribed or what the biomarker results were.
The researchers tracked the participants’ depression symptoms throughout an eight week period to see how they progressed. They anticipated that individuals matched with the correct drug would show a stronger improvement. “We expected people to do better when they received the specific medication indicated by their biomarker profile, but that was not what we found,” Pizzagalli told PsyPost.
“Instead, the markers appeared better at distinguishing people who were generally more likely, or less likely, to respond to either medication: the response rate was approximately 71% among participants with positive markers for both drugs, compared with about 43% among those with neither marker,” Pizzagalli said. The data showed that assigning a drug strictly based on an aligned marker did not produce a statistically significant difference in symptom relief compared to an unaligned assignment. The specific medication a participant received seemed to matter less than the overall presence of positive biomarkers.
Regardless of the drug assignment, participants who possessed positive markers for both drugs, or at least one of the drugs, experienced greater reductions in their depression symptoms than those who had two negative markers. Individuals with a marker for just one of the drugs had a response rate of 65.4 percent. The study defined a response as a 50 percent or greater reduction in clinical depression scores from the beginning to the end of the trial.
“Our results offer an encouraging but preliminary step toward making depression treatment more personalized,” Pizzagalli said. “Participants whose profiles included at least one positive treatment-response marker improved more than those whose profiles suggested that neither medication was likely to help; however, the markers did not reliably tell us whether sertraline or bupropion was the better choice for a particular individual.”
There are some potential misinterpretations that readers should avoid. “These findings do not mean that a brain scan or behavioral test can currently tell a patient which antidepressant to take,” Pizzagalli said. “This was a relatively small study, and our primary comparison, whether patients did better when matched to the medication indicated by their markers, was not statistically significant; the findings therefore need to be replicated in substantially larger and more diverse samples.”
The reliance on functional magnetic resonance imaging presents another practical barrier for widespread clinical application. Brain scans are expensive, time consuming, and require specialized equipment that is not easily accessible in everyday primary care settings. Additionally, variations in how different MRI machines process data can make it difficult to generalize the algorithm across different hospitals or research centers.
Addressing these logistical hurdles will require establishing standardized procedures across the medical field. “The next step is to test and refine these prediction models in larger, multisite clinical trials,” Pizzagalli said. “We also want to develop markers that are easier and less expensive to use in routine care, and to determine whether different profiles predict response not only to conventional antidepressants, but also to treatments such as brain stimulation, ketamine, and other rapid-acting therapies.”
The study, “A precision medicine trial of bupropion and sertraline for major depressive disorder using a biomarker-guided sequential multiple-assignment design,” was authored by Peter Zhukovsky, Manuel Kuhn, Lauren R. Borchers, Boyu Ren, Sarah E. Woronko, Mohan Li, Choi Sze Tracy Lam, Ethan M. Zhang, Kerry J. Ressler, Brian P. Brennan, Gordana Vitaliano, and Diego A. Pizzagalli.