Analysis of selected parameters of tremor recorded by a biaxial accelerometer in patients with parkinsonian tremor, essential tremor and cerebellar tremor.

Background And Purpose: Accelerometric registration of tremor with the subsequent estimation of its parameters allows objective collection of information. The aim of the study was an analysis and visualization of the signal of the recorded tremor, including assessment of spectral components at the dominant frequency and selection of those parameters that are most useful in differentiation of tremor types.

Material And Methods: We examined 44 patients with parkinsonian tremor (PT), 39 patients with essential tremor (ET), and 13 patients with cerebellar tremor (CT), diagnosed clinically using the criteria of the Movement Disorders Society and the Tremor Investigation Group. The control group consisted of 26 healthy persons. A biaxial accelerometer mounted at the dorsal surface of the hand was used. Spectral analysis was performed. The analysis focused on determination of spectrum shape, frequency of spectral peaks, centre frequency, standard deviation of the centre frequency, harmonic index, as well as changes in frequency over time.

Results: In patients with pathological tremor (PT, ET, and CT), a single spectral peak was observed that persisted for the whole registration period. In the control group, tremor was characterized by a wide range of spectral frequencies with marked variability in time. Statistical analysis revealed significantly lower frequencies of the highest peak and of the central frequency, a significantly smaller standard deviation of the central frequency and a significantly higher harmonic index in cases of pathological tremors as compared to the control group.

Conclusions: The presented method enables distinction between pathological and physiological tremor and helps in differentiation of various types of pathological tremor.