We tested the repeatability of a finger tapping device with magnetic sensors to determine its reliability. This device, which was developed to assist in the diagnosis of movement disorders such as Parkinson's disease (PD) and strokes, measures the distance between the first and index fingers during finger tapping movements (opening and closing the fingers repeatedly). We evaluated three types of repeatability based on ICC (interclass correlation coefficient) and Welch's test (test for equal means in a oneway layout): repeatability when measured at different times, when using different devices, and when using different measurers. We calculated these three types for three finger tapping tasks on both hands for 21 characteristics calculated from finger tapping waveforms. Results demonstrated that the repeatability when using different devices is high regardless of the task or hand. The repeatability when measuring at different times and when using different measurers is high at some tasks, but not all. One of the finger tapping tasks (finger tapping movement with the largest amplitude and highest velocity), which is used in a conventional PD diagnosis method (UPDRS), does not have enough repeatability, while other tasks show high repeatability. Results also showed that five characteristics have the highest repeatability (ICC ≥ 0.5 or significance probability of Welch's test ≥ 5% in all tasks): “total moving distance,” “average of local minimum acceleration in opening motion,” “average of local minimum acceleration in closing motion,” “average of local maximum distance” and “average of local minimum velocity”. These results clearly demonstrate the strong repeatability of this device and lead to more precise diagnosis of movement disorders.
