Functional electrical stimulation (FES) uses electrical currents to externally activate the nerves and finally activate the muscles responsible for the movements of interest. FES applied via transcutaneous electrodes is a common rehabilitation technique for assisting grasp in patients with central nervous system lesions. In upper extremities, FES can be used as neural-prosthesis, or as a therapeutic tool. In the former, some of the missing functionality is achieved by externally controlled electrical stimulation, wherein the latter FES provides functional exercise which is proven to lead to relearning of motor control, as well as muscle strengthening, thus aiding reestablishment of some of the functionality. Technology that will be used in this study enhances the capacity of FES through the application of multi-pad electrodes and smart protocols for rehabilitation exercises. To improve the stimulation effectiveness of conventional FES, introduced technology relies on spatio-temporal distributed stimulation over the multi-pad electrodes and a battery of predefined, field-tested protocols for rehabilitation.

The participants will be split into two balanced groups (randomly assigned), and one of the groups will start with AUTO intervention followed by FUNCTION, and the other group will start with FUNCTION intervention followed by AUTO intervention. The total duration of the study will be 6 months.

Each participant will receive 28 therapy sessions (5 sessions per week approx.) in total, that are divided into four blocks of 7 sessions. During the first 7 sessions the participant will receive standard care treatment (S). This block will be followed by standard care treatment (S) and first FES intervention block with 7 sessions of protocol A (AUTO or FUNCTION). Third block will consider additional 7 sessions of standard care (S). Fourth block will include standard care (S) and 7 sessions of FES intervention with other type of treatment B (FUNCTION or AUTO).

Each of the sessions of FES will have a total duration of 45 minutes. This 45-minute session contains system donning/doffing, adjusting stimulation amplitude (AUTO and FUNCTION), defining VEs for 8 movement primitives (FUNCTION) and exercises.

During assessment sessions the investigator will evaluate the condition of the hand of the participant in different aspects. Clinical scales and data collected from a inertial sensor system will be used.

Clinicians will first of all apply the Modified Ashworth scale to different extensors and flexors of the wrist and fingers in order to determine the spasticity level.

After this they will perform the Fugl Meyer assesment for the upper-limbs to estimate the sensorimotor function of the whole upper-limb (from shoulder to hand).

Then, participants will perform the ARAT test to evaluate the arm and hand function reaching and grasping different types of objects.

Finally, the BEAGLE system will be used for measuring the active and passive ranges of motion of the wrist and fingers.

Anthropometric measures and questionnaire data will be collected. The Visual Analog Scale (VAS) will be filled by the participants, where they will indicate their level of pain before starting the therapy.

The electrode will be attached on the participants forearm always taking the head of ulna as the reference.

The Fesia Walk stimulator will be used in this study. It is a wireless stimulator with a single current source capable of emitting pulses and controlling up to 32 cathode and 8 anode pads independently. It is designed to be controlled remotely via Bluetooth communication. It is connected to the multi-pad electrode by means of a pressure connector, and the connection is secured by means of a rigid plastic socket attached to the electrode. As for battery charging, it is carried out using any generic charger with microUSB connection. The software used at the start of the study will be the NeuroResearch software. This is a multi-platform software developed in html5. It provides means to control the stimulation parameters of the Fesia Walk stimulator as well as creating and saving virtual electrodes (VEs), patterns and functions. It is a versatile software that can be used with different types of multi-field electrodes and it brings the possibility of combining and testing different VEs and patterns to perform several functions. In parallel, an adaptation of this software for its use in clinics is being developed. This new software, called NeuroClinics Grasp, will be designed for its specific use with forearm garments for hand and wrist applications. It will have a simplified user interface and a limited set of configurable parameters to increase its ease of use. This software will be introduced during the study and feedback from therapists will be taken into account to make possible further modifications on it. If the NeuroClinics Grasp software gets ready at any time during the study and therapists think that it might ease the calibration procedures of the study, then, this software will replace the NeuroResearch software. Two different multi-pad electrodes (A and B) will be used during this study. Both were specifically designed for covering the extrinsic muscles of the hand, located on the forearm, and are comprised of 8 anodes and 32 cathodes. Electrode A or B will be used, alternating them over sessions with each participant.

Then the calibration procedure will take place, which will imply defining the virtual electrodes (VE) for each of the 8 movement primitives: wrist extension, wrist flexion, index flexion, index extension, mass finger extension, mass finger flexion, thumb extension and thumb flexion.

A VE is a group of electrode pads which, when stimulated together produce a movement primitive, and in this case it will be configured with two parameters: the selection of pads (cathodes) that should compound the VE and the amplitude (mA) of the VE.

All the anodes will be active for all the VEs except those located closer than 2 cm from the selected pads (cathodes). This anode configuration will be automatic.

In the AUTO case, the pads will be preset for each VE, based on results of previous studies for achieving different movement primitives. So, only the amplitude should be tuned for each VE. This will be done increasing the amplitude in 1mA steps until patient's tolerance or until a full function is achieved, whichever happens first.

Only if there is no movement at all and the therapist considers that it is due to a bad VE pad configuration, then the therapist will be allowed to adapt the best pads for that specific participant during the first therapy session only, but relying only on the knowledge of anatomy and not by trial and error tests of different configurations. Then this VE configuration will be set for the rest of AUTO therapy sessions for that participant.

Once all the amplitudes are defined for each of the VEs or movement primitives, then the FES response questionnaire (attached in the annex) will be filled by the therapist and at the same movements will be recorded with BEAGLE. Testing each of the movement primitives with the configured parameters and classifying the motor response as none, partial or full function.

Then, the therapist will run the AUTO mode, which will consist of cyclic stimulation of the pre-configured VEs. The VEs will be activated in a pseudorandom order in series of 5 repetitions of each movement with 50% duty cycle. Number of series will be calculated based on the remaining time after amplitude calibration and questionnaire filling. The amplitude will start being 10% less than configured and it will end being 10% more than configured to compensate any fatigue that could happen during the sessions. Once the stimulation is finished, the therapist will test each of the movement primitives again (as configured before), fill the FES response questionnaire (attached in the annex) and record movements with BEAGLE, and the participant will fill the VAS scale again. Like this we will be able to check the short-term effect of the therapy in terms of pain and motor response.

During the FUNCTION therapy sessions the therapist will always be able to configure or adapt the VE pads to the participant. Moreover, instead of a pseudo-random stimulation, FES will target those movement primitives that are most impaired in the participant for generating different grasps.

In the FUNCTION case, both the pads and the amplitude should be tuned for each movement primitive and each participant. The first session there will be a pad configuration preset that will serve as a guidance for the therapists, and from then on the configuration for that specific participant will be saved. The amplitude will be set for each VE increasing the amplitude in 1mA steps until participant's tolerance or until a full function is achieved, whichever happens first.

The therapist will be allowed at the beginning of each session to adapt the best pads for that specific participant and movement primitive. This will be done by modifying the number and location of the selected pads until the target movement primitive is achieved. Then this VE configuration will be set for the rest of the session.

Once all the amplitudes and pad configurations are defined for each of the VEs or movement primitives, then the FES response questionnaire (attached in the annex) will be filled by the therapist and at the same time movement will be recorded with BEAGLE. Testing each of the movement primitives with the configured parameters and classifying the motor response as none, partial or full function.

Then, the therapist will run the FUNCTION mode, and she/he will reduce the amplitudes (below motor threshold) of the movement primitives that are part of the participant's residual voluntary movements. Like this we will focus on the movement primitives that are most impaired and we will take advantage of the residual function of the participant. Once these amplitudes are adapted, the FUNCTION mode will be turned on. In a FUNCTION, VEs will be temporally arranged to form demanding hand function (palmar and lateral grasp) or simple task (wrist extension, wrist extension...). Functions will be chosen based on the capabilities and needs of the patient. A therapist will choose individual FUNCTION and number of repetitions.

Duration of the FUNCTION will be calculated based on the remaining time after VEs calibration and questionnaire filling. The amplitude will start being 10% less than configured and it will end being 10% more than configured to compensate any fatigue that could happen during the sessions.

During the FUNCTION mode the participant will be instructed to follow the stimulation patterns to make the grasps and thus, assist the FES with movement intention and residual voluntary contractions. Targeted movement will be shown on the monitor and synchronized with stimulation.

Once the stimulation is finished, the therapist will test each of the movement primitives again (as configured before) and fill the FES response questionnaire, and the participant will fill the VAS scale again. Like this we will be able to check the short-term effect of the therapy in terms of pain and motor response.

Inclusion Criteria:

• stroke confirmed by MRI or CT scan
• time from stroke < 6 months
• age above 18
• able to comply with study protocol
• able to communicate effectively
• able to provide informed consent
• existence of a residual voluntary extension movement either on the wrist or any of the fingers (even if very weak)
• to sit in a position determined for exercises

Exclusion Criteria:

• - dependent on care, prior to stroke, for activities of daily living
• severe medical condition in any arm that precludes participation in the study
• previous injury, disease, or contracture affecting paretic or normal arm or hand
• pre-existent neurologic disease or injury
• a metal implant in the affected upper limb
• electrical life support devices
• Modified Ashworth > 3
• Finger and wrist Fugl Mayer score = 2
• current participation in another study involving physical rehabilitation of the arm.

Subacute stroke subject with initial wrist extension.