Interrater reliability of Tremor Characterization

Background: Tremor is a repetitive, oscillatory movement with multiple subtypes requiring accurate characterization for diagnosis and treatment. While frequency, amplitude, and context are routinely used to describe tremor, recent recommendations have emphasized oscillation shape and regularity. However, these descriptors lack strict operational definitions, and differences in how clinicians interpret them introduce inconsistency into tremor classification. Objective: We quantified inter-rater reliability for commonly used tremor descriptors and determined which objectively measured signal parameters: peak amplitude, peak frequency, and frequency variability best explained inconsistency in rater assessments. Methods: We developed the Tremor Waveform Morphology Questionnaire (TWMQ) and obtained standardized video recordings of 34 individuals with upper-extremity tremor performing rest, postural, and kinetic tasks, with simultaneous kinematic recordings using an inertial sensor. Movement disorder experts rated tremor presence, shape, symmetry, regularity, rhythmicity, and diagnosis. We assessed inter-rater reliability, cross-feature dependencies, and modeled the influence of objective signal parameters on each subjective rating. Results: Tremor presence and symmetry showed high agreement, whereas jerkiness, rhythmicity, amplitude stability, and clinical diagnosis showed lower reliability. Modeling revealed that peak amplitude primarily influenced judgments of tremor presence, jerkiness, and diagnosis, while frequency variability drove ratings of rhythmicity and amplitude stability. Conclusions: Morphology-based descriptors remain subjective, but disagreement systematically reflects measurable signal properties. Objective definitions based on peak amplitude and frequency variability may improve the reliability and reproducibility of tremor characterization.