Vibration perception in normal man and medical patients.

Interest in sensitivity to vibration as an aid to neurological diagnosis began in 1889 when Rumpf tested this function in a case of syringomyelia (Fox and Klemperer, 1942). Subsequent investigators found altered vibratory sensitivity in various disorders of the nervous system (e.g., Williamson, 1922). Reduced sensitivity, for example, has been reported in peripheral neuritis, diabetes mellitus, and pernicious anaemia (e.g., Cosh, 1953), while hypersensitivity has been seen in the Parkinsonian syndrome (Gordon, 1945). 'Changes in vibratory sensitivity have been used as positive or negative diagnostic indicators (e.g., Canelas, 1958). Unfortunately, efforts to establish a simple, standard clinical routine for testing vibratory sensitivity have been plagued by a variety of problems. The first one concerns age and sex differences. Although vibratory sensitivity diminishes with ageing (e.g., Newman and Corbin, 1936; Laidlaw and Hamilton, 1937), few clinical investigations have allowed for this effect. This is quite surprising when one considers the high mean age of patients displaying certain clinical neurological symptoms. Possible sex differences also have been investigated (Mirsky, Futterman, and Broh-Kahn, 1953; Steiness, 1957; Detre, Feldman, Rosner, and Ferriter, 1962), but the results are inconclusive. Many approaches to clinical testing of vibration also have suffered procedural difficulties, such as inadequate control of the stimulus or of the routine (see discussion in Steiness, 1957). In earlier investigations, furthermore, thresholds were measured at one. or, at most, a few frequencies. Since vibratory thresholds are a U-shaped function of frequency (Goff, 1959), measurements at one frequency give much less information than is necessary to assess vibratory functions. Attempts to secure more clinically sensitive tests of vibratory sensitivity have prompted introduction of such techniques as the use of a pneumatic cuff (Steiness, 1959), the in-