Stanford Researchers & Collaborators Develop SELSER to Identify Brain Wave Signatures to Predict SSRI Response

Stanford Researchers & Collaborators Develop SELSER to Identify Brain Wave Signatures to Predict SSRI Response

The National Institutes of Mental Health (NIMH), part of the National Institutes of Health (NIH), have funded research that has led to the discovery of a neural signature predicting whether individuals with depression could likely benefit from sertraline, a commonly prescribed antidepressant. Led by Wu Tsai Neuroscience Institute, Stanford University and UT Southwestern, the findings suggest that new machine learning techniques can help to pinpoint complex patterns in the brain that correlate with meaningful clinical outcomes. SELSER is born.

The Challenge

About 7% of adults in the United States are affected by common mental disorders including depression. However, symptoms can vary widely—from person to person. Some may experience many standard characteristic features such as persistently sad mood and feelings of hopelessness to other people with few such symptoms. Although a number of evidence-based approaches for treating depression exist, it is difficult for physicians to assess which option is superior for the patient—treating depression almost becomes a trial and effort process for many, reports National Institutes of Health (NIH).


Published recently in journal Nature Biotechnology the researchers pursue new machine learning techniques to aid in the identification of complex patterns in a person’s brain that correlate with meaningful outcomes. The team capitalizes on past research that implies the specific components of brain activity—measured by restating-state electroencephalography (EEG) can potentially produce insight into how behavioral health patients could respond to targeted depression treatments. But researchers haven’t been able to develop sufficient predictive models able to differentiate between responses to antidepressant medication and response to placebo and that can predict outcomes for patients. As reported by the NIH press release, both features are essential for the neural signature to have relevance.

SELSER is Born

The research team—involving members from Stanford University  and University of Texas Southwestern Medical Center, Dallas drew on insights from neuroscience, clinical science and bioengineering to build an advanced predictive model. Based on machine learning algorithms, they designed an advanced tool for analyzing EEG data called SELSER (Sparse EEG Latent SpaceE Regression)

The multi-center team hypothesized that this advanced algorithm can identify robust and reliable neural signatures of antidepressant treatment response. Once SENSER was developed, the investigators harnessed the advanced machine learning algorithms to analyze data from the NIMH-funded and University of Texas, Southwestern Medical Center conducted Establishing Moderators and Biosignatures of Antidepressant Response in Clinic Care (EMBARC) study, a large randomized clinical trial of the antidepressant medication sertraline, a widely available selective serotonin reuptake inhibitor (SSRI).

During the EMBARC study, participants with depression were randomly assigned to receive either sertraline or placebo for eight weeks. By applying SELSER to participants’ pre-treatment EEG data, studying whether the machine learning technique could produce a model that predicted participants’ depressive symptoms after treatment.

SELSER Delivers

The team found that SELSER was able to reliably predict individual patient response to sertraline based on the measurement of alpha waves, a form of brain signal that is recorded when participants had their eyes open. Interestingly, this EEG-based model outperformed conventional models that used either EEG data or other types of individual-level data, such as symptom severity and demographic characteristics. More promising, the researchers believe that based on an analyses of independent data sets, using several complementary methods, predictions produced by SELSER may extend to broader clinical outcomes beyond sertraline response.

Next Steps

These investigators are now pursuing the replication of these findings in larger, independent samples to determine the true value of SELSER as a diagnostic tool. The present research highlights the powerful potential of machine learning for advancing personalized approach to treatment in depression.

Commercial Angle?

Note, study lead Amit Etkin of Stanford is also CEO of Alto Neuroscience which includes SELSER on its website. It would appear that Alto Neuroscience has the rights to distribute the non-commercial software.

What is the mission of Alto? They are developing a new generation of brain biomarker-based diagnostic tests and personalized treatments for use in mental health, based on a platform (SELSER et al) built from data aggregated at scale and cutting-edge artificial intelligence analysis tools.

Lead Research/Investigator

Amit Etkin, MD, PhD, professor of psychiatry and behavioral sciences, Stanford University, Wu Tsai Neuroscience Institute,  (CEO of Alto Neuroscience)

Madhukar Trivedi, MD, UT Southwestern

Call to Action: Interested in commercial opportunity with Alto Neuroscience? See a link to their website