The relationships between memory processes and oscillatory electroencephalography (EEG) are well established. Neurofeedback training (NFT) may cause participants to better regulate their brain EEG oscillations. The present study is a double-blind sham-controlled design investigating the effect of NFT on memory. NFT included up-training upper alpha (UA) band, up-training sensory-motor rhythm (SMR) band and sham protocol. Thirty healthy adult volunteers were randomly divided into three treatment groups. NFT sessions (30 min each) took place twice weekly for a total of 10 sessions while memory testing took place pre- and post-training. The results indicate dissociation between SMR and UA NFT and different memory processes. While the SMR protocol resulted in improving automatic, item-specific and familiarity-based processes in memory, the UA protocol resulted in improved strategic and controlled recollection. The implications of the results are discussed.