Stroke: Hand Rehabilitation
Top Contributors -
Introduction[edit | edit source]
Stroke is defined by the WHO, World Health Organization, as a syndrome characterized by the sudden onset of focal or global neurological deficits (coma), lasting more than 24 hours, or with a lethal outcome, not attributable to any other apparent cause except for cerebral vasculopathy.
The Ministry of Health distinguishes three types of stroke :
- Ischemic stroke: occurs when the cerebral arteries are blocked by the gradual formation of an atherosclerotic plaque an /or a blood clot, which comes from the heart or from another vascular area (thrombo-embolic stroke). About 87% of all strokes are ischemic.
- Hemorrhagic stroke: occurs when a brain artery ruptures, thus causing non-traumatic intracerebral hemorrhage (this form represents 13% of all strokes) or characterized by the presence of blood in the sub-arachnoid space (this form accounts for about 3% of all strokes). Hypertension is almost always the cause of this very serious form of stroke.
- Transient ischemic attack or TIA, differs from ischemic stroke due to the shorter duration of the symptoms (less than 24 hours, although in most cases the TIA lasts a few minutes, from 5 to 30 minutes).
It is estimated that 40% of people who have a TIA will experience a full-blown stroke in the future. According to the 7th edition of the Italian Guidelines on Cerebral Stroke (2012) of the Stroke Prevention and Educational Awareness Diffusion (SPREAD), stroke is the third leading cause of death in Italy after cardiovascular disease and neoplasms, causing 10% -12% of all deaths per year, and is the main cause of disability. The prevalence rate of stroke in the Italian elderly population (aged 65-84 years) is 6.5%, higher in men (7.4%) than in women (5.9%).
The incidence of stroke increases progressively with age, reaching its maximum value in the over eighty-five year olds. 75% of strokes are found in people over 65 years of age. Ischemic stroke represents the most frequent form of stroke (about 80%), while intraparenchymal haemorrhages concern 15% -20% and subarachnoid haemorrhages about 3%. Ischemic stroke affects people with an average age of over 70 years, more often men than women; intraparenchymal haemorrhage affects slightly less elderly subjects, always with a slight prevalence for males; sub arachnoid haemorrhage most often affects female subjects, with an average age of about 50 years.
Every year, approximately 196,000 strokes occur in Italy (data extrapolated from the 2001 population), of which 80% are new episodes (157,000) and 20% relapses, affecting previously affected subjects (39,000). It is estimated that demographic evolution will lead, in Italy, if the incidence remains constant, to an increase in stroke cases in the near future. The number of subjects who have had a stroke (2001 population data) and survived it, with more or less disabling outcomes, can be calculated in Italy as about 930,000. In the world, the number of deaths from strokes is expected to double by 2020. Acute mortality (30 days) after stroke is approximately 20% -25% while that per year amounts to approximately 30% -40%. One year after the acute event, about one third of stroke survivors, regardless of whether it is ischemic or hemorrhagic, has a high degree of disability, which makes them totally dependent.
After a stroke, it is estimated that about 20% of subjects do not recover the use of the upper limb and that most of the subjects (from 65 to 85%) undergo a partial recovery. Loss of functionality in the use of the upper limb is one of the main factors affecting disability, an independent predictor of quality of life. The impairment in the upper limb is in fact one of the first obstacles to the recovery of competence in the management of ADL activities. This is why people who survive stroke perceive the loss of functionality of the upper limb as one of their major problems and it follows that it is essential to promptly implement regular and intense treatment aimed at its recovery. A 2008 review by Oujamaa et al. and more recent studies by Pomeroy et al., describe various techniques for treating the upper limb damaged by stroke. Some are specific for the rehabilitation of the upper district and the hand, while others are more generalized.
Conventional Techniques[edit | edit source]
Mental Imagery[edit | edit source]
Mental imagery is defined as the conscious representation of an action and is based on the activation, below the perception threshold, of the motor neurons of the nervous system. It has been shown by numerous neuroimaging studies that seeing oneself mentally performing a task (visual image), or above all, imagining the somatosensory consequences of a movement (motor image), activates specific areas of the cortex, the same ones that would be activated if the movement was actually performed. Through these techniques, internal visual and somatosensory inputs can be created that allow for a sort of movement planning, without execution. There are several studies that show that adding mental exercise to the practice can have positive effects on improving the motor performance of the upper limb.
Tactile Stimulation, Soft Tissue Mobilisation and Passive Movements[edit | edit source]
Tactile stimulation and soft tissue mobilization are techniques that provide sensory stimuli that attract the patient's attention and direct him towards the affected side. Passive mobilization has similar effects but in addition it induces the patient to create sensory patterns of movement.
Action Observation - Mirror Therapy[edit | edit source]
Action observation, mirror therapyt-hrough the observation of specific movements or tasks performed by the physiotherapist or by the unaffected limb reflected in a mirror positioned at the level of the midline of the body, the patient receives visual movement inputs that activate certain circuits nervous.
Repetitive Transcranial Magnetic Stimulation[edit | edit source]
Repetitive transcranial magnetic stimulation (rTMS a non-invasive neuromodulation technique capable of generating changes in the excitability of the motor cortex that persist even at the end of the stimulation and which are dependent on the stimulation frequency. Low frequency (1 Hz) magnetic stimuli that inhibit excitability or high frequency (10-50 Hz) stimuli that increase the excitability of the underlying cortex can be released on the scalp. For example, it can be used to inhibit the contralesional cortex and to stimulate the ipsilesional one. However, it should be emphasized that rTMS is a methodology with poor prospects for large-scale clinical use in patients due to the high cost of the equipment and the need for use in highly specialized clinical centers.
Transcranial Direct Current Stimulation[edit | edit source]
Transcranial direct current stimulation (TDCs) is another non-invasive technique that appears more advantageous and promising, as the equipment for its execution is low cost and easy to use (even at home). It uses a low intensity current (1 or 2 mA) developed between two electrodes placed on the scalp. Depending on the polarity of the electrodes, the neurons are polarized to become more or less excitable.
Static and Dynamic Splints[edit | edit source]
Static and dynamic splints were initially used with the aim of reducing spasticity and preventing the onset of contractures, but a recent 2011 review showed that wearing a splint for the entire night and for several consecutive days does not effectively reduce spasticity and does not prevent the onset of contractures, but on the contrary, it has a negative effect. The main reason is that maintaining splints for an extended period of time tends to reduce wrist mobility, induce disuse and consequent muscle atrophy; moreover, the splints tend to be passive and unable to stimulate neural plasticity and motor learning. Then there are the dynamic splints that use additional components (springs, threads, elastic,) to mobilize the joints, in order to provide a controlled and delicate force to the soft tissues for a long period of time, so as to facilitate the remodeling of the tissues without tear them off. However, it is difficult to determine the amount of force to be released while preventing further damage to the joints and stimulating voluntary movement.
Pharmacological[edit | edit source]
Drugs used to modify neurotransmission and excitability of the motor cortex. However, it must be remembered that the action of drugs is not hemisphere-specific like other types of treatment described above.
Constraint-Induced Movement Therapy[edit | edit source]
Constraint-induced movement therapy (CIMT) develops from the idea that the patient tends not to use the paretic limb, damaged by the injury. This non-use of the affected limb occurs due to a reduction in neural excitability in the vicinity of the lesion, continuous negative feedback following unsuccessful attempts at action, an alteration of the inter-hemispheric interaction with increased excitability of the contralateral cortex, etc. The CIMT aims to rebalance the activity of the motor cortex, limiting the activity of the healthy side (asking the patient to wear a glove or a containment bandage) while the affected limb is intensely trained. The physiotherapist must propose exercises appropriate to the patient's abilities and at the same time provide feedback and comfort in case of errors.
Bilateral Training[edit | edit source]
This type of intervention that includes the execution of activities, such as reaching and grasping, simultaneously with both arms, since the ability to coordinate both upper limbs during a bilateral task in stroke patients is partially preserved . It is believed that this approach can improve the performance of motor tasks with the paretic limb thanks to the healthy part of the motor system that facilitates the motor representation of the paretic musculature. Although there is still no clear evidence, it seems that patients with greater impairment and who tend to recruit more the contralesional motor cortex during movements of the paretic limb, have a more significant benefit. Bilateral training can be combined with robotic therapy.
Functional Electrical stimulation (FES) and Transcutaneous Electrical Nerve Stimulation (TENS)[edit | edit source]
These can be used to activate muscles and generate sensory inputs. Normal movements are produced by delicate and complex patterns of muscle activation which, however, are not reproducible by the massive and generalized activation produced by TENS. It seems that the FES allows instead to facilitate the generation of voluntary movements especially at the level of activation of the elbow, wrist and extensor fingers during functional tasks. A limitation is that the motor neurons that are more easily activated for low electrical intensities induced from the outside are those with a larger diameter that control the more full-bodied muscle bundles, contrary to what happens physiologically. It follows that through external stimulation it is currently not possible to activate the so-called "fine movements".
Electromyographic (EMG) Biofeedback[edit | edit source]
This method provides visual and auditory feedback to patients regarding the time and strength of muscle activation recorded thanks to surface electromyography. Although the apparent neuronal mechanisms underlying this technique are not known, recent evidence indicates that biofeedback can modulate the patterns of cortical activity.
Electromechanical or Robotic-Assisted Therapy[edit | edit source]
Different types of devices have been designed to help the movement of the paretic upper limb. These technologies provide the motor system with the sensory feedback it would receive if it were able to move normally; such feedback allows the patient to create internal models of desired movement patterns, increasing the excitability of the motor pathways. Beyond the feedback, robotic therapy can deliver large amounts of therapy relatively cheaply. Early generations of robots provided motion assistance with a fixed torque, regardless of the patients' voluntary contractions. Newer models can instead sense the amount of voluntary movement of the patient and correct the degree of assistance. Numerous studies have analyzed the effects of robotic motion assist therapy compared to other interventions, yielding weak to robust results for ADL, strength and motor function, but more studies are needed to better define the intervention "dose" and consequent effects.
Virtual Reality[edit | edit source]
Virtual reality offers more sensitive feedback while the patient is immersed in a virtual environment seeing their body in motion. We are trying to reduce the time delay between the visual information received by the user and his movements performed during the exercise. The difficulty of the exercises for the upper limb can be modulated according to the performance and the patient's motivation is greater because the treatment takes on a playful aspect. A study has shown that training in a two-dimensional environment (1 hour a day for 4 weeks) increased the functional capabilities of the arm of 5 chronic patients. The control group consisted of 5 patients who were not treated with virtual reality and showed no improvement.
Task-Specific Practice[edit | edit source]
Task-specific practice exercise is the heart of physiotherapy. Repetition of a functional task can produce significant improvements in function and motor skills. Exercise improves performance because it provides the nervous system with repeated opportunities to estimate the state of the body and integrate this with a movement goal, to produce ordered, regulated, progressive motor commands and to adjust motor output based on sensory feedback. The scientific rationale of the task-specific exercise is based on the principles of the plasticity of the motor cortex dependent on experience. It is essential to guide the patient to identify the key components of a task (for example speed and distance) and to provide positive verbal feedback with each small improvement of these components, while the demand for the task is progressively increased. Several efficacy studies have been carried out in relation to the "dose" of therapy also for this type of intervention and double effects were found in patients with more than 20 hours of exercise compared to patients with 0 to 20 hours of exercise. By focusing our attention on the upper limb, this method aims to make the patient able to perform voluntary movements, such as reaching a distant object while sitting or bringing a glass to his mouth. A recent study found the importance of avoiding "compensatory" movements in the execution of exercises from the first moments after the stroke, because these movements could then be difficult to eliminate. In particular, it was found that significant improvements have been achieved, even by patients in the chronic phase, trying to avoid / minimize trunk movements: these improvements result in a reduction in disability and an increase in limb functionality. The objectives of task-specific training are: to promote the acquisition of skills, strength, speed, coordination and modulation of effort.
Robot Techniques[edit | edit source]
Rehabilitation with robotic systems offers patients high dosages and training intensities, allowing a repetitive practice of specific movements and / or functional tasks. Indeed, it is precisely through repetition that cortical reorganization can be induced after the stroke. In the last 30 years, in the field of rehabilitation of patients suffering from stroke outcomes, the development of research has been directed, among the various technologies, towards the experimentation of robotic systems for the recovery of motor function of the upper limb. Initially, the attention of the researchers focused in particular on the proximal extremities, thus creating specific robots for the recovery of the functionality of the shoulder and elbow. In recent years, however, there has been an increase in the development of robotic tools for the assistance and training of the distal extremities of the upper limb, therefore of the wrist and fingers.
The secondary interest in the distal end of the upper limb can also be traced back to the fact that designing robotic systems for hand training is a rather difficult task, by virtue of the complexity and versatility of the human hand. Just think that the fingers and the thumb have 21 degrees of freedom while the arm, from the wrist to the shoulder, has only 7.
There are basically two different types of robots used for the motor recovery of the hand: End-effector devices and exoskeleters.
- The End-effector devices, also known as haptic interfaces, are robots consisting of devices designed to be connected with the end of the robotic arm, so that the patient holds these devices to interact with the environment. Using these robots, it is possible to create environments that reproduce activities of daily life or functions of the hand, assuming that it is the patient himself who determines the movements at the level of the joints. The advantages of the end-effectors are the simplicity of assembly, the easy adaptability to the different sizes and characteristics of the hands with small adjustments, the possibility of being used by both right-handed and left-handed patients. A disadvantage is that these robots allow limited control of the proximal joints of the arm and this can favor the establishment of abnormal movement patterns during training.
- The exoskeleters, on the other hand, are external cybernetic devices that follow the anatomy of the patient's upper limb, to which they are attached in some points, thus constituting a sort of "artificial musculature". One of the advantages of exoskeletons is the possibility of directly controlling the affected joints as the robotic axes are aligned with the anatomical ones of the patient, thus minimizing the compensation. Among the disadvantages are the complexity in the design, the high costs, the difficulty in modulating them according to the needs of the various patients.
Despite the different characteristics of the two types of robots, it should be emphasized that recent studies on the efficacy of robot-assisted upper limb rehabilitation with exoskeletons and end-effectors, have shown improvements in the motor function of the hand, equivalent or superior to conventional therapy in chronic patients. The complex structure of the wrist allows you to adapt the orientation of the hand according to the required tasks and to lock its position while interacting with the external environment. Movement generally occurs around two main axes: flexion / extension and abduction / adduction. The pronation / supination movements are instead performed by the radio-ulnar complex, with some contribution from the shoulder in specific positions of the arm. The importance of the rehabilitation of wrist movements is also underlined by a study by Butefisch et al.  which states that repetitive extension movements of the affected wrist, in stroke patients, not only allow to improve the degree of efficiency of biomechanical parameters such as grip strength, isometric extension and isotonic extension speed, but they also improve the motor function of the entire upper extremity, assessed with the Rivermead Motor Assessment.
- Wrist Robot: 3 degrees of freedom exoskeleton that allows flexion-extension, abduction-adduction, pronation-supination movements of the wrist. The patient grasps a handle connected to the robot with his hand while the forearm is contained within a rigid support. One study demonstrated improvements in wrist ROM especially for flexion-extension and abduction-adduction rather than pronation-supination movements.
- MIT - Manus: Device created in 1991, initially only for shoulder and elbow; subsequently a module for the rehabilitation and exercise of the wrist was added, which allows the movements of adduction - abduction (30 ° / 30 °), flexion-extension (60 ° / 60 °) and pronation-supination (70 ° / 70 °). It can be used alone or connected to the original MIT-Manus, having in the latter case 5 active and 2 passive degrees of freedom. Even for the most seriously ill, the latest version allows you to prepare everything you need in less than two minutes. The patient's hand is placed in a power grip position around the knob, with the minimum necessary supports. The robot can provide continuous passive mobilization, sensorimotor or wrist strength training. Studies on the application of both devices, for the shoulder and elbow and for the wrist, have observed improvements of 10% in a 12-week training program. In other studies improvements were also observed in both acute, subacute and chronic patients.
- Bi-Manu-Track: 2 degrees of freedom device; allows you to perform pronation-supination and flexion-extension movements. Both hands are tied to a handlebar that moves with a rotary motion; the exercises can be mirrored (the patient observes the sound hand in a mirror while making the movements) or parallel (both hands perform the same movement). There are three modes of control: passive-passive, active-passive (the healthy hand guides the diseased one) and active-active (the diseased hand must overcome a certain initial resistance to be able to perform the movement). Each hand grasps a 3 cm diameter knob and is held in place by a 6 cm velcro strap. The two knobs are connected to each other by an axis, connected in turn to the motor. Some studies have shown significant improvements in various parameters such as ROM, Fugl - Meyer and strength, other studies have shown reduction of the spasticity of the wrist, resulting in a better chance of hand hygiene and pain relief.
- RiceWrist: electrically operated exoskeleton for the forearm and wrist. It is an evolution of the MAHI exoskeleton, which can interface with the MIME (Mirror Image Movement Enabler), an exoskeleton for the rehabilitation of the shoulder and elbow, to provide various modes of interactions. Work can be set in passive, active-assisted or forced mode. Devices for the recovery of the functionality of the thumb and fingers
- The Hand Wrist Assistive Rehabilitation Device or HWARD: `exoskeleton with 3 degrees of freedom (flexion-extension of the fingers, thumb and wrist) that assists (through pneumatic actuators) the hand in the" grasp and release "movement. The patient is seated in front of a monitor, the hand is secured to the robot by three soft Velcro strips, and the forearm is secured inside a padded splint. The palm of the hand is left free allowing the grasping of real objects, as well as virtual ones. The most important result of the study is the improvement, in chronic patients with moderate motor deficits, of the functionality of the distal extremity of the upper limb, improvements that persisted for at least one month after the end of treatment.
- Rutgers Hand Master II glove: device powered by pneumatic pistons positioned in the palm of the hand that simulate the forces of contact with objects in a virtual environment. Gains in ROM, speed and movement fractionation were found in most of the patients tested, while only in some were improvements in the Jebsen-Taylor Test of Hand Function. 
- The Amadeo: it's a robot for rehabilitation of hand function
- Reha-Digit: device consisting of 4 mutually independent cylinders, each fixed to the main driving axis, forming a kind of camshaft. Two smaller cylinders per finger hold it in place, exerting a force on the back thanks to elastic springs that “pull” towards the main cylinder. The forearm is supported by a support. The device, in its current form, is limited since it allows only passive finger movement. However, it is useful for patients with severe disabilities, as it has been shown that even a purely passive sensory-motor stimulation increases cortico-spinal excitability and induces long-term strengthening of the synapses. In chronic patients, the device did not lead to significant improvements in active movements, but it did reduce finger and wrist spasticity. In the subacute phase, however, more consistent improvements were found, but always in the passive movement.
- The Hand Mentor: commercially available exoskeleton that uses artificial musculature to flex or extend all fingers simultaneously.
- The Haptic Knob: presents a terminal device that allows you to train the opening and closing of the hand and the manipulation of the "button" to simulate interaction with objects. The robot can provide assistance or resistance, and various types of "buttons" can be mounted on the interface to vary the type of handling.
- The HandCARE system: robotic device that uses cables for each finger, driven by a motor and a pulley system. The five fingers can be trained either all together or in different combinations to improve both coordination and singular movements.
According to some studies, for effective rehabilitation of the upper limb and hand after stroke, motor training should be focused on realistic tasks that require reaching and manipulating real objects. Several robotic systems have been developed according to these theories. Between these :
- The Activities of Daily Living Exercise Robot or ADLER (2006): robotic device that allows assisted training of ADLs, guiding the affected limb along trajectories to perform real-life tasks such as drinking water, eating, fixing hair, etc. ;
- The Hand Wrist Assistive Rehabilitation Device or HWARD: described above, allows patients to interact with both virtual and real objects;
- The Haptic Knob: robot described above, allows you to train the opening and closing of the hand and the interaction with various objects that can be mounted on the interface;
- The HandCARE system: described above, robot with a cable-operated mechanism that allows you to train the grip, release and isolated control of the fingers necessary for ADLs;
- The Adaptive and Automatic Presentation of Task or ADAPT: robotic system that stands out from most other devices due to its particular design based on task-oriented training; this allows patients to be provided with adaptable realistic functional tasks such as ringing a bell, opening or closing a lock, opening a jar, etc.
In a review  comparing data from multiple studies on hand rehabilitation performed with different devices (BiMANU Track, PneuGlove, HWARD, Hand Mentor, HapticKnob) comforting results were obtained. In fact, the robotic rehabilitation of the hand has determined a reduction of motor impairments, measured with the Fugl-Meyer, both in patients in the sub-acute and chronic phase. In addition, improvements in the functional use of the limb were also highlighted, evaluated with the Action Research Arm Test, the Box and Blocks Test and the Wolf Motor Function test. Although it must be said that it is not clear whether these improvements later resulted in greater use of the affected extremity in daily life. A further particularly interesting fact that emerged from the review is that the rehabilitation of the hand with these robotic devices has made it possible to obtain a reduction in motor dysfunctions of the entire upper limb, while the rehabilitation of the proximal extremities resulted in localized improvements exclusively in the shoulder and elbow. The possible explanations are different, such as the fact that the training of the distal extremity of the upper limb determines a greater general activation of the sensorimotor cortex compared to the training of the proximal extremity, just think of the sensorimotor homunculus and the greater territory of distribution of the hand relative to the rest of the arm; but also the fact that the proximal end is still activated by the patient during therapy for the hand or to stabilize the entire limb and allow more selectivity of movement at the distal level or to compensate for the motor deficit of the hand. However, it is too early to draw firm conclusions on this and further studies are certainly needed. These results are very promising and show the importance of hand rehabilitation, which traditionally was not considered among the main objectives of rehabilitation after stroke. But to build definitive conclusions on the efficacy of robotic therapy for the hand, more extensive clinical investigations are needed since the studies carried out to date have mainly been pilot studies with small samples. Furthermore, to make these devices accessible to all, it would be necessary to make them small in size, easy to use and cheap.
References[edit | edit source]
- World Health Organization. Stroke. http://www.who.int/en/ (accessed 13/10/2020)
- Donnan GA, Fisher M, Madeod M, Davis SM. Stroke [Seminar]. Lancet, 2008, 371:1612-23
- Oujamaa L, Relave I, Froger J, Mottet D, Pelissier JY. Rehabilitation of arm function after stroke. Literature review. Annals of physical and rehabilitation medicine. 2009 Apr 1;52(3):269-93.
- Pomeroy V, Aglioti SM, Mark VW, McFarland D, Stinear C, Wolf SL, Corbetta M, Fitzpatrick SM. Neurological principles and rehabilitation of action disorders: rehabilitation interventions. Neurorehabilitation and neural repair. 2011 Jun;25(5_suppl):33S-43S.
- Lannin NA, Ada L. Neurorehabilitation splinting: theory and principles of clinical use. NeuroRehabilitation. 2011 Jan 1;28(1):21-8.
- Bütefisch C, Hummelsheim H, Denzler P, Mauritz KH. Repetitive training of isolated movements improves the outcome of motor rehabilitation of the centrally paretic hand. Journal of the neurological sciences. 1995 May 1;130(1):59-68.
- Balasubramanian S, Klein J, Burdet E. Robot-assisted rehabilitation of hand function. Current opinion in neurology. 2010 Dec 1;23(6):661-70.
- Chang WH, Kim YH. Robot-assisted therapy in stroke rehabilitation. Journal of stroke. 2013 Sep;15(3):174.
- Squeri V, Masia L, Taverna L, Morasso P. Improving the ROM of wrist movements in stroke patients by means of a haptic wrist robot. In2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2011 (pp. 1268-1271). IEEE.
- Krebs HI, Volpe BT, Williams D, Celestino J, Charles SK, Lynch D, Hogan N. Robot-aided neurorehabilitation: a robot for wrist rehabilitation. IEEE transactions on neural systems and rehabilitation engineering. 2007 Sep 17;15(3):327-35.
- Hesse S, Schulte-Tigges G, Konrad M, Bardeleben A, Werner C. Robot-assisted arm trainer for the passive and active practice of bilateral forearm and wrist movements in hemiparetic subjects. Archives of physical medicine and rehabilitation. 2003 Jun 1;84(6):915-20.
- Gupta A, O'Malley MK, Patoglu V, Burgar C. Design, control and performance of RiceWrist: a force feedback wrist exoskeleton for rehabilitation and training. The International Journal of Robotics Research. 2008 Feb;27(2):233-51.
- Takahashi CD, Der-Yeghiaian L, Le V, Motiwala RR, Cramer SC. Robot-based hand motor therapy after stroke. Brain. 2008 Feb 1;131(2):425-37.
- Bouzit M, Popescu G, Burdea G, Boian R. The Rutgers Master II-ND force feedback glove. InProceedings 10th Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems. HAPTICS 2002 2002 Mar 24 (pp. 145-152). IEEE.
- Koeneman EJ, Schultz RS, Wolf SL, Herring DE, Koeneman JB. A pneumatic muscle hand therapy device. InThe 26th Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2004 Sep 1 (Vol. 1, pp. 2711-2713). IEEE.
- Lambercy O, Dovat L, Gassert R, Burdet E, Teo CL, Milner T. A haptic knob for rehabilitation of hand function. IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2007 Sep 17;15(3):356-66.
- Dovat L, Lambercy O, Gassert R, Maeder T, Milner T, Leong TC, Burdet E. HandCARE: a cable-actuated rehabilitation system to train hand function after stroke. IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2008 Dec;16(6):582-91.
- Schweighofer N, Choi Y, Winstein C, Gordon J. Task-oriented rehabilitation robotics. American Journal of Physical Medicine & Rehabilitation. 2012 Nov 1;91(11):S270-9.
- Wisneski KJ, Johnson MJ. Quantifying kinematics of purposeful movements to real, imagined, or absent functional objects: implications for modelling trajectories for robot-assisted ADL tasks. Journal of NeuroEngineering and Rehabilitation. 2007 Dec 1;4(1):7.
- Takahashi CD, Der-Yeghiaian L, Le V, Motiwala RR, Cramer SC. Robot-based hand motor therapy after stroke. Brain. 2008 Feb 1;131(2):425-37.
- Squeri V, Masia L, Giannoni P, Sandini G, Morasso P. Wrist rehabilitation in chronic stroke patients by means of adaptive, progressive robot-aided therapy. IEEE transactions on neural systems and rehabilitation engineering. 2013 Mar 13;22(2):312-25.