Support in spinal cord injury: A focus on robotics

Introduction

The incorporation of robotics in the field of neurorehabilitation is taking place rapidly, both in research and in its clinical applications, and is presented as a very promising tool that is changing therapeutic paradigms. In the late 1980s and early 1990s, basic research findings constituted a major change in therapeutic intervention in neurorehabilitation. One of most relevant was that in experimental models with cats subjected to a spinal cord injury (SCI), the subsequent training of the locomotor function applied to them offered good results. In fact, it was shown that these cats with SCI walked effectively when placed on a treadmill with partial weight support ( ). The suggested mechanism is the activation of the basic neuronal circuitries sufficient to generate efficient stepping patterns and independent standing. Indeed, the operations underlying the elaboration of motor patterns for walking and standing are essentially achieved by the neuronal networks embedded within the lumbosacral segments of the spinal cord ( ). These findings led to the concept of spinal learning via activity-dependent plasticity. Following this concept, it was found that locomotor activity can be activated in patients with severe SCI via passive activation of the legs on a treadmill ( ). Synchronous reciprocal movements of both legs, simulating normal walking are required to activate the locomotor centers in the spinal cord. The repetitive and simultaneous activation of certain sensory and motor pathways with task-specific training can select and reinforce those spinal circuits improving the ability to perform the practiced movement successfully. Thus, functional rehabilitation (i.e., walking) had to be intensive and task-oriented. Intensive and task-oriented are the two are the pillars of the motor learning neuroplasticity-based neurorehabilitation concepts that also justify the development of robotic therapy ( ; ).

Although these interventions appear promising, in order to translate them into clinical practice in humans, a great effort is needed to standardize the assessments of the therapies applied ( ). Gait training using partial weight bearing systems on treadmills in patients who had suffered a stroke or SCI was extended in the early 1990s following motor learning principles. This therapy initially presented high costs in terms of personnel and effort, as it required the participation of at least two physiotherapists to mobilize the paralyzed lower extremities of the patient with the intention of reproducing the treadmill walking cycle ( ). The great effort that this activity demanded from the physiotherapists limited the duration of the treatment sessions. This limitation led to the idea that a robotic device could serve as an alternative to manual treatment and that such a device could cover the demands of functional training ( ). This led to the first robotic systems for walking training with weight suspension on treadmills.

These robotic assistive devices enable to start a functional and task-oriented training as soon as possible after the injury and allow an intensive application of adequate afferent feedback and a high number of repetitions of functional movements ( ).

Furthermore, the outcome of rehabilitation is better if the patient is more motivated and involved in the treatment ( ). All these without forgetting one of the most evident shortcomings of conventional systems, which is the need to incorporate sensors that provide objective variables of the patient’s condition or of the execution of the task, need to be trained. These issues are satisfactorily addressed by robotic devices. This therapy can be applied alone or in combination with other new technologies such as functional electrical stimulation (FES) or virtual reality.

Robotic therapy has experienced a huge boom in the last 15 years. In fact, different clinical guidelines approved its use as a complementary element to conventional therapy in the rehabilitation of patients with upper limb deficits after suffering a stroke ( ). Robotic devices are appropriately adapted to the need to assist limb movements based on their ability to perform simple, repetitive tasks in a consistent manner that facilitates functional recovery and adaptive plasticity ( ). There are two main categories: distal end effector devices and exoskeleton-type devices. Distal end effectors were the first to appear and are characterized by the fact that they use a single distal point of contact to guide the movement of the entire limb. In the upper extremity, it can make contact in the hand or forearm, facilitating the movements of the elbow and shoulder. They produce combined movements being difficult to isolate pure simple movements. The operation of exoskeletons is different. They are structures located in parallel to the different parts of the extremities with more than one point of interaction with the person. They provide direct control over each segment of the limb by incorporating individualized motors, also called actuators, which coincide with the anatomical axis of each joint. Thus, each actuator triggers the movement of each joint on which it is located. The design of exoskeletons seems to be more suitable than that of distal effector systems to achieve large joint paths ( ).

In this chapter, we will focus on upper limb robots, stationary and ambulatory lower limb exoskeletons.

Upper limb robots

Cervical SCI can result in partial or complete tetraplegia. Each small improvement in motor control of the upper extremity can translate to significant ameliorations in function and increases independence for the individual. As mentioned above, this type of therapy offers new possibilities in the rehabilitation not only for the lower limbs but also for the upper limbs. The robotic devices allow the application of high-intensity sessions during longer periods of time, remaining invariable certain physical parameters such as speed, strength, or precision ( ; ). There is evidence that suggests task-based therapy specifically designed to deal with lost abilities produce better results than resistance strengthening exercises ( ). This task should be performed by the patient as far as possible. That’s why the devices should be equipped with a controller that provides the least assistance needed to accomplish the movement (assist as needed) and reproducible treatment protocols.

Some studies point out that by focusing the improvement of robotic therapy more on the proximal recovery of the upper limb (shoulder and elbow), it does not translate into improvement of the functional ability that depends on hand control. However, the best results seem to be found by adding the application of both types of therapies ( ). Despite the low number of studies, results from these studies suggest that robotic training protocols are feasible and well tolerated and have a positive impact on improving arm and hand functions in selected patients with cervical SCI, but the results must be interpreted with caution ( ). In any case, studies with larger samples are needed, especially those that analyze the distal region of the upper limb, in order to have solid conclusions about the effectiveness of these devices.

Most of the current devices include a virtual reality module with visual or haptic feedback to improve sensory feedback, as well as patient motivation and engagement. They also have the capability to obtain movement kinematics that can provide precise information about movement quality that otherwise is not included in functional assessments ( ).

Although there is a number of different robotic devices currently used for neurorehabilitation of the upper extremities following SCI ( Fig. 1 ), we will now focus on the most commonly used: