Major depression is currently defined based on clinical criteria and encompasses a heterogeneous mix of neurobiological phenotypes. This heterogeneity may account for the modest superiority of antidepressant medication over placebo.
Antidepressants are widely prescribed, but their efficacy relative to placebo is modest, in part because the clinical diagnosis of major depression encompasses biologically heterogeneous conditions.
Past research has proposed that specific components of brain activity, as estimated by resting-state electroencephalography (EEG), could yield knowledge into how individuals will react to particular medications. Be that as it may, scientists have yet to develop prescient models that can differentiate between response to antidepressant medication and response to placebo, and that can likewise anticipate results for individual patients. Both features are essential for the neural signature to have clinical relevance.
Amit Etkin, M.D., Ph.D., a professor of psychiatry and behavioral sciences at Stanford University along with Madhukar H. Trivedi, M.D., a professor of psychiatry at the University of Texas Southwestern Medical Center, Dallas, and first author Wei Wu, Ph.D., an instructor at Stanford University, California, drew insights from neuroscience, clinical science, and bioengineering to build an advanced predictive model.
They developed a new machine learning algorithm specialized for analyzing EEG data called SELSER (Sparse EEG Latent SpacE Regression).
Scientists noted, “This AI can predict from people’s brainwaves whether an antidepressant is likely to help them. The technique may offer a new approach to prescribing medicines for mental illnesses.”
Scientists used SELSER to analyze data from the NIMH-funded 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). As a feature of the examination, members with depression were arbitrarily assigned to get either sertraline or placebo treatment for about two months.
The scientists applied SELSER to members’ pre-treatment EEG data, inspecting whether the AI could create a model that anticipated participant’s depressive symptoms after treatment.
The AI was able to reliably predict individual patient response to sertraline based on a specific type of brain signal, known as alpha waves, recorded when participants had their eyes open. This EEG-based model beat regular models that utilized either EEG information or different sorts of individual-level data, for example, symptom severity and segment attributes. Examinations of independent data sets, using a few basic strategies, proposed that the predictions made by SELSER may stretch out to more extensive clinical results past sertraline response.
Scientists found that the EEG-based SELSER model predicted more significant improvement for participants who had shown partial response to at least one antidepressant medication compared with those who had not responded to two or more medications, in line with the patients’ clinical outcomes.
Work is now underway to replicate these findings in large, independent samples further to determine the value of SELSER as a diagnostic tool. According to Etkin, Trivedi, Wu, and colleagues, the present research highlights the potential of machine learning for advancing a personalized approach to treatment in depression.
Etkin said, “While work remains before the findings in our study are ready for routine clinical use, the fact that EEG is a low-cost and accessible tool makes the translation from research to clinical practice more possible in the near term. I hope our findings are part of a tipping point in the field concerning the impact of machine learning and objective testing.”
The study is published in the journal Nature Biotechnology.