EEG May Speed Antidepressant Choice

MDD remission declines with number of medication trials

In adults with major depressive disorder (MDD), analysis of data from EEG recordings was able to predict response to a popular antidepressant with an accuracy rate of 79.2%. The findings, published online January 3 in JAMA Network Open demonstrate the potential utility of the EEG as a treatment planning tool in depression.

Treatment of depression typically is characterized by prolonged periods of pharmacologic trial and error to identify optimal therapy for an individual patient. The resulting cost is both social and personal as the burden of depression may be prolonged for months or years.

Authors of the new study propose to reduce the time spent in “failed trials” by using a tool that identifies biologic predictors of antidepressant response. The research team was led by Faranak Farzan, PhD, scientific director, Center for Engineering-Led Brain Research, Simon Fraser University, Burnaby, British Columbia, Canada.

Commenting in their introduction on previous EEG studies, the authors commend the compelling findings but note results were limited by small sample size  (10 to 15 participants from a single center), reports of poor prediction accuracy, and prediction accuracy biased by lack of an independent testing set.  

Large, multicenter study

To control for similar limitations, Farzan and colleagues used data from the Canada-wide multicenter study, the Canadian Biomarker Integration Network in Depression (CAN-BIND) study.  

Resting state EEG recordings were collected before treatment began from 122 participants meeting criteria for MDD (mean age, 36.3 years; 62.3% women) and for 115 of these patients 2 weeks after treatment initiation with the selective serotonin reuptake inhibitor escitalopram.

Mean baseline score on the Montgomery-Åsberg Depression Rating Scale (MADRS) for the 122 participants was 30.1.

All participants completed 8 weeks of open-label treatment with escitalopram. Responders were identified as subjects who demonstrated a ≥50% reduction in MADRS score from baseline to week 8.

For responders, the mean MADRS score was 29.5 at baseline and 7.9 at week 8.

The primary outcome measure was the ability of EEG data to predict treatment outcome, measured as accuracy, specificity, and sensitivity of the computer model at baseline and after the first 2 weeks of treatment.

EEG data were analyzed to detect patterns that could predict treatment response. Analysis was performed using a support vector machine-a supervised machine learning algorithm that probes data for classification and regression analysis.

Results

Results showed that, using only the baseline EEG data, the classifier could identify responders with an estimated accuracy of 79.2% (sensitivity, 67.3%; specificity, 91.0%). No significant difference in accuracy was found between study centers.  

Among the 115 patients whose EEG data were recorded after 2 weeks of treatment, classifier accuracy increased to 82.4% (sensitivity, 79.2%; specificity, 85.5%).

Is this the future of depression Rx?

In the study’s discussion the authors say that their findings suggest use of machine learning might speed identification of the right medication for MDD patients. Research shows that remission rates decline progressively with the number of medications tried, beginning with 30% to 40% remission after a first medication trial and up to 55% after a second trial of a different antidepressant.

In discussing the results, they caution that:  “In this study, we demonstrated feasibility of predicting an outcome of escitalopram treatment using resting-state EEG. Of note, the prediction was that of the treatment outcome and not of the patient’s response to escitalopram.”

To definitively attribute predicted differences in treatment outcomes to individual differences in patients’ response to escitalopram would require a randomized clinical trial.

(image ©lightpoet/stock.adobe.com)

Stay in touch with Patient Care^® Online:
→  Subscribe to ourNewsletter
→  Like us on Facebook 
→  Follow us on Twitter →  Follow us on LinkedIn  →  Write or Blog for Patient Care^® Online