Notwithstanding these uncertainties, the premise that an episode of current MDD is associated with AA has been the starting point for AA manipulation with the application of NF as a treatment for MDD. To date, case studies, and a small randomized open trial indicate that the increase of right relatively to left alpha activity at F3-F4 with the use of neurofeedback may be associated with a reduction in depressive symptomatology. This previous work has several limitations. First, NF treatment in the case studies was combined with psychotherapeutical sessions and lacked the use of state-of-the-art clinical instruments to assess psychiatric diagnoses and clinical change. The study by Choi et al, delivered only a total of 10 NF sessions during 5 weeks, which is considerably lower and less frequent than typically offered in the case studies and by NF practioners. Additionally, their participants suffered from subclinical levels of depression severity. Although the authors concluded that significant clinical change occurred in their active treatment group, clear criteria were not defined. We decided to carry out a pilot study to address several questions. First, we aimed to examine whether NF is effective in the treatment of moderate severe MDD using current clinical instruments based on clearly defined response and remission criteria. Second, we investigated whether AA indeed decreased during the course of the NF sessions. Third, we examined the association between changes in clinical state and changes in AA. Lastly, the optimal duration of a single NF session for depressed subjects is unknown. Given fatigue and difficulties in concentration in MDD, we investigated the time-course of changes in AA during NF sessions to assess optimal session duration. In the Presentation paradigm, the last 20 values of the asymmetry are used in a moving average to prevent in the feedback. Participants received feedback with visual feedback; they were instructed to increase the level of a thermometer that was shown on a flatscreen. Additionally, a numerical score below the thermometer indicated their actual total performance. This score was adjusted continuously by a number ranging from 0 and 128, depending on the level of the thermometer. In this way a good actual performance resulted in an increasing total score. A big shift in the desired direction resulted in a rapidly increasing total score, whereas a small shift in the desired direction resulted in a slow increasing total score. A shift in the undesired direction produced no change in total score. The purpose of this total performance score was to give participants feedback on the differential effect of the sessions. Upon arrival in our laboratory, subjects were shown the facility and the monitor that displays the thermometer and the numerical score. After we had established a good EEG signal, they were just instructed to try to increase the level of the thermometer by trial and error. A typical response of subjects was the question if we could provide any detailed instructions that could be helpful in increasing the thermometer level. We outlined that there are no specific strategies known to be helpful in this respect based on our experience; we had tested this with healthy volunteers.