A mechanical linkage device was used to measure the three-dimensional posit
ion of the fingertip during a postural task. Thirty patients with essential
tremor were tested simultaneously with the device, uniaxial accelerometry.
and clinical tremor measures. Eighteen patients were tested again 16 +/- 4
days later. The device accurately recorded the three-dimensional behavior
of essential tremor. Measures from the device included mean three-dimension
al velocity, mean three-dimensional dispersion, and power of the three-dime
nsional acceleration. The logarithms of these measures were strongly correl
ated (r = .841-.984) with all clinical measures including self-reported tre
mor disability. The device measures were reliable within and between testin
g sessions (intraclass correlation coefficients = .971-.977). The performan
ce of the device was superior to uniaxial accelerometry, most likely as a r
esult of the three-dimensional nature of the measurements. We conclude that
essential tremor can be validly and reliably quantified during a postural
task providing the recording device records movement in three dimensions an
d the measurements are logarithmically transformed.