In order to study the effect of the alpha EEG power increasing training at heart rate variability (HRV) as the index of the autonomic regulation of cognitive functions there were follow tasks: (1) to figure out the impact of biofeedback in the voluntary increasing the power in the individual high-frequency alpha-band effect on heart rate variability and related characteristics of cognitive and emotional spheres, (2) to determine the nature of the relationship between alpha activity indices and heart rate variability, depending on the alpha-frequency EEG pattern at rest (3) to examine how the individual alpha frequency EEG pattern is reflected in changes HRV as a result of biofeedback training. Psychometric indicators of cognitive performance, the characteristics of the alpha-EEG activity and heart rate variability (HRV) as LF/HF and pNN50 were recorded in 27 healthy men aged 18-34 years, before, during, and after 10 sessions of training of voluntary increase in alpha power in the individual high-frequency alpha band with eyes closed. To determine the biofeedback effect on the alpha power increasing training, data subjects are compared in 2 groups: experimental (14) with the real and the control group (13 people)--with mock biofeedback. The follow up effect of trainings was studied through month over the 10 training sessions. Results showed that alpha biofeedback training enhanced the fluency and accuracy in cognitive performance, decreased anxiety and frontal EMG, increased resting frequency, width and power in individual upper alpha range only in participants with low baseline alpha frequency. While mock biofeedback increased resting alpha power only in participants with high baseline resting alpha frequency and did change neither cognitive performance, nor HRV indices. Biofeedback training eliminated the alpha power decrease in response to arithmetic task in both with high and low alpha frequency participants and this effect was followed up over the month. Mock biofeedback training has no such effect. The positive correlation between the alpha-peak frequency and pNN50 in patients with initially low, but negative--those with high baseline alpha frequency explains the multidirectional biofeedback effects on HRV in low and high alpha frequency subjects. The individual alpha-frequency EEG pattern determines the effectiveness of the alpha EEG biofeedback training in changing heart rate variability, which provides a basis for predicting the results and develop individual approaches to the biofeedback technology implementation that can be used in clinical practice for treatment and rehabilitation of psychosomatic syndromes and in educational training.