Systematic review and meta-analysis of neurofeedback training efficacy and neural mechanisms in the treatment of posttraumatic stress disorder

Daniel E Berman^1,2Kiriana P Cowansage^1,2Dawn M Bellanti^1,2Reshmi Nair^1,2Courtney C Boyd^1,2Erin H Beech³Madhavi K Reddy⁴Robyn Recker⁴Bradley E Belsher⁵Marija S Kelber^1,2*

• ¹Defense Health Agency (DHA), Silver Spring, United States
• ²Psychological Health Center of Excellence, Silver Spring, United States
• ³VA Portland Health Care System, Portland, United States
• ⁴Walter Reed Army Institute of Research, Silver Spring, United States
• ⁵Carl T Hayden VA Medical Center, Phoenix, United States

Neurofeedback in the treatment of psychological disorders has gained increasingly widespread interest in recent years. As the use of neurofeedback training expands, it is important to elucidate its treatment efficacy, especially for prevalent and debilitating psychopathologies such as posttraumatic stress disorder (PTSD). Likewise, furthering our understanding of the underlying neural mechanisms by which neurofeedback acts is also necessary. Here, we present the results of a PROSPERO registered (protocol number: CRD42020184659) meta-analysis of randomized controlled trials (RCTs) of neurofeedback training for treatment of PTSD in adults. We evaluate the efficacy of different neurofeedback modalities used to treat PTSD, including functional magnetic resonance imaging neurofeedback (fMRI-NF) and electroencephalogram neurofeedback (EEG-NF). We also differentiate active control (sham or yoked sham) studies from passive control (waitlist, treatment as usual, no treatment, and any non-neurofeedback based intervention) studies. Our results show that EEG-NF has a moderate to large effect in reducing PTSD symptoms pre-to post-treatment compared to passive controls (k=5). Two fMRI-NF RCTs, both using sham controls, showed no improvement in PTSD symptoms, pre-to post-treatment. However, our confidence in these findings is very low to low due to concerns regarding risk of bias, imprecision, and conflicts of interest. Neurofeedback in passive control studies outperformed neurofeedback in active control studies relative to their respective control treatment arms. This highlights the need for improved controls in studies examining neurofeedback for PTSD to reliably determine whether neurofeedback training, or other factors, are the basis for improvements in PTSD symptoms. We also synthesized the neural results from these studies and found that between-group neural effects were generally inconclusive. We elaborate on some of the underlying neural mechanisms by which neurofeedback training shows potential in improving PTSD symptoms to guide future developments and provide recommendations for future neurofeedback interventions in treating PTSD.