Tremor is a motor phenomenon defined as an involuntary, rhythmic oscillation of one or more body parts. One such type of tremor is essential tremor. Its prevalence is estimated at approximately 1% of the general population, or 5% among individuals older than 65 years.
For the purposes of classification and diagnosis of tremor, all three axes from the accelerometer are usually analysed. However, the aim of this master’s thesis was to validate a new online tool for tremor analysis that is based on the analysis of only a single axis. The study included 32 recordings of patients with essential tremor. The main research question was whether Tremoroton provides results comparable to manual analysis despite its different spectral settings. More specifically, the goal was to determine whether the results of the analysis of the most pronounced axis in Tremoroton are comparable to the analysis of the same axis using a manual procedure.
Each participant’s recording was first processed in the Spike2 software, after which the signal of the most pronounced axis for each hand was uploaded to Tremoroton. Following Fourier transformation, peak frequency and amplitude of the main peak were determined separately for each hand. The obtained values were then compared with the manually calculated values of the most pronounced axis. To assess agreement between the two methods, paired statistical testscorrelation analyses, as well as ICC and Bland–Altman analyses were performed.
The results of the statistical analyses showed good agreement between how the two methods analyse accelerometer signal. The findings suggest that Tremoroton could in the future contribute to faster, simpler, and more accessible analysis, and consequently to more efficient tremor diagnostics.
|
