The aim of this study was to investigate whether brain information storage and processing in one-month abstinent heroin addicts differ from healthy controls, as reflected by the P600 component of Event-related potential (ERP) signals elicited during a working memory test, as well to search deeper into the P600 signals by extracting new features and, by employing powerful classification procedures, to develop a pattern recognition system for discriminating drug users from controls. Sixteen one-month abstinent heroin addicts and twenty normal controls were examined. All subjects were evaluated by a computerized version of the digit span subtest of the Wechsler Adult Intelligence Scale. EEG activity was recorded from 15 scalp electrodes (leads). From the P600 component of each signal nineteen waveform-features were calculated. The cubic least-square minimum-distance classifier was developed and it was fed with features from all leads. The system was evaluated by means of the exhaustive search and leave-one-out methods. Highest single-lead precision (86.1%) was found at the P3, C5 and F3 leads at left. When leads were grouped into compartments, highest accuracies were achieved at the temporo-central region (88.9%). These findings may be indicative that one-month abstinent heroin addicts present deficits in working memory processes, as reflected by the P600 component.