TABLE 1.

Human experimental studies of tremor suppression using non-invasive peripheral stimulation.

Study	n (Number of Patients)	Wearable/Non-wearable Hardware	Electrodes (Number Excluding Negative Electrode, Placement, Type)	Sensors (Number, Type, Sampling Rate)	Tremor Suppression	Comments
Sensory Stimulation
S. Dosen et al., 2015 [68]	6 (4 PD, 2 ET)	Non-wearable	2, point with the strongest motor response from a probing electrode; for wrist extensors this was close to lateral epicondyle and for wrist flexors, proximal to the elbow crease and medially, PALS Platinum electrodes	2, sEMG, 1kHz	35-48% (sensory stimulation in 5 patients) 46-81% (motor stimulation in 5 patients)	1 ET patient failed to have any tremor suppression. Approach: 3 second stimulation window, 1 second recording window. Iterative Hilbert Transform used for tremor demodulation.
Pahwa et al., 2019 [17]	77 ET patients (40 received treatment stimulation, 37 received sham stimulation)	Wearable	2, overlying the path of the median and radial nerves close to the wrist, 2.2 cm × 2.2 cm square hydrogel electrodes	N/A, accelerometers, N/A	46% reduction in tremor amplitude in the treatment group compared to 24% reduction in tremor amplitude in the sham group.	Approach: 40 minutes of stimulation after a short calibration to the patient’s dominant tremor frequency with tasks and tremor ratings completed immediately following this.
Isaacson et al., 2020 [53]	263 ET patients enrolled, 205 patients completed at least one stimulation session per day on 78% of the days on average.	Wearable	2, overlying the path of the median and radial nerves close to the wrist, 2.2 cm × 2.2 cm square hydrogel electrodes	N/A, accelerometers, N/A	92% of patients had an improvement ratio greater than 1 from pre- to post-session.	Approach: 40 mins stimulation twice per day over 3 months.
Kim et al., 2020 [18]	9 ET patients (1 participated in 2 separate sessions)	Wearable	1, overlying the site of the radial nerve close to the wrist, 0.8” round transcutaneous electrical nerve stimulation unit electrodes (Syrtenty)	1, 3-axis Accelerometer (LSM303D), 100 Hz	42.17% +− 3.09% reduction in tremor power.	Approach: For the closed-loop trials, active tremor periods and tremor phase were analysed in real-time to determine when to apply the nerve stimulation.
Motor Stimulation
Javidan et al., 1992[66] and Prochazka et al., 1992 [65]	24 (3 ET, 4 PD and 6 other)	Non-wearable	2, overlying biceps muscle and overlying triceps muscle, 2 cm × 3cm pre-gelled self-adhesive electrodes (Chattanooga Corp.)	1, low-noise optical displacement transducer monitoring a pivoting armrest or strain gauge, N/A	73% tremor suppression in ET patients.	Approach: Manual calibration to ensure equal flexor and extensor offset torques was done initially for each subject. A filter designed to maximize open-loop gain in the 2-5 Hz frequency range and minimize phase lag in the 0-10 Hz frequency range was used and the stimulation was applied out-of-phase to the tremorogenic muscles.
Gallego et al., 2011[50]	1	Non-wearable	N/A, array of electrodes on forearm and arm, sewn electrodes in e-textile (SMARTEX)	1, 16-channel EEG, N/A 1, 128-channel sEMG, N/A 2, IMU, N/A	60% tremor attenuation	Approach: open-loop stimulation using an electrode array to increase joint impedance.
Widjaja et al., 2011 [69]	1	Non-wearable	2, image shows electrodes on the dorsum of the forearm, N/A	2, sEMG, 1 kHz 1, 3-axis accelerometer, 1 kHz	57% tremor power reduction	Approach: real-time closed-loop ON/OFF control. An algorithm determines first if the sEMG signal is large enough to be considered tremor, then combines accelerometer and sEMG phase estimations to determine when to start the signal to ensure it is out of phase.
Popović Maneski et al., 2011 [47]	7 (some PD, some ET)	Non-wearable	2, “dorsal and volar side of the forearm over the motor points of the wrist extensors and flexors”, PALS round electrodes	2, gyroscope,1kHz	67 +/− 13% (average tremor suppression in 6 patients)	1 patient failed to have any tremor suppression. Approach: 3 second stimulation window, 1 second recording window

Note: N/A = not available, sEMG = surface electromyography.