Types of nonwearable assistive robotic devices.

There is a variety of nonwearable assistive robotic devices. Some of these nonwearable assistive robotic devices are summarized in Box 38-10.^35–43 This group of robotic devices primarily includes powered wheelchairs with autonomous intelligence, body-weight–supported mobile walking aids, robots for body support with indoor and outdoor navigation, hands-off service robotic devices, and body-weight–supported treadmill systems (BWSTSs) with and without robotics.^37,41–45

Purpose of nonwearable assistive robotic devices.

Nonwearable assistive robotic devices were designed primarily to facilitate mobility and gait training. More specifically, body-weight–supported gait training systems were initially designed to unweight the body, decrease ground reaction forces (GRFs), and protect against falling (Figure 38-3). Under these assisted conditions, it is easier for patients to achieve intense levels of exercise such as walking, skipping, and running with less pain and less trauma to the joints.⁴⁶ These systems also allow runners to exceed the speed of overground running while being protected from falling.⁴⁷ Unweighting also allows patients to increase their heart rate more slowly while running or jogging at a higher speed.⁴⁷ For patients with spinal cord injuries, body unweighting over a treadmill was a translation of basic science findings to clinical practice. In animal studies, walking ability could be restored after induced spinal cord injury by facilitating the spinal generator for stepping through movement of the treadmill belt.^48–50 Through unweighting, it is also possible to maintain metabolic activity while allowing healing and recovery after joint inflammation, muscle injuries, degenerative joint conditions, bone fractures, surgical repairs, joint replacements, osteoporosis, stroke, or head trauma. Most of the research on bone density is based on studies involving patients with spinal cord injuries. In these studies, walking on a BWSTS does not significantly increase bone density. However, those who walk on the BWSTS lose less bone density than those who do not exercise on the treadmill.^51,52 On the other hand, although body-weight–supported treadmill training (BWSTT) may not necessarily increase bone density, particularly in patients with spinal cord injuries, exercise (with or without BWSTT) can increase muscle mass and/or prevent atrophy. On the other hand, in one study the use of shorter, frequent mechanical loading sessions was correlated with enhanced bone mass.⁵³ More research is needed to determine the time and frequency of standing and walking exercises and their beneficial effects on bone density.⁵⁴

Key features of unweighting.

A key feature of BWSTT is the degree, comfort, and convenience of adjustable support.⁵⁵ Whereas harness systems are the most commonly used for unweighting, when high levels of body support are needed or patients are jogging, these harnesses can be uncomfortable and cause pressure chaffing. This has led to the development of new body-weight–support systems such as the AlterG trainer (Figure 38-4), which uses a lower-body air distribution system, or the GlideTrak and Glide Cycle, which use a suspended bicycle seat–type unweighting system (Figure 38-5). The movement of the treadmill support surface provides an automatic stretch stimulus for the individual to step. In addition, the speed can be modified to stimulate faster or slower lower-extremity movements, which is very beneficial for individuals who need to vary the rate and responsiveness of their motor movements, such as clients with Parkinson disease, aging adults, or individuals with balance impairments. See Table 38-1 for a summary of the characteristics of some of the current unweighting systems.

Effectiveness of body-unweighting treadmill training.

Various biomechanical studies have been done to confirm the parameters of unweighting. After use of early lower-body positive pressure support (LBPPS) prototypes, Grabowski^46,47 reported that GRFs were reduced while metabolic power demands were maintained during running. Individuals could achieve faster running speeds when under conditions of unweighting. With slow-speed running on the LBPPS (1.0 to 1.5 mph), it is possible to reduce GRFs by 50% when the body is unweighted to 27% to 48% of normal body weight. At faster speeds (3 to 5 m/s; 6.6 to 11 mph), individuals must be unweighted to 25% of their body weight to maintain a GRF similar to that during walking over ground.

During progressive levels of unweighting, Ruckstuhl⁵⁵ compared cardiorespiratory performance, gait parameters, and comfort between a laboratory prototype device similar to the LBPPS AlterG trainer and a traditional harness BWSTS. Subjects reported significantly greater comfort using the AlterG trainer and greater tolerance for high unloading. Subjects had significantly lower heart rates when training on the LBPPS versus the harness-based body support system. Other clinical studies have been carried out to confirm the benefits of body unweighting for patients with neurological problems (Box 38-11).

The outcomes studies on BWSTT (in patients after stroke, after spinal cord injury, and with Parkinson disease) have generally positive findings. BWSTT enhances gait speed, endurance, and potentially quality.56–58 It is important to force patients to walk at community-level velocities of at least 0.8 m/s (faster than 2.0 mph).^59–63 With forced intensity, patients are more likely to achieve physiological and possibly neuroprotective benefits.^64,65 For those who make the greatest gains, the improvement has been noted as increased single-limb support, particularly on the most affected limb of individuals after stroke.⁶⁶ Although there are positive mobility benefits of BWSTT for patients with spinal cord injuries,^{58,59,67–73} guidelines still have not been generated regarding the best parameters for training (e.g., speed, amount of unweighting, time of intervention). Furthermore, one randomized clinical trial found that intense task-specific gait training over ground was equally as effective as intense body-weight–supported gait training for patients after stroke and after spinal cord injury.⁷⁴ In another study, BWSTT produced similar gains in mobility compared with locomotor training at home.⁵⁶ However, there were more falls in those patients who were started on locomotor training early rather than late. In addition, although the outcomes of improved gait performance are similar for BWSTT and aggressive bracing-assisted walking over ground for patients poststroke, the patients with the most severe impairments made the greatest gains with BWSTT.⁷⁵ Other researchers have reported statistically significant gains in balance, balance confidence, and quality of life after chronic poststroke patients were trained on a BWSTS. However, over the long term the gains were not necessarily considered to meet the criteria for minimal detectable change.⁷⁶ It appears that the gains in speed after BWSTT should be 0.16 m/s to achieve a minimal clinically important difference.⁷⁷ Furthermore, there may be additional training activities that can be done to increase the effectiveness of BWSTT, such as VR training,⁷⁸ electrical stimulation,⁷⁹ constraint-induced therapy and/or robotics,⁷⁰ or dual-task learning-based training.⁸⁰

One BWSTS device allows the individual to walk overground while having one arm attached to the apparatus. The KineAssist gait trainer–destabilizer is an example of this type of body-supported gait trainer (Figure 38-6). Its primary function is to retrain balance to prevent falling. It is considered a BWSTS because the attached arm perturbs the individual while walking and can also provide some support. It perturbs the individual to fall and then catches the individual before he or she falls to the ground.⁴⁰

General issues of wearable robots.

Wearable assistive robots are person oriented. The distinctive aspect of the wearable robot is that the exoskeleton is mapped onto the body or limb(s) with touch, pressure, visual, auditory, or movement sensors. The exoskeleton is worn by the individual, but the internal or external interface must be mapped to the anatomy, a cognitive control mechanism, or the brain. Wearable assistive robotic technology extends, complements, empowers, replaces, or enhances the function or capability of the individual. Ideally, the wearable orthotic moves with the individual in open space. However, some of these devices are in the developmental phase and are temporarily tethered to electronic systems or workstations.

Classification of wearable assistive robotic devices.

Wearable assistive robots can be classified in a number of ways. Most simply, they can be classified by body part, such as the upper limb, lower limb, or trunk. They can also be classified by how they are powered (e.g., static orthotics for stability, prosthetic robots, or empowered exoskeletons for task performance or mobility for gait training). Empowered exoskeletons, called extenders, are controlled by the individual and are usually designed to enhance the performance of the subject beyond natural ability. This is a type of master slave. These robotic devices usually require an external force. Most often, these master slave robotics are used in space and certain types of research rather than rehabilitation and will not be elaborated on in this chapter.

In designing a wearable robotic device, biological principles must be followed. The wearable robot must be adaptable with minimal weight, muscle fatigue, and energy consumption. It should be compact, miniaturized, portable, and energy efficient. In addition, there should be minimal damage imposed on the tissues. Redundancy of the joints and degrees of freedom allow a variety of ways patients can perform a task. However, these redundancies may be hard to control. Wearable robotic devices will vary by type of human interface system incorporated (e.g., brain neurons, cognitive, sensory [tactile, pressure], physical [movement], or breath). At a minimum, the physical interface of a wearable robot requires an actuator and a rigid structure to transmit forces to the neuromusculoskeletal system.

Some common types of wearable assistive devices are summarized in Box 38-12^{10,11,82–114} for the upper extremities (Figure 38-7) and Box 38-13^{10,11,35,42,115–121} for the lower extremities (Figure 38-8). In general, they are divided into three categories of wearable robots: single-task robots, workstation robots, and wheelchair-mounted manipulators (see Figure 38-7). They can also be classified by whether they use neuromuscular stimulation or biofeedback to assist the movement (Figure 38-9). Some patients use these wearable assistive robotics for training and recovery. Those who have reached a plateau in recovery may continue to use these assistive robotic devices for ongoing functional assistance and safety.

In 2012, there were a variety of wearable upper-limb robotic devices on the market, and most were unilateral (see Box 38-12). At the same time, there were fewer lower-limb wearable robotic devices on the market; however, a variety of lower-limb exoskeletons are under development^10,11,35,94 (see Box 38-13).

Wearable assistive robotic devices with neuromuscular stimulators.

Wearable assistive robotics with neuromusculoskeletal stimulation are available for both the upper and lower limbs. Ideally, both muscle stimulation and active voluntary muscle contractions are used to increase functional use. These robotic devices may facilitate muscle contractions in one or more directions around one joint such as the elbow, wrist, or ankle.

The primary impairment that can be addressed with neuromuscular stimulation is muscle weakness, but the patient may also have neuromotor control problems or abnormal synergies of movement. These devices work only when peripheral nervous system function is preserved. However, in the case of patients with spinal cord injuries, multiple electrodes may need to be used to sequentially activate a series of muscle contractions to enable walking. Similar robotic devices are available for patients poststroke. The electrodes are placed on the skin within the wearable robotic device and programmed for stimulation during specific movements.

Effectiveness of wearable assistive robotic devices.

The effectiveness of wearable assistive robotic devices is promising. Assistive robotic devices for the lower limb have been positively associated with short- and long-term gains in gait speed, quality of gait, endurance, time in single leg stance, toe clearance, and balance.^{35,115,119,120} There is also positive evidence supporting the effectiveness of training with a neuroprosthetic ankle-foot gait trainer, with short- and long-term gains reported in gait speed, endurance, safety, and quality.^122–126

For upper-limb assistive robotic training, there are positive trends for improving motor control, particularly in patients poststroke.^{84,96,105,106,127–143} Unfortunately, based on a systematic review of upper-limb robotics,^132,133 improved function in task performance and object manipulation was not necessarily correlated with improved motor control. The gains in motor control appear to be greatest when one-on-one care is supplemented with robotic training.

Research studies also report improved motor control with the integration of a neuroprostheses in functional arm training (e.g., includes neuromuscular stimulation or biofeedback).^82,144–146 However, although there are greater gains in motor control, these gains are not necessarily associated with improved function.¹⁴⁶

For both upper- and lower-limb neuroprostheses, greater gains are made by patients with the most spontaneous recovery before robotic training. In addition, earlier training appears to be associated with greater gains in motor control. Training needs to be repetitive and intense (several hours a week) for a period of time (e.g., 2 to 8 weeks). Unfortunately, some of the robotic devices for the upper limb do not have an end effector, limiting the gains in object manipulation.

For both upper- and lower-limb assistive wearable robotics, more research is needed in a variety of areas: comparing the effectiveness of unilateral versus bilateral training; functional benefits of wearable exoskeletons for the upper limb that have an end effector to allow better object manipulation; the benefits of one-on-one therapy supplemented with robotic training compared with one-on-one therapy or robotic training alone; timing of the robotic intervention; and intensity of training. It will also be necessary to determine whether a device should be leased for temporary use or purchased for long-term use.