Proceedings of The Physiological Society

University of Manchester (2010) Proc Physiol Soc 19, PC234

Poster Communications

Human physiological tremor has a mechanical basis

M. Lakie1, T. M. Osborne1, R. F. Reynolds1

1. Sport and Exercise Sciences, Birmingham, Birmingham, United Kingdom.


  • Fig1. The gain between rectified extensor EMG and tremor acceleration. There are 3 postural conditions (Ptop, Pmid and Pbot) and 3 dynamic conditions (D0.1, 0.2 and 0.5 Hz). The gain is sharply peaked at the tremor frequency. Gain is approximately an order of magnitude higher under dynamic conditions and peaks at a lower frequency. Values are the means of data from 9 subjects. ANOVA and post hoc comparison (Tukey) shows that all means are significantly different at a level of at least p<0.005

The predominant frequency of human tremor was first accurately noted by Schäfer (1886): “..about 10 per second may be taken as the average”. Much effort has gone into explaining this rhythm as a centrally driven (Elble, 1996), or reflexly driven (Lippold, 1970)neurogenic phenomenon. The part played by the mechanical resonant properties of the limb, tendon and muscles has received relatively little attention (Lakie et al 1986). With ethical permission we have made an extensive study of tremor acceleration, EMG and their relationship in over 30 subjects under postural and dynamic (slow movement) conditions. Here we describe results for the middle finger. In postural conditions the rectified extensor EMG frequency spectrum is relatively flat, the acceleration frequency spectrum is sharply peaked between 6 - 12 Hz and the gain is consequently sharply peaked where the acceleration size is maximal. Under conditions of slow movement (cf Vallbo and Wessberg (1993) the gain becomes an order of magnitude higher and the frequency at which the acceleration and gain peaks becomes lower by approximately 2 Hz (Fig 1). There is no change in the EMG except that low frequency modulation is observed which drives the slow movement. These are the characteristics of a lightly damped mechanically resonating system. The frequency will be determined by the limb inertia and the stiffness of the muscles and tendons. Under conditions of movement, stiffness and damping reduce. Accordingly the resonant peak occurs at a lower frequency and the gain becomes generally higher. We conclude that the most parsimonious explanation of postural and dynamic physiological tremor is that it is a consequence of broad band forcing of a resonant system. Surprisingly, the necessary gain calculation to show this has not previously been made. There is no special relationship between the EMG and the tremor, and such coherence as there is at the tremor frequency may well result from the movement modulating the EMG rather than the converse. These findings probably explain why the only way of altering tremor frequency is by mechanical means (altering the load characteristics) and why the search for physiological central tremor generators has proved unsuccessful.

Where applicable, experiments conform with Society ethical requirements