Will You Respond to Antidepressant Medication? A New AI Technology Might Be Able To Tell You
August 7, 2020
Main image courtesy of Catholic Health.
Research at Stanford University has brought about a promising form of technology that could potentially be applied to the future diagnosis protocol for depression. By using simple electroencephalography (EEG) in conjunction with a little bit of artificial intelligence, the Stanford researchers were capable of predicting which of the patients suffering from depression would most benefit from treatment with the antidepressant medication sertraline, known more commonly as Zoloft.
With depressed patients responding quite differently across the board to different treatments, and especially with many patients suffering from treatment-resistant depression, these findings could prove extremely beneficial for the future of depression treatment. In seeing which patients specifically will respond well to prescription antidepressants like Zoloft, and even which patients will not (or who may respond better to an alternate treatment), this research has unlocked key information to apply to psychiatry moving forward.
What Exactly Did the Researchers Learn Through All of This Data?
There is much to learn about how depressed patients respond differently to various treatments, and this study’s technology just may have made some major advancements in our understanding of this phenomenon.
In this recent study, Stanford University was able to look at data from the EMBARC study, which was considered to be the biggest placebo-controlled antidepressant study of its kind that utilized image-guiding technology. The study included 309 patients suffering from depression, either with or without the common antidepressant medication Zoloft. The patients’ brain activity was recorded using EEG technology to look at the electrical activity before they began treatment.
The researchers then utilized artificial intelligence to build an advanced predictive model that was based on a new machine-learning algorithm referred to as SELSER. This particular technology is great for analyzing EEG data, so they designed their model using theories based in neuroscience, biotechnology, and clinical science. The researchers used SELSER by applying this technology to all of the EEG data that they had collected from the patients, hoping that it would be able to therefore predict the patients’ depressive symptoms following their treatment.
To the researchers’ surprise, the SELSER technology was capable of reliably predicting how individual depressed patients would respond to Zoloft treatment. It was able to do so based on alpha waves, a specific type of brain signal that is associated with reduced processing capacity of a particular brain area, in conjunction with general states of relaxation.
One of the brain areas that is of particular interest when it comes to those suffering from depression is the prefrontal cortex, an area of the brain that plays an important role in our emotional reactions. This brain region specifically tends to be impaired in depressed individuals. Interestingly, the researchers noted from the data that the patients who responded best to the antidepressant treatment tended to have a more active and excitable prefrontal cortex.
In previous research, EEG data has been used in an attempt to predict symptom severity and demographic information. The researchers in this study were pleasantly surprised to observe that this particular EEG-based model that they created had essentially surpassed these more conventional models in their abilities to make these types of predictions.
How Does This Research Relate to TMS Therapy?
Not only was this EEG-based model able to reliably predict patients’ response to antidepressant medication, but it was also associated with how well they would respond to other treatments like TMS.
Another independent data set was able to gather some important findings regarding treatments outside of just antidepressant medications like Zoloft. Using this data, the researchers were able to determine that the patients who SELSER predicted a small improvement in after the antidepressant treatment were also more likely to respond to transcranial magnetic stimulation (TMS) therapy alongside psychotherapy. Psychotherapy is a form of “talk therapy” in which the patient participates in a series of therapy sessions to work through the issues and symptoms that they are experiencing.
For those who are not entirely familiar with TMS therapy, it is a non-invasive, alternative treatment which relies on electrical stimulation of specific areas of the brain that may be associated with a person’s symptoms. TMS therapy is an FDA-approved, safe treatment for severe and treatment-resistant depression. By restoring your body’s balance of neurotransmitters, TMS therapy can help depressed patients lessen their symptoms and start feeling like themselves again.
The fact that the research data found implications for patient response to other types of therapy like TMS therapy is extremely promising. Combining various treatment techniques is often a better route for patients with treatment-resistant depression. Even more importantly, some patients have particular preferences when it comes to what treatment they feel most comfortable with.
So What is the Importance of This Research and What Does It Mean For the Future of Depression Treatment?
Finding that a depression treatment does not work is extremely disheartening and discouraging for patients, but these research findings may be able to give them more hope.
The researchers in this study note that one in five people in the United States experience or will experience depression in their lifetime, and unfortunately only 30% of these patients will actually respond to the treatment that is prescribed to them. Usually, when patients who begin to suffer from depression seek treatment, the first step is prescription of an antidepressant medication. If this medication fails to work, then the second step is typically another prescription for a different antidepressant medication.
What is particularly unfortunate about antidepressant prescriptions is that in order to figure out whether or not a specific medication is going to work well for a patient, they need to be taking the medication consistently for at least eight weeks. In waiting this long just to make an assessment as to whether or not symptoms have been reduced, patients can become extremely discouraged, which can actually contribute and feed into their depression.
Once it is decided that antidepressant medications are not working for a patient, the next step is usually TMS therapy or psychotherapy, which we have touched on previously in this article. The researchers mention that current diagnostic methods for depression are just too imprecise and often just subjective, and that it makes it difficult for the physicians to accurately determine the proper treatment for their patients.
Fortunately, the findings of this research are possibly beginning to resolve some of these discrepancies in the current methods for prescribing treatments to patients. An antidepressant medication failing to work can be quite devastating to the patient, but the hope is that patients will have to endure much less trial and error situations before finding a treatment that works for them.
This research study has proven to be extremely promising in terms of the future of depression treatment. Stanford researchers have demonstrated that a combination of artificial intelligence and EEG technology is capable of making useful diagnostic predictions in depressed patients, more specifically as to whether or not they will respond well to the antidepressant medication Zoloft. In addition, they learned that other treatments like TMS therapy and psychotherapy also produce a response in these patients who responded well to the prescription medication. As we continue to learn more about the patient response to the many types of treatments that exist for depression, we can hopefully learn more ways to predict these responses and get patients quicker access to the treatments that are best suited to them as an individual.