Personalized upper limb training combined with anodal-tDCS for sensorimotor recovery in spastic hemiparesis: study protocol for a randomized controlled trial

In the personalized practice group, the Jintronix system will be calibrated prior to each training session according to the individual’s ability to reach within the active zone only, whereas in the non-personalized practice group the calibration will be done based on the individual’s full active ROM. In this way, the calibration procedure is used to set and progress the sagittal reaching distance in each game from near to far according to the length of the subject’s effective sagittal reaching distance. Success in each reaching task depends on a combination of reaching distance, speed and precision according to Fitts’ law [24]. Thus, difficulty levels of each activity will be increased by altering the location of the virtual targets and by decreasing the allowable movement time to complete each reach (i.e., by increasing movement speed), while target size remains the same. Levels will be progressed when the participant reaches at least 95% success rate in three rounds. Game difficulty will be progressed according to the Challenge Point Theory of motor learning [25]. This theory suggests that learning is enhanced by optimally challenging the individual through manipulation of task difficulty according to their motor skill level and cognitive capacity. Since learning is also affected by performance-related feedback, participants will receive concomitant visual and auditory Knowledge of Results about task success when movements are made accurately and within a specified lapse of time programmed by the therapist. Participants will also receive negative Knowledge of Performance about the use of compensatory trunk movements during reaching through the Jintronix system. Thus, if the participant leans their trunk more than 5° in the sagittal direction, a yellow square will appear on the screen and a reach success will not be registered. The 5° trunk-flexion range was chosen since it corresponds to the cutoff value identified in Subramanian et al. [26] as indicating compensatory trunk movement. Regardless of the training zone and difficulty level, all participants in each group will follow the same training guidelines and complete the 50 min of active movement time.

A-tDCS (Soterix, New York, NY, USA) will be provided by a constant direct current stimulator at an intensity of 1.5 mA, applied via saline-soaked surface electrodes (5 cm × 7 cm) on the scalp. Stimulation will occur during the initial 30 min of the 50-min training session. The anode will be placed over the sensorimotor areas of the affected hemisphere (placed over C3/C4 using the electroencephalogram (EEG) 10–20 referencing system), while the cathode will be placed on the contralateral supraorbital area. For the sham stimulation, the same electrode placement will be followed, but the device pre-programmed sham setting will apply stimulation only for the first minute. In the first 30 s, a stimulation ramp-up is expected to cause a mild itching sensation underneath the anodal electrode. This sensation is expected to fade between 40 and 60 s, at the end of which the intensity reaches zero level. No stimulation will be provided for the rest of the practice session.

As this trial intervention is carried out in different centers, the adherence to the intervention will be ensured by recording the frequency and the intensity of the games on standardized documents, supervision by the lead investigator at each site and supervision by the co-principal investigators (PIs) via regular meetings.

Since active elbow extension (active control zone) is limited by the location of the ST in the angular range of the elbow, the ST angle of the elbow flexors (biceps brachii, BB) will be measured. This will allow us to determine the range of angles in which BB cannot be relaxed (spasticity zone) and the range in which reciprocal agonist/antagonist muscle activation is possible (active control zone). The ST angle will be defined using the MSRT, a portable clinical device consisting of a two-channel electromyography (EMG) system (Procomp 5, Thought Technology, Montreal, QC, Canada), an electrogoniometer (servo-type rotational-position potentiometer P2200; Novotechnik U.S. Inc., Southborough, MA, USA) and dedicated software implemented on a laptop computer (Fig. 4). The MSRT has moderate-to-good intra- and inter-evaluator reliability for the measurement of elbow spasticity in post-stroke participants [23].

BB and triceps brachii (TB) EMG recordings will be obtained with surface electrodes (Ambu® Blue Sensor P, Ballerup, Denmark), from motor points indicated in the Surface ElectroMyoGraphy for the Non-Invasive Assessment of Muscles (SENIAM) guidelines [28]. EMG signals will be band-pass filtered (10–1000 Hz), amplified (×500) and sampled at 2048 samples/s. Seated participants will perform a maximal voluntary contraction of the elbow flexors to adjust EMG gain and initial elbow angle will be specified. Then, 20 stretches will be performed manually by stretching the elbow from flexion (approximately 50^o) to full extension. Participants will be instructed to relax completely without assisting or resisting the angular displacement. Adequate rest (10 s) will be allowed between stretches to minimize effects of fatigue and muscle thixotropy [29]. Before each stretch, a randomly assigned speed, equally distributed between slow, moderate and fast velocities, will be indicated by an auditory signal from the MSRT to avoid the participant anticipating the upcoming stretch. Each evaluation will last for approximately 20 min. At the end of each stretch, MSRT identifies the angle at which the EMG begins, called the dynamic stretch reflex threshold (DSRT) [23]. At the completion of at least 20 correct stretches, a linear regression line is computed through the DSRTs on the angular velocity/displacement plot from which the correlation coefficient (r²), slope and x-axis intercept are determined. The x-axis intercept corresponds to the tonic stretch reflex threshold (TSRT) angle. A low TSRT angle corresponds to a high level of spasticity.

Voluntary arm motor ability and coordination will be evaluated using the valid and reliable Fugl-Meyer Assessment (FMA) [30‐32] that evaluates reflexes, volitional movements and rapid alternating movements. UL scores range from 0 to 66. Sensation (light touch), kinesthesia and passive ROM will be evaluated using FMA scales.

Clinical spasticity in elbow flexors and extensors will be measured with the MAS [18, 19] on six-point ordinal scales (0–1, 1 +, 2–4). The MAS has poor-to-good inter-rater reliability [19, 33, 34].

Arm functional ability will be assessed with the six-item Streamlined Wolf Motor Function Test (S-WMFT) consisting of (1) speed of completion and (2) movement quality scores. Items 1 and 2 are activities involving elbow and shoulder movements and items 3 to 6 are functional tasks. Each item has a maximal time of 120 s and the mean will be reported. In the case of inability to complete a task, a time score of 121 s will be assigned [35]. Functional ability is rated on a six-point scale from 0 to 5 where 0 indicates no attempt to move the affected UL and 5 indicates normal movement. Mean scores will be reported. The S-WMFT has good concurrent (Spearman’s ρ = 0.69) and predictive validity (Spearman’s ρ = 0.68) with FMA. Responsiveness, as assessed by the standard response mean, was 0.41 [36].

A kinematic assessment of a common daily reach-to-grasp task requiring different amounts of elbow extension will be recorded with a room- and body-calibrated five-sensor wireless electromagnetic tracking system G4 (Polhemus, Colchester, VT, USA; RMS static accuracy 0.20 cm for position and 0.5° for orientation at a distance of 1 m from the transmitter). Each sensor is tracked at a rate of 120 Hz providing six degrees of freedom (three Cartesian and three polar). Sensors will be placed on the index finger metacarpo-phalangeal joint, the proximal third of the dorsal forearm, the mid-lateral surface of the arm, the mid-point of the superior-lateral border of the acromion, and mid-sternum.

Participants will sit in front of a table on a standard chair with no armrests. Arm position will be adjusted so that in the initial position the elbow will be in 30° flexion alongside the body and feet will be supported on the floor. Participants will be instructed to reach to grasp a 6-cm diameter cone “as fast and as precisely as possible,” hold or touch the cone for 2 s, lift the cone or arm towards the chin and then return it to the table. A mid-sagittal reference frame will define four target locations. Targets 1 and 2 will be placed at two thirds and full arm’s length in the mid-sagittal plane, respectively, for which different levels of elbow extension are required. Targets 3 and 4 will be placed at approximately 20 cm to the right and left of Target 2. Arm length will be measured from the medial axillary border to the distal wrist crease with the elbow extended [37]. After several initial practice trials (two trials/target), participants will perform two sets of 40 trials (2 × 10 trials × 4 targets = 80 movements; randomized), in approximately 1 h. Rest between sets and trials will be provided.

Motor behavior will be measured at two levels: endpoint performance and arm movement quality [38]. Endpoint performance includes peak speed; trajectory straightness (index of curvature, IC) [39] where IC > 1.57 describes a semi-circle; and smoothness calculated as the number of velocity peaks in the tangential velocity profile.

For movement quality, joint angles involved in reaching will be measured including shoulder flexion, abduction/adduction and rotation, elbow extension, wrist supination and extension using Euler angles based on standard 3D kinematic reconstruction [40]. Analysis of inter-joint coordination, compensations and pathological synergies will include consideration of trunk and arm plane motion [41].

In a secondary analysis, the relationship between specific lesion type, size and location and treatment outcomes will be assessed.