Objectives. According to some studies, a putatively calming effect of EEG neurofeedback training could be useful as a therapeutic tool in psychiatric practice. With the aim of elucidating this possibility, we tested the efficacy of a single session of ↑sensorimotor (SMR)/↓theta neurofeedback training for mood improvement in 32 healthy men, taking into account trainability, independence and interpretability of the results. Methods. A pre-post design, with the following dependent variables, was applied: (i) psychometric measures of mood with regards to anxiety, depression, and anger (Profile of Mood State, POMS, and State Trait Anxiety Inventory, STAI); (ii) biological measures (salivary levels of cortisol); (iii) neurophysiological measures (EEG frequency band power analysis). In accordance with general recommendations for research in neurofeedback, a control group receiving sham neurofeedback was included. Results. Anxiety levels decreased after the real neurofeedback and increased after the sham neurofeedback (P < 0.01, size effect 0.9 for comparison between groups). Cortisol decreased after the experiment in both groups, though with significantly more pronounced effects in the desired direction after the real neurofeedback (P < 0.04; size effect 0.7). The group receiving real neurofeedback significantly enhanced their SMR band (P < 0.004; size effect 0.88), without changes in the theta band. The group receiving sham neurofeedback did not show any EEG changes.