Quantification of Hand Tremor in the Clinical Neurological Examination

A compensatory tracking task using a visual display and augmented auditory feedback and requiring subjects to maintain a 500-gm force on a control stick has been modified for clinical use. Integrated absolute values of the subjects' force error/second (tremor scores) have been used as a measure of steadiness, and power spectra of the force error have been used to establish the predominant tremor frequencies. Twenty parkinsonian patients, 7 age-matched controls, and 20 young controls have been evaluated, and representative tremor-time records, tremor scores, and tremor-power spectra are presented. The quantitative measures have proven sensitive to individual differences among normals, yet having a dynamic range sufficient to include the most severely afflicted patient tested. Normative values of both tremor scores and predominant tremor frequencies have been established and are significantly different from the parkinsonian values, the parkinsonian patients having a higher average tremor score [8.31 versus 3.76 (gram . seconds/second)] and a lower mean predominant tremor frequency (6.6 Hz versus 7.9 Hz). Representative tremor-power spectra have proven useful in describing physiological and parkinsonian tremors. Reevaluation over 1-3-week intervals indicates that the quantitative measures of tremor scores are reliable (r= 0.83, p < 0.01 for control subjects; r = 0.87, p < 0.01 for parkinsonian patients). In addition, the predominant tremor frequency as measured from the tremor-power spectra is found to be independent of learning effects. It is emphasized that such tests are not designed as diagnostic tools in themselves.