According to [21], while the peripheral nervous system (PNS) has the capacity to regenerate itself, the central nervous system (CNS) has limited capacity of regeneration after a damage occurs. In order to make the correct connections, i) cells have to be alive or be replaced (neural or tissue transplantation), ii) the environment has to be permissive for axonal growth, iii) correct signaling to select the desired target has to be provided, and iv) axonal remyelination has to be allowed. Some studies have demonstrated that application of electrical fields can directly affect the orientation and regeneration of axons [24]. Apart from this, spinal cord prosthetic devices to promote host tissue regeneration, rewiring, and plasticity are also being developed [22]. Passive SC prostheses are usually based on artificial scaffolds impregnated with cytokines to promote the surrounding tissue regeneration. On the other hand, the use of active SC prostheses is also aimed to the treatment of pain, bowel or bladder dysfunction, spasticity management, and motor control of limbs and trunks. Regeneration is considered as the solution for the complete repair and restoration; however, it is still under investigation and some promising results have been recently published [57]. In this Phase 1 clinical study, three human subjects with SCI were transplanted with autologous mucosal olfactory cells in the injured SC segments, and were provided with post-operative neurorehabilitation. Interestingly, neurophysiological examinations showed improvement in spinal cord transmission and activity of lower extremity muscles in surgically treated patients but not in patients receiving only neurorehabilitation. These initial results are encouraging and further investigations with a larger population will elucidate the mechanism of neural recovery after such transplantation, and will help develop more targeted and effective treatment approaches.

When there is a severe paralysis due to SCI, assistive devices normally help achieving independence in activities such as grasping, driving a wheelchair, or triggering an active orthosis. In this context, neuroprostheses have attracted considerable attention, as they can serve as a potential tool for intuitive user-driven control of an assistive device. This approach is typically useful for providing intuitive control in extreme disability to provide potential command signals to operate switches, joysticks, etc., via voice, eye movement and gaze tracking, electromyography, electroneurography, electrooculography, electroencephalography, etc. Neuroprostheses typically consist of electrodes that interface with the nervous system and aim to restore lost motor, sensory, or autonomic function. To this end, neuroprostheses can stimulate specific muscles, nerves, or the spinal cord to produce function in end effectors below the level of spinal lesion. Alternatively, the intended movement-related neural activity that is recorded via one of the aforementioned approaches can serve as the control signal for an active orthosis that compensates for the loss of limb function [36]. In this context, neural prosthetic devices that decode intended motion trajectories from the motor cortex to control functional electrical stimulation (FES) devices have been developed [24]. For example, the Freehand System (NeuroControl, Cleveland, OH, USA), shown in Fig. 2, consists of an implantable neuroprosthesis intended to restore hand function in individuals with motor complete injury at the C5/C6 level [59]. It can stimulate specific muscles in the arm and hand in order to produce a useful grip and key pinch. The system consists of a surgically implanted receiver unit incorporating an eight-channel stimulator and electrode combination with an external controller and power supply/microprocessor [31]. Even though it has been shown to positively impact ADL in patients, the Freehand product has been discontinued, primarily due to issues with low market size and high production costs. Neuroprostheses can also stimulate intact peripheral nerves, providing functional restoration of various body organs [24], such as bladder function. In summary, neuroprostheses serve as compensatory functional tools in patients when voluntary movement is impossible due to paralysis. Therefore, this technology has the potential to significantly increase functional quality of life (QOL) in patients with complete SCI and paralysis, and even more so in chronic SCI.

Another recent approach that has contributed to improving SCI therapy is brain–computer interfacing (BCI). These BCIs can record movement-related neural command signals directly from the brain, via intracortically with microelectrode arrays implanted in the motor areas of the brain or extracortically with electrocorticogram or electroencephalogram electrodes. Most of the current BCI designs are based on the principle of recording neural signals and decoding the user’s movement intention to operate an assistive device. BCI has been shown to not only be useful for assistive device control ([34, 49]), but also suitable for rehabilitation due to its therapeutic effects as an adjunct with physical therapy [48, 49]. Therefore, BCIs can be used for neural biofeedback-based training that can assist in rehabilitation as well as act as control mechanisms for neuroprostheses. This dual functionality provides the advantage of using this technology for rehabilitation in patients with acute complete and incomplete SCI to help engage intrinsic neuroplastic recovery mechanisms, and in patients with chronic complete SCI and paralysis. An important caveat to consider in the clinical translation of BCI technologies is that accuracy of decoding movement-related signals has been found to be lower in patients when compared to healthy subjects; further research in developing robust BCI decoding algorithms is necessary [19, 41]. Recently, promising results have been demonstrated in augmenting powered robotic gait assistive exoskeletons with noninvasive electroencephalography (EEG)-based BCIs for direct control by SCI patient user’s movement intents [11, 32]. Such findings suggest that decoding algorithms for BCI-based therapies are improving, and future developments could significantly enhance the use of BCI technologies in rehabilitation of patients with complete SCI.

An alternative to EEG-based BCIs is the use of chronically implanted intracortical, microelectrode arrays to record motor cortical single-unit or multiunit ensemble spiking activity from a person with tetraplegia and use it to control an assistive device such as a robot arm or computer cursor. This approach generally uses the relationship between intended limb movement kinematics and neuronal spiking [39] to decode movement intention [37, 63] in contrast with EEG-based BCIs, which typically map the relationship between whole brain cortical dynamics and movement intention. The intended movement commands are then used to produce corresponding real-time movements in an external device. Ongoing clinical trials are exploring the feasibility of this approach for clinical rehabilitation of tetraplegia due to SCI and other severe movement disorders [10, 26]. Due to challenges posed by the lack of stability of single-neuronal ensemble recordings [45], other approaches have recently explored the use of larger-scale intracortical signals such as local field potentials [20, 28], multiunit activity [55], and electrocorticograms [53, 61]. We refer the reader to Chapter 7 of this book for further information on applications of BCI systems in neurorehabilitation, which further describes current approaches, issues, and applications of this technology.