Martin Bares1, Tomas Novak2, Premysl Vlcek3, Martin Hejzlar4, Martin Brunovsky5. 1. National Institute of Mental Health Czech Republic, Topolova 748, 250 67 Klecany, Czech Republic; Department of Psychiatry and Medical Psychology of Third Medical Faculty, Charles University, Ruská 87, 100 00 Prague 10, Czech Republic. Electronic address: martin.bares@nudz.cz. 2. National Institute of Mental Health Czech Republic, Topolova 748, 250 67 Klecany, Czech Republic; Department of Psychiatry and Medical Psychology of Third Medical Faculty, Charles University, Ruská 87, 100 00 Prague 10, Czech Republic. Electronic address: tomas.novak@nudz.cz. 3. National Institute of Mental Health Czech Republic, Topolova 748, 250 67 Klecany, Czech Republic; Department of Psychiatry and Medical Psychology of Third Medical Faculty, Charles University, Ruská 87, 100 00 Prague 10, Czech Republic. Electronic address: premysl.vlcek@nudz.cz. 4. National Institute of Mental Health Czech Republic, Topolova 748, 250 67 Klecany, Czech Republic; Department of Psychiatry and Medical Psychology of Third Medical Faculty, Charles University, Ruská 87, 100 00 Prague 10, Czech Republic. Electronic address: martin.hejzlar@nudz.cz. 5. National Institute of Mental Health Czech Republic, Topolova 748, 250 67 Klecany, Czech Republic; Department of Psychiatry and Medical Psychology of Third Medical Faculty, Charles University, Ruská 87, 100 00 Prague 10, Czech Republic. Electronic address: martin.brunovsky@nudz.cz.
Abstract
BACKGROUND: The study evaluated the effectiveness of EEG alpha 1, alpha 2 and theta power, along with prefrontal theta cordance (PFC), frontal and occipital alpha 1, alpha 2 asymmetry (FAA1/2, OAA1/2) at baseline and their changes at week 1 in predicting response to antidepressants. METHOD: Resting-state EEG data were recorded from 103 depressive patients that were treated in average for 5.1 ± 0.9 weeks with SSRIs (n = 57) and SNRIs (n = 46). RESULTS: Fifty-five percent of patients (n = 56) responded to treatment (i.e.reduction of Montgomery-Åsberg Depression Rating Scale score ≥ 50%) and 45% (n = 47) of treated subjects did not reach positive treatment outcome. No differences in EEG baseline alpha and theta power or changes at week 1 for prefrontal, frontal, central, temporal and occipital regions were found between responders and non-responders. Both groups showed no differences at baseline PFC, FAA1/2 and OAA1/2 as well as change of FAA1/2 at week 1. The only parameters associated with treatment outcome were decrease of PFC in responders and increase of OAA1/2 at week 1 in non-responders. There was no influence of the used antidepressant classes on the results. The PFC change at week 1 (PFCC) (area under curve-AUC = 0.75) showed only a numerically higher predictive ability than OAA change in alpha 1 (OAA1C, AUC = 0.64)/alpha 2 (OAA2C, AUC = 0.63). A combined model, where OAA1C was added to PFCC (AUC = 0.79), did not significantly improve response prediction. CONCLUSION: Besides PFCC, we found that OAA1C/OAA2C might be another candidate for EEG predictors of antidepressant response.
BACKGROUND: The study evaluated the effectiveness of EEG alpha 1, alpha 2 and theta power, along with prefrontal theta cordance (PFC), frontal and occipital alpha 1, alpha 2 asymmetry (FAA1/2, OAA1/2) at baseline and their changes at week 1 in predicting response to antidepressants. METHOD: Resting-state EEG data were recorded from 103 depressivepatients that were treated in average for 5.1 ± 0.9 weeks with SSRIs (n = 57) and SNRIs (n = 46). RESULTS: Fifty-five percent of patients (n = 56) responded to treatment (i.e.reduction of Montgomery-Åsberg Depression Rating Scale score ≥ 50%) and 45% (n = 47) of treated subjects did not reach positive treatment outcome. No differences in EEG baseline alpha and theta power or changes at week 1 for prefrontal, frontal, central, temporal and occipital regions were found between responders and non-responders. Both groups showed no differences at baseline PFC, FAA1/2 and OAA1/2 as well as change of FAA1/2 at week 1. The only parameters associated with treatment outcome were decrease of PFC in responders and increase of OAA1/2 at week 1 in non-responders. There was no influence of the used antidepressant classes on the results. The PFC change at week 1 (PFCC) (area under curve-AUC = 0.75) showed only a numerically higher predictive ability than OAA change in alpha 1 (OAA1C, AUC = 0.64)/alpha 2 (OAA2C, AUC = 0.63). A combined model, where OAA1C was added to PFCC (AUC = 0.79), did not significantly improve response prediction. CONCLUSION: Besides PFCC, we found that OAA1C/OAA2C might be another candidate for EEG predictors of antidepressant response.
Authors: Ian A Cook; Aimee M Hunter; Marissa M Caudill; Michelle J Abrams; Andrew F Leuchter Journal: J Psychiatr Res Date: 2020-02-26 Impact factor: 4.791