Essential Tremor (ET), a condition characterized by postural and kinetic tremor in the upper limbs, is one of the most prevalent movement disorders. While pharmaceutical and surgical treatment options exist, they are not ideal. Assistive devices have the potential to provide relief to patients but are largely unexplored for ET. Furthermore, prior characterizations of essential tremor have focused on endpoint tremor and provide insufficient detail for designing such a device. We propose and demonstrate a novel method for characterizing essential tremor in the 7 proximal degrees of freedom (DOF) of the upper limb in various postures. In addition, we provide a preliminary characterization in a small number of patients with mild ET. We collected data from 10 patients with ET. Subjects were instrumented with four electromagnetic sensors that recorded orientation of upper limb segments. After a calibration, each subject positioned his/her upper limb in 16 different postures for 15 seconds each. This procedure was repeated 4 times for each subject, with each repetition being considered a run. Sensor data were converted to angular kinematic data for each DOF using inverse kinematics, a practice unique to this study. These data were then analyzed in the frequency domain to calculate the power associated with the tremor in each DOF and posture. More specifically, we computed the area of the periodogram over the 4-12 Hz frequency band typically associated with ET [narrow-band area (NBA)] and over the wider frequency band from 2 Hz to the Nyquist frequency [wide-band area (WBA)]. If significant peaks were found in the 4-12 Hz band, their frequency and amplitude were reported. Mixed-model ANOVA tests were used to investigate effects of DOF, posture, run, gravity, and patient characteristics on reported measures. NBA and WBA varied significantly between DOF, being lowest in the wrist, intermediate in the shoulder, and greatest in the elbow and forearm (pronation-supination). NBA and WBA also varied significantly with posture. Only 5% of observations had significant peaks, with 49% of peaks occurring in wrist flexion-extension and 39% occurring in wrist radial-ulnar deviation. Peak frequency was quite stereotyped (5.7 Hz ± 1.3Hz). Run had no significant effects, indicating that tremor measures were consistent over the duration of the experiment. Effects of gravity and demographic factors on measures were mixed and did not present a discernible pattern. This preliminary characterization suggests that tremor may be focused in a subset of upper limb DOF, being greatest (in terms of power) in elbow flexion-extension and forearm pronation-supination, and most concentrated (with peaks at a stereotyped frequency) in wrist flexion-extension and radial-ulnar deviation. Our method of 7 DOF characterization through inverse kinematics, in conjunction with future research (isolation studies, EMG, and finger DOF) may allow for optimal tremor suppression by an orthosis.



College and Department

Ira A. Fulton College of Engineering and Technology; Mechanical Engineering



Date Submitted


Document Type





essential tremor, upper limb, motor control, inverse kinematics