70 patients with probable Alzheimer's disease were randomly allocated to four groups: 17 patients received only social support, 18 cognitive training twice a week, in 17 cognitive training was combined with pyritinol 2 x 600 mg/day and in 18 cognitive training was combined with phosphatidylserine 2 x 200 mg/day. Treatment duration was 6 months. Before and after treatment, the patients underwent neuropsychological testing as well as measurement of the regional cerebral metabolic rate for glucose using positron emission tomography and 18F-2-fluoro-2-deoxy-D-glucose. Before treatment the groups were comparable in respect to resting and activated glucose pattern achieved by a visual recognition task. Electrophysiological changes were assessed as EEG power, globally and in 4 frequency bands. This 6-month study in four groups of patients with Alzheimer's disease indicated that phosphatidylserine treatment has an effect on different measures of brain function. Since neuropsychological improvements were best documented after 8 and 16 weeks and faded towards the end of the treatment period, it must be concluded that this symptomatic therapy is mainly of short-term benefit and was overcome by the progressive pathological changes at the end of the treatment period.