Heart rate variability: An autonomic correlate of reaction time performance

Heart rate variability: An autonomic correlate of reaction time performance* STEPHEN w. PORGESt University of Illinois, Champaign, Ill. 61820 Resting level heart rate variablity was measured and correlated with subsequent reaction time performance in male college students. The Ss were required to respond as rapidly as possible following an extended visual warning signal of either fixed or variable duration. Resting level heart rate variability was related to reaction time only in the variable-foreperiod condition. The data are interpreted as illustrating the sensitivity of the relationship between heart rate variability and reaction time to slight changes in task demands. Lacey & Lacey (1958) hypothesized a relationship between spontaneous autonomic fluctuations and reaction time. They proposed that the frequency of spontaneous heart rate and skin resistance fluctuations would be correlated with reaction time performance. This hypothesis assumed that the autonomic afferents initiated by these spontaneous fluctuations would facilitate the cortical mechanisms associated with attentional processes. In response to the Laceys' work, there have been reports of experiments relating spontaneous fluctuations in skin resistance (Rice, 1970; Wallace & Fehr, 1970) and heart rate (porges, 1972) to reaction time performance. This research has supported the Lacey and Lacey hypothesis by demonstrating that in adults and normal children greater fluctuations in skin resistance and heart rate were associated with faster reaction times. In the present study, heart rate variability was used as a measure of spontaneous heart rate fluctuation. The study was designed to investigate the relationship between heart rate variability and reaction time performance in a simple reaction time task with an extended preparatory interval. Independent groups of Ss were tested under two different reaction time conditions, variable foreperiod or fIXed foreperiod. The two foreperiod schedules were used to differentiate the predictive value of the resting heart rate variability under conditions of varying attentional demands. It was assumed that greater attentional demands were required when the termination of the foreperiod could not be anticipated. METHOD Subjects Thirty-six male volunteers from introductory psychology courses received extra course credit for serving as Ss. As they appeared at the laboratory, they were assigned to the experimental conditions according to a predetermined random schedule. *This work was supported in part by Grant MH-20329-01 from the National Institute of Mental Health. tAddress requests for reprints to Stephen W. Porges, Department of Psychology, University of Illinois, Champaign, Illinois 61820. 270 Apparatus Stimuli were programmed by means of magnetic tape and were presented automatically. Reaction time was measured by a Standard electronic clock. The ambient noise level of the experimental room was approximately 50 dB (re .0002 dynes/cm 2 ). Room temperature was maintained at approximately 70°F. Heart rate was recorded on a Beckman Type RS dynograph at a paper speed of 5 mm/sec. The EKG recording sites were cleaned with 700/0 ethanol prior to the application of the electrodes. Zinc cup electrodes with a surface area of 3.14 sq cm and filled with cotton soaked in a 1% ZnSO. solution were used to record heart rate from EKG lead II. The heart rate was measured using a Beckman 9857 cardiotachometer. Procedure The Ss were randomly assigned to one of two groups of 18 Ss each. The two groups were instructed to respond as rapidly as possible following the termination of the warning signal (green light) that remained illuminated for the duration of the foreperiod. Concurrent with the termination of the warning signal, the respond signal (red light) was illuminated. The respond signal was terminated by a buttonpress response. The S was seated in a comfortable armchair in a sound-attenuated room. After the pickups were attached, E read the instructions which informed S of the task and then calibrated the recording equipment. Each S then received two sessions of 10 reaction time trials, separated by a 2-min rest period. The rest period was signaled to the Ss by the illumination of a white pilot light. The fIrst session was used as practice to adapt the Ss to the experimental situation. The foreperiod duration was 16 sec for the fIxed-foreperiod group and varied among 16, 22, and 28 sec, according to a predetermined randomized schedule, for the variable PI group. Each group had the same schedule of intervals between the respond signal of one trial and the onset of the warning signal for the next trial. These mtervals varied among 45, 60, and 75 sec, according to a predetermined randomized schedule. The latency between the termination of the warning signal and the buttonpress was recorded as S's reaction time. Quantification and Analysis of the Data One minute into the rest period, 25 interbeat intervals were measured. These R-R intervals were transformed into heart rate by the cardiotachometer. Within each group the variance and the mean of the 25 beats were calculated and correlated with the mean reaction time performance of each S during the second session. The heart rate variance for each S was ranked within each group. The Ss were divided into three equal-sized within-group levels of heart rate variance, i.e., high, mid, or low. Since the heart rate variability characteristic of the three levels was not equivalent for the two groups, the levels were nested under the factor of groups in a hierarchical repeated-measures analysis of variance on reaction time performance. RESULTS Correlational Analyses Heart rate, heart rate variance, and mean reaction time were intercorrelated separately for the two groups. The only significant relationship was in the variable-foreperiod group, between the heart rate Bull. Psychon. Soc., 1973, Vol. 1 (4) 430 /j. LOW HR crt o MID HR crt o HIGH HR crt :.: 380 f/I 2 I&J ! 330 z Q I- ~ 280 I&J 0: 2 4 5 6 7 8 10 TRIALS Fig. 1. Reaction time across trials for the three levels of heart rate variability in the variable-foreperiod coJ)dition. variance and mean reaction time, r(18) = -.70, P < .005. This indicated that individuals exhibiting the larger heart rate variance during the resting period exhibited faster reaction times. This contrasted with the negligible -.03 correlation in the fIxed-foreperiod group. A comparison of these two correlation coefficients indicated that the variable-foreperiod group exhibited a signifIcantly greater concordance between heart rate variability and reaction time, z(30) = 2.4, p < .025. Also of interest was the striking similarity between the two groups on the other comparisons, i.e., heart rate vs heart rate variance [VI, r(18) = .16; FI, r(18) = .11] and heart rate vs reaction time [VI, r(18) =- .06; FI, r(18) = -.03] . Analyses of Variance A hierarchical analysis of variance with repeated m (...truncated)