PERIODIC NATURE OF SPONTANEOUS PERIPHERAL NERVOUS SYSTEM ACTIVITY

Journal/Book: Reprinted from Nature Vol. 197 No. 4873 pp. 1169-1170 March 23 1963. 1963;

Abstract: DR. A. A. ALEXANDER PROF. ROBERT ROESSLER and DR. NORMAN S. GREENFIELD Wisconsin Psychiatric Institute and Department of Psychiatry University of Wisconsin RHYTHMIC central nervous system activity and spontaneous autonomic nervous system activity are well documented1 2. The literature relating the central and the autonomic nervous systems functions also quite extensive (cf. review by Dell and Bonvallet3). Yet little is known of rhythmic activity in peripheral nervous system functions (cyclic fluctuations in variables such as heart rate or skin temperature are usually specified an the order of hours rather than minutes or seconds). It was therefore predicted that relatively high-frequency rhythmic activity could be discerned in measures mediated by peripheral physiological systems. This hypothesis was tested by means of a variance (or power) spectrum analysis of resting-levels of the psycho-galvanic response finger blood volume heart rate and electromyograms. The power spectrum is the integrated Fourier transform of the auto-correlation function of a time series4 5 and allows an estimate of the variance at each of a series of frequency intervals. 6 min of continuous data were scored at every second for each of the four measures an each of 40 subjects. Power spectra of each 360-sec series were then derived an a CDC 1604 computer following the formulae of Blackman and Tukey6. Sixty computational lags were used allowing the examination for periodicity at each of 60 frequency-intervals. The frequencies studied ranged from 0·5 c/s to 0·008 c/s. Their reciprocals or periods range from 2 sec to 120 sec. Periodicity when present was indicated by a hump in the spectrum when plotted as the proportion of the total variance per frequency interval (Fig. 1). The degrees of freedom appropriate to evaluating the significance of any such hump was 12 and based an a 2 distribution and a Gaussian assumption the value at the top of the hump had to exceed that of the bottom of the valley by a factor of 4 to differ at the 95 per cent level of confidence. ... ___MH

Zurück | Weiter