Effects of central cortical EEG feedback training on incidence of poorly controlled seizures.

This study examined the clinical effects of central cortical EEG feedback training in 8 patients with poorly controlled seizures. After base-line recordings, patients were trained in the laboratory and then initiated on a double or triple crossover design using portable equipment at home, with bimonthly laboratory test sessions. Performance at home was monitored by a strip chart recorder with the portable unit. Training was based on the simultaneous detection of two central cortical (C3--T3) EEG frequency bands (6--9 Hz and either 12--15 or 18--23 Hz), with reward provided for the occurrence of one in the absence of the other. The design consisted of successive 3 month periods of training, with reward contingencies reversed after each period without the subject's knowledge. Seizure incidence records were compared statistically before, during, and after the design. Six of the 8 patients reported significant and sustained seizure reductions, which averaged 74%, following reward for either 12--15 or 18--23 Hz in the absence of 6--9 Hz. Response to positive reward for 12--15 Hz was specific, with seizure rates returning to base line when reinforcement contingencies were reversed. Reduced seizure rates following positive reward for 18--23 Hz were not altered with contingency reversals. A nonspecific interpretation of these effects is rejected in favor of an EEG normalizing hypothesis.