Objective. To study the specificity of physiological mechanisms determining the efficiency of modelled purposeful activity in healthy people and patients with epilepsy. Material and methods. The sample consisted of 235 people, including 72 healthy people (47 men and 25 women) and 163 patients with epilepsy (84 men and 79 women). Emotional and cognitive disorders were scored. Purposeful activity was modelled using the Gorbov—Shulte test. EEG, visual and auditory evoked potentials, P300 wave were recorded. The «NS-Psychotest» complex was used to record a simple visual-motor reaction, a choice of difference and finger-tapping test. Heart rate variability and external breath functions were studied. An artificial neural network method was used for the prognosis of purposeful activity. Results and conclusion. The «resultative» and «low resultative» groups were selected on the basis of the Gorbov—Shulte test performance. Patients with symptomatic forms of epilepsy, higher frequency of generalized seizures and cognitive disorders were mostly represented in the low resultative group. Artificial neural network classification into groups with different efficiency of activity reveals the greater role of the parameters of afferent and associative mechanisms in healthy people, parameters of cross correlation function of EEG and motor system function in patients with epilepsy.