Background and purpose: There are several instrumental and clinical methods to assess hand tremor. The clinical methods, e.g. rating scales, have the advantage that they are available to most clinicians; however, they require experience, and are not as repeatable as instrumental methods. The study describes the use of a method based on a digitizing tablet and artificial neuronal networks in the assessment of tremor. The Automated Computer Tremor Score (ACTS) is based on spiral drawings on a graphic digitizing tablet. The aim of the study was to evaluate a new method and compare it with the standardized methods of tremor assessment.
Material and methods: A hundred and one patients with idiopathic Parkinson's disease (IPD) and 52 patients with essential tremor (ET) were examined. All subjects were asked to draw an Archimedes spiral on the graphic tablet. The drawn spirals were evaluated using ACTS and clinically by three independent raters according to a ten-point scale. Tremor was additionally assessed using the volumetric method. The Automated Computer Tremor Score correlated considerably with tremor rates provided by every rater (r=0.68 vs. r=0.76, p<0.0001), and with measures obtained using the volumetric method (r=0.63, p=0.01 and r=0.56, p= 0.03). ACTS also correlated with ADL score among ET patients (r=0.56, p=0.0004).
Conclusions: The study shows that neuronal networks may be taught to rate tremor severity analogically to human rating and automated scoring may be a useful method in clinical practice.