Biofeedback training of heart rate in essential hypertensives

The objective of this study was to help clarify the disparate results reported in the biofeedback literature as to whether, under conditions of biofeedback training of heart rate (HR), HR and blood pressure (BP) are uncorrelated (dissociated) or positively correlated. This is relevant to the question of whether the physiological response is specific and limited to the parameter constituting the feedback signal or whether the response is nonspecific and generalized. Essential hypertensive subjects were used to assess the predicted positive shift in HR-BP correlations from early to later sessions since a primary clinical objective of cardiovascular self-regulation is the possible therapeutic application with this patient population. Thus, six hypotensively medicated essential, and one renal, hypertensive subjects were given ten biofeedback training sessions in HR slowing. For each subject within-session mean HRr systolic BP and diastolic BP values were computed for the training trial and alternating rest periods. Within-session changes in these parameters were calculated relative to that session's pretraining resting baseline levels, and, as a control, relative to adjusted progressively changing baselines in four of these same subjects. These within-session changes were used to calculate training and rest period HR-SBP and HR-DBP correlation coefficients across within-subject sessions grouped into early (2-4), middle (5-7) and late (8-10) sessions. Determinations were then made of the shifts in the correlation from early to middle, from early to late and from middle to late sessions. This was accomplished by computing the differences between the Fisher transforms of each of the twenty-four (6 Ss x 2 BPs x training and rest conditions) correlational determinations in each of the early, middle, and late sessions. The experimental findings showed that in the early sessions HR and BP were negatively correlated in twenty-three of the twenty-four determinations, and were significantly negative (p <.05) in seventeen of these. By the middle sessions there had been twenty-three positive shifts in correlation with seventeen of these being significant. And by the late sessions there were positive shifts in all twenty-four determinations, with twenty-one of these being statistically significant. These relationships were maintained under the control condition of adjusted baselines. Thus, the experimental findings support both the prediction of a negative HR-BP correlation in the early sessions and the prediction of a positive shift in these correlations as a function of the number of training sessions. These results do not lend support to the view that feed back training results in a feedback-parameter-specific physiological response.