Virtual defense: https://us02web.zoom.us/j/89879746768?pwd=MnV3L3BCU0lHTzZrWndUSTQ5MTBKQT09 Meeting ID: 898 7974 6768, Password: 310520
Professor Poul Jennum (Chairperson)
Dr Martin Brunovský
Professor Martijn Arns
Professor Gitte Moos Knudsen
Clinical Pharmacology Scientist Søren Rahn Christensen
Associate Professor Sándor Beniczky
Major depressive disorder (MDD) is a heterogeneous disorder with potentially diverse pathophysiological mechanisms. This diversity may account for the observation that far from all patients benefit from the same treatment. Consequently, identification of MDD subgroups would allow for treatment stratification which could improve clinical trial outcomes and lead to a precision medicine strategy in patients. However, previous studies examining whether electroencephalography (EEG) can predict the effect of antidepressant treatment have included low sample size, and independent replication are needed to confirm these findings.
The purpose of this PhD work was to a) examine the test-retest reliability of an EEG battery in healthy males who were given different antidepressants, b) to validate EEG candidate biomarkers that have previously shown some predictive value in our own cohort of unmedicated patients with moderate to severe MDD (NeuroPharm Trial) c) explore the effect on EEG measures after 8 weeks of selective serotonin reuptake inhibitor/ serotonin noradrenalin reuptake inhibitor (SSRI/SNRI) treatment.
In study Ⅰ, we investigated the test-retest reliability of an EEG battery in 32 healthy males who were given four different antidepressant regimens. We compared baseline EEG recordings from the four interventions to assess whether EEG/ERP (event-related potentials) were stable over time. We found that middle frequency bands (θ, α and β) of continuous EEG were highly reliable while evoked power of task-related potentials was less stable. Furthermore, though the reliability of ERP measures in general was lower compared to power measures, large components such as P300 and Pe still exhibited fair to excellent reliability. Our results support that these EEG/ERP parameters are reliable over a three-week interval. II
In study Ⅱ, we aimed to replicate previously reported EEG predictive biomarkers for treatment outcome by using an independent cohort of 91 antidepressant-free outpatients and 35 healthy controls. We found that only 2 out of 6 chosen biomarkers could be partially validated; both of which involved alpha asymmetry. The results indicate that measurement of alpha asymmetry carries information that improves prediction of treatment efficacy.
In study Ⅲ, used EEG data from the NeuroPharm data to determine the predictive value of vigilance regulation, we found that patients with MDD showed a hyperstable EEG-wakefulness regulation compared to healthy controls, replicating prior work. Treatment responders showed faster decline in vigilance regulation in comparison to non-responders at pretreatment. Furthermore, patients with good treatment response after 8 weeks of SSRI/SNRI treatment had their EEG-wakefulness regulation patterns reverted to look more like that of controls.
These findings support that EEG vigilance measures adds value diagnostically as well as in predicting treatment outcome in patients with MDD. Overall, the low cost of and methodological simplicity of EEG makes it a good tool for the optimization of patient stratification in future clinical trials and may even have value when choosing drug treatment for MDD patients.