Recent Advances in Stroke Recovery and Rehabilitation


Wuwei Feng, MD, MS, Samir R. Belagaje, MD | Semin Neurol. 2013;33(5):498-506.

Copyright © 2014 by WebMD LLC. All rights reserved.

Read The Full Article Here


Stroke is the fourth leading cause of death in the United States, but remains a leading cause of disability. As more stroke victims survive with advanced acute care, effective strategies and interventions are required to optimize poststroke outcomes. 

In recent years, knowledge with respect to stroke recovery has expanded greatly through completion of preclinical and clinical trials. Emerging technology may provide further treatment options beyond the standard therapy and practices. In this article, the authors review recent advances in stroke recovery and rehabilitation, including the major determinants of poststroke recovery, challenges in translational stroke recovery research, and several emerging rehabilitation modalities such as noninvasive brain stimulation, brain–computer interface, biotherapeutics, and pharmacologic agents. Potential future directions in research are also addressed.


Stroke is the fourth leading cause of death and a leading cause of long-term disability in the United States.[1,2] The economic burden of stroke is huge. As of 1990, the average lifetime cost of caring for one stroke patient was estimated at $103,576,[3] including care costs for all phases (acute treatment, rehabilitation, ambulatory, and nursing home).

With improvement in acute stroke care and the establishment of stroke centers nationwide, more stroke victims survive, but with varying degrees of disability.[4,5] The common deficits associated with stroke are motor impairment (including limb spasticity), sensory impairment, language impairment (aphasia and/or dysarthria), dysphagia, cognitive impairment, visual impairment, and poststroke depression. Motor impairment is the most common physical deficit. Although the demand for effective and accessible rehabilitation therapies has tremendously increased over time, only a limited set of effective neurorehabilitative modalities are currently available while a few promising therapeutic options are still in research stages.[6,7]

In this article, we will review recent advances in stroke recovery and rehabilitation. We will provide (1) a summary of the major determinants of poststroke recovery, (2) a discussion of the challenging issues in translational stroke recovery research, (3) a review of several promising rehabilitation modalities; and (4) and overview of future research directions.

Understanding Major Determinants of Poststroke Recovery Process

Stroke recovery is a sophisticated biological process with many factors influencing the recovery trajectory. Rehabilitative strategies aiming to improve poststroke recovery outcomes require a comprehensive understanding of those factors. The recent advances in imaging tools, neurophysiological methods, and genetics have increased our knowledge about the poststroke recovery process. In this section, we will summarize the major determinants of recovery process, including initial injury, sociodemographic determinants, poststroke depression, rehabilitation therapeutics, and genetic factors (Fig. 1).


Figure 1.
Major determinants of poststroke recovery.

The Initial Injury

Initial injury is the most important prognostic factor for subsequent stroke recovery. The more severe the initial injury as defined by motor function, the more impairment patients will experience in the chronic phase (i.e., the lower Fugl-Meyer Upper Extremity score at baseline, the lower Fugl-Meyer Upper Extremity score 90 days post-stroke).[8] A recent study also showed that most stroke survivors (with the exception of patients with initial severe disability) can recover ~ 70% of their maximal recovery potential.[9] The corticospinal tract (CST) is a major motor pathway connecting the motor cortex and the limb muscles through the spinal cord. An assessment of the degree of initial injury of CST can be used to predict the motor impairment in the chronic phase. Several methods, including bedside motor assessment,[8,9] electrophysiological assessment (by transcranial magnetic stimulation [TMS] to induce motor evoked potential on the unaffected limb),[10,11] and advanced imaging tools (functional magnetic resonance imaging [MRI] to detect the brain activation pattern)[12] were all found to have some predictive value for motor impairment. Comorbidities such as diabetes, degree of periventricular white matter disease, and prior stroke can also adversely affect outcomes.[13–15]

Poststroke Depression

The prevalence of poststroke depression (PSD) can be > 30%.[16,17] The interaction between PSD and stroke recovery is complex, but studies have demonstrated that PSD can impede the poststroke rehabilitation and recovery process and adversely impact quality of life. Early treatment with antidepressants can promote motor recovery, which is supported by several clinical trials.[18–21]
Rehabilitation Therapeutics

The type, dosage, and timing of rehabilitation therapies play important roles in stroke recovery, but optimization of these parameters needs further refinement. For example, patients who suffered a stroke in the previous 3 to 9 months with some degree of wrist and finger movements can have a greater benefit from constraint-induced movement therapy (CIMT) over traditional therapy,[22] but such therapy is only equally effective and not superior to customary therapy if it is given in the very acute phase.[23] More discussion on this topic is continued in the next major section.

Sociodemographic Determinants

Sociodemographic factors are important determinants of stroke recovery. Although some studies have shown that increased age is a significant prognostic factor for poorer outcomes,[24–26] this remains controversial.[27,28] Overall, the influence of age on recovery is probably minimal after adjusting for other factors.[29] Women are less likely to achieve functional independence and more likely to be disabled than men,[30–32] but the mechanism for this difference is not clear. Racial disparity is also well documented in stroke. Blacks have a significantly higher stroke incidence,[33] greater initial stroke severity,[34] higher stroke mortality,[35] and poorer recovery outcome compared with Whites.[36] Socioeconomic status (insurance coverage, educational level, household income, etc.) is tied with access issues and subsequent recovery outcomes. The relative or complete lack of health insurance coverage may delay or limit access to rehabilitative services and is likely associated with less optimal recovery.

Genetic Factors

Genetic variation may account for some of the interindividual variability in recovery.[37] Several gene candidates such as brain-derived neurotrophic factor (BDNF) polymorphism,[38–44] apolipoprotein E (Apo-E) genotype,[45–47] catechol-O-methyl transferase (COMT) polymorphism[48] and mitochondrial DNA gene variations[49] may influence poststroke recovery. Of these gene candidates, BDNF is the most widely investigated. Brain-derived neurotrophic factor plays a major role in synaptic plasticity,[38] as well as in learning and memory,[39] thereby affecting stroke recovery. In animal studies, administration of exogenous BDNF either by intranasal[40] or by intravenous approach[41] has been shown to enhance recovery. In human studies, a functional SNP (rs6252) has been identified, in which a G-to-A substitution at nucleotide 19 results in amino acid switch from valine to methionine at codon 66. Brain-derived neurotrophic factor Val66Met polymorphism has been shown to partially affect activity-dependent BDNF secretion and impair motor skill acquisition.[42] Stroke patients with BDNF Val66Met polymorphism appear to have poorer outcomes.[43,44] Overall, more research is needed on the topic of genetic influence on poststroke recovery.

Challenges for Translational Stroke Recovery Research

As stroke rehabilitation progresses, the challenges for advancing the field become more apparent. In this section, we highlight several key questions and issues researchers and clinicians must face while conducting studies in stroke recovery and rehabilitation.

The first question is “When is the best time to start rehabilitation following a stroke?” In the most widely applied stroke-care models, rehabilitation is initiated immediately after the acute hospitalization. While therapists are involved in the acute hospitalization, rehabilitation intensity and goals frequently vary. For instance, some caregivers believe that stroke patients who receive tissue plasminogen activator (tPA) following stroke should be placed on bedrest to minimize postthrombolytic therapy complications such as falls to avoid risk of bleeding. Clinicians may be wary of increasing physical activity in patients if such activity might lead to an elevation in blood pressure, risk of falling, or early neurologic deterioration. Therapists may feel pressured to address practical goals, such as hygiene and transfers, while being less focused on rewiring the brain.

Recent trials have attempted to determine the optimal timing of rehabilitation, particularly with regard to early mobilization. In the A Very Early Rehabilitation Trial (AVERT), a multicenter, phase II randomized trial, hospitalized stroke patients were randomized to receive customary therapy or a very early intervention (VEM),[50] in which patients were mobilized as soon as practical after randomization with the goal of first mobilization within 24 hours following stroke onset. The VEM group also received additional assistance with the aim of being upright and out of bed at least twice daily. Patients received twice the “mobilization dose” compared with standard care. The intervention was delivered by a trained nurse and physiotherapy team for the first 14 days after stroke or until discharge from the acute stroke unit. The authors estimated that ~ 10% of the participants received tPA. As a phase II trial, the primary endpoints of AVERT were safety measures. There was no significant difference in the number of deaths between groups (standard care, 3 of 33; VEM, 8 of 38; p = 0.20). The number of falls, occurrence of early neurologic deterioration, and patient fatigue were similar between groups.[50] Subsequent analyses have demonstrated that exposure to very early and intensive mobilization led to a significantly faster return to walking than did standard stroke care (median 3.5 vs. 7.0 days, p = 0.032).[51] Moreover, VEM was independently associated with good functional outcome on the Barthel Index at 3 months (p = 0.05).[51] Comparable findings were found in a similarly conducted trial, named Very Early Rehabilitation or Intensive Telemetry after Stroke (VERITAS), albeit with a smaller number of participants.[52] A major drawback of both trials is small sample size. The phase III multicenter AVERT trial with a larger sample size is currently underway. Confirmatory findings could prompt standardization of earlier rehabilitation initiation. The results could also arguably be extrapolated to rehabilitation of other deficits, including aphasia, neglect, and upper extremity deficits.

The optimal intensity or “dosage” of rehabilitation also remains unclear. Both the duration and intensity of rehabilitation therapy are reported to influence stroke recovery outcomes.[53] However, a systematic review[54] concludes that “current evidence provides some, but limited, support for the hypothesis that higher therapy dosage enhances motor recovery after stroke.” Prospective dose-finding studies are required. Additionally, it is not clear whether there is a linear correlation between therapy dosage and the motor improvement after stroke. For example, based on the results of the Very Early Constraint Induced Movement Therapy (VECTORS) trial, it appears that more therapy does not always equate with better outcomes. VECTORS was a phase II randomized trial comparing traditional upper extremity therapy with dose-matched and high-intensity CIMT protocols administered over the course of 2 weeks. Patients were randomized 9.65 ± 4.5 days after stroke onset. Although all groups improved with time as measured by the Action Research Arm Test (ARAT) score, the high intensity CIMT group had significantly less improvement at day 90. There were not any significant differences found between the dose-matched CIMT and control groups at day 90.[23] Similarly, animal models have also demonstrated enlargement in areas of ischemia when constraint therapy is applied early after a stroke that correlates with poor function.[55] Based on these results, it appears that more data are needed before any definite conclusions about early application of intense therapy, such as CIMT, can be fully made.

In recent years, new technology and devices have been applied to stroke rehabilitation and are commercially available. In the Locomotor Experience Applied Post-Stroke (LEAPS) Trial, Duncan et al tested the role of body-weight-supported treadmill against standard home physical therapy program and also attempted to provide further answers into the timing of rehabilitation. In this single-blinded trial, 408 subjects were randomly assigned to one of three arms lasting 12 to 16 weeks: a home-based exercise program starting 2 months after the stroke or to the body-weight-supported treadmill locomotor program, starting either at 2 or 6 months after the stroke. The primary outcome of the study was improved walking function one year after the stroke. Most participants (52%) improved their walking function, but there were no significant differences between the three arms.[56] Serious adverse events were similar in all three arms, as were minor events, except that there was significantly more dizziness or faintness (p = 0.008) in the locomotor groups.[56] These results suggest that there are not any differences between body-weight-supported treadmill devices and a home-based physical therapy program. There were not any differences in outcomes in terms of the timing of therapy, but the earlier therapy groups seemed to recover faster than the later therapy group. The home-based program may be more pragmatic after consideration of the costs and accessibility. Similarly, robotic-assisted technology has not been shown to be useful in the rehabilitation of the upper extremity compared with standard therapy.[57]

Success in stroke animal models has not been consistently translated to success in human studies in terms of acute treatments, neuroprotection, or secondary prevention; this unfortunately also holds true for rehabilitation. Some deficits that humans experience after a stroke cannot be modeled well in animals. For instance, it is difficult, if not impossible, to develop an aphasia model in animals that could enable testing of rehabilitation therapeutics. It is now clear that poststroke processes are influenced by a variety of factors as summarized in here earlier. Only a few of these factors can be included in animal models and this will remain a constant challenge for future rehabilitation trials.

Emerging Rehabilitation Modalities

There are several novel rehabilitation modalities that demonstrate promising initial results and may represent future interventions for stroke. Among these, the most exciting techniques are noninvasive brain stimulation (transcranial direct current stimulation [tDCS], repetitive transcranial magnetic stimulation [rTMS]), brain–computer interfaces, biotherapeutics, and pharmacologic agents.
Noninvasive Brain Stimulation

The theoretical model serving as a basis for brain stimulation is a phenomenon called “interhemispheric interaction” between the two primary motor cortices.[58,59] This normal interaction is altered after a unilateral stroke and contributes to motor impairment.[60] This theoretical model takes into consideration (1) the imbalance of interhemispheric motor interactions after stroke, (2) decreased motor activity in the lesional hemisphere, and (3) excessive motor activity in the contralesional hemisphere. It is conceivable that modulation/correction of such an imbalance by tDCS or rTMS stimulation may lead to motor improvement in stroke patients (Fig. 2). Several studies have shown that targeting one hemisphere, such as upregulating excitability in the ipsilesional motor cortex[61–63] or downregulating excitability in the contralesional motor cortex,[64–66] improved motor function in patients with chronic stroke. This suggests that an approach that modulates the activity of both motor cortices simultaneously, such as applying anodal stimulation (by tDCS) or high-frequency stimulation (by rTMS) to the remaining motor cortex in the lesional hemisphere (upregulating excitability) and cathodal stimulation (by tDCS) or low-frequency stimulation (by rTMS) to the contralesional motor cortex (downregulating excitability) may be a better montage for enhancing motor recovery. This combined approach has been tested in several studies with bihemispheric tDCS or rTMS for stroke motor recovery.[67–69]


Figure 2.
Schematic model of poststroke interhemispheric imbalance and stimulation montages. (A) Interhemispheric imbalance as a result of stroke affecting one hemisphere, with excessive inhibition from the contralesional hemisphere. There are three options to correct such an imbalance. (B) Anodal stimulation or high-frequency repetitive transcranial magnetic stimulation (rTMS) of the lesional hemisphere to upregulate excitability. (C) Cathodal stimulation or low-frequency rTMS of the contralesional hemisphere to downregulate excitability and (D) bihemispheric stimulation of the lesional hemisphere with the anodal stimulation or high-frequency rTMS to upregulate excitability and stimulation of the contralesional hemisphere with the cathodal stimulation or low-frequency rTMS to downregulate excitability.

Although tDCS and rTMS both deliver brain stimulation via noninvasive approaches, the physical processes are different. Transcranial magnetic stimulation uses a magnetic field to induce an electric current in the motor cortex to trigger an action potential, while tDCS directly delivers weak electrical currents. The current is applied through conductive pads attached to the scalp. It does not trigger an action potential; rather, it modulates the membrane potential as well as the cortical excitability depending on the polarity of the stimulation. The tDCS instrumentation is portable and can deliver simulation while patients are simultaneously receiving rehabilitation therapy whereas rTMS can only deliver stimulation while patients are immobilized in a chair and cannot be applied simultaneously with therapy. Transcranial direct current stimulation and rTMS are compared further in Table 1 .

Feng et al[70] reviewed all of the tDCS studies associated with stroke recovery published before 2012. They found that tDCS has been investigated in poststroke motor recovery in almost 300 stroke patients (12 studies with 224 patients), dysphagia (1 study with 14 patients), aphasia (6 studies with 44 patients), and vision impairments (1 study with 8 patients). There are at least nine tDCS clinical trials on stroke recovery published since that review.[71–79] Most of the studies are proof-of-concept with small sample sizes, except for one large multicenter randomized study,[62] and the majority of study results are positive. ClinicalTrials.gov currently lists 25 ongoing tDCS studies associated with stroke recovery.[80] We expect to see a conclusive answer for tDCS application in stroke recovery soon. The challenging issues in tDCS research for stroke recovery include optimal dosage and montage, long-term safety in stroke patients, and the effect size of tDCS.

Repetitive transcranial magnetic stimulation was approved by the U.S. Food and Drug Administration (FDA) for “drug-resistant depression,”[81] but its clinical utility in stroke recovery is still unclear.[82,83] It is known to modulate cortical excitability and has been tested as a therapeutic approach for improving motor recovery. Generally, either a low-frequency stimulation is applied to the motor cortex of the contralesional hemisphere to down regulate the excitability of the contralesional hemisphere[65,66] or a high-frequency stimulation to the motor cortex of the lesioned hemisphere to increase excitability of the ipsilesional hemisphere.[63,84] In general, rTMS is safe for clinical application.[85] One of the safety concerns for stroke patients are seizures induced by rTMS, especially with the high-frequency stimulation.[84,86] Despite the encouraging results, there are several unanswered questions. Whether the lesional or contralesional hemisphere is a better target for rTMS in stroke patients is unclear, though stimulation on either hemisphere has demonstrated positive effect on outcomes. Double-blinded, sham-controlled phase III clinical trials with adequately powered sample sizes are needed to prove clinical efficacy.

Brain–Computer Interface

Another emerging technology is brain–computer interface (BCI) or brain–machine interface (BMI). In stroke survivors with motor impairment, the circuitry between desire to perform a motor task is disconnected from the circuitry to carry out that actual function. Therefore, the concept of BCI technology is to breach this disconnect by recording neural signals from the brain, decoding this neural activity, and using those decoded plans to control an external device.[87] The components to carry out the functions include a sensor to record cortical signals (usually an electroencephalogram [EEG] or invasive microelectrodes), a processor to extract appropriate signal (such as hand movement) and decode it, and an effector to carry out the intended signal (usually a computer-screen cursor, robotic limb, or wheelchair).[87] In some systems, there is sensory feedback provided to the patient to improve their motor functions. For these reasons, BCI is commonly classified as a neural prosthetic.

The application to stroke patients and its efficacy has been tested in recent years, albeit in preliminary and pilot studies, but the results are promising thus far. Currently, there is some evidence that BCI can be used in stroke rehabilitation to control functional electrostimulation devices to improve foot drop,[88] assist with robot-controlled therapy of upper extremities,[89] and assist with motor imagery therapy techniques.[90] Further research is needed to verify these preliminary results and better understand which stroke patients would best benefit from this type of therapy. Limitations of this therapy include limited signal decoding for complex movements, device cost, and the need for tailoring the therapy to individual patients.

Biotherapeutics and Pharmacological Agents

A biotherapeutic is a therapeutic material produced using biological means. In stroke recovery, this most commonly is associated with stem cells, but trophic factors are also included in this category. There have also been several recent advances in pharmacological interventions.

Because stem cells have the capacity to differentiate into various types of cells, these cells, once transplanted, could develop into neurons and glial cells and thus replace the brain tissue damaged from the stroke.[91] However, in the past decade, some evidence has suggested that differentiation into neural cells is not absolutely necessary to achieve favorable outcomes.[92] Increasing evidence suggests that development of stem cells into neurons and glial cells may not be the primary event leading to recovery. First, the beneficial effect observed after transplantation appears too early for the differentiation and integration of cells into local circuits. Second, in some studies, although neuronal differentiation is seen, the degree of differentiation and integration of the transplanted cells does not correlate with changes in functional outcome. As further support, peripherally transplanted cells do not have to cross the blood–brain barrier at all to induce neurorestorative effects.[93] In an observational animal model, after endogenous stem cells migrate to an area of injury, they proliferate, but then undergo apoptosis. It is likely that there are alternative mechanisms contributing to the therapeutic benefit of cell transplantation in stroke.

The delivery route for stem cells is also still being investigated, with delivery via either intravenous or intra-arterial route. In addition to transplanting exogenous cells, an alternative approach may be to recruit endogenous stem cells. Such endogenous stem cells have been seen in rat models of stroke where they reside in the subventricular zone.[94] Stromal cell derived factor-1 (SDF-1) and its receptor, CXC-chemokine receptor-4 (CXCR4) are important molecules in stem cell mobilization and migration.[95] Erythropoietin (EPO) appears to also serve as a migration signaling molecule for stem cells.[96] An infusion of EPO may help promote stem cell migration and may represent a novel treatment option for stroke recovery. The safety and practicality of these treatments are under investigation as well.

There have also been advances in the production of various trophic factors by stem cells. Factors shown to be a by-product of stem cells include vascular endothelial growth factor (VEGF), fibroblast growth factor (FGF), glial cell derived neurotrophic factor (GDNF), and BDNF.[95] These factors play a role in synaptogenesis by promoting axon and/or dendritic growth. Conceptually, rather than replacing damaged brain tissue, neural networks can be restored by a rewiring process. Given the role of trophic factors in the mechanisms of stroke recovery, exogenous administration of these factors could be efficacious. Some of these factors may be responsible for inducing migration of natural progenitor cells.

In addition to biotherapeutics, medications can also play an important role in promoting stroke recovery. This was most apparent in the Fluoxetine for Motor Recovery after Acute Ischemic Stroke (FLAME) Trial.[21] In this double-blind, placebo-controlled trial, 118 stroke patients with hemiplegia or hemiparesis were randomized to receive fluoxetine or placebo for 3 months starting 5 to 10 days after stroke onset. The main outcome measure was change in motor function as measured by the Fugl-Meyer Motor Scale (FMMS). In the fluoxetine group, the change score with FMMS was significantly higher than the placebo group, 34.0 (95% confidence interval [CI]: 29.7–38.4) versus 24.3 points (95% CI:19.9–28.7); p = 0·003.[21] Beyond simply treating poststroke depression (addressed in the trial), selective serotonin receptor inhibitors (SSRI) may help with rehabilitation, likely through neuroplasticity. Other antidepressant or neuromodulating agents have been examined as well with positive benefits. For instance, clinical trials utilizing cholinesterase inhibitors and glutaminergic medications suggest improvements in aphasia rehabilitation.[97–99] Dopaminergic medications may also address poststroke depression and attention.[100,101] The trials are limited by small sample sizes and heterogeneity of lesion size and locations.

Caution is warranted in interpreting results of such initial pharmacological studies; subsequent better-designed studies may contradict conclusions from earlier ones. An example is the study of amphetamines for stroke recovery. Studies in healthy humans demonstrated that premedication with amphetamines enhanced the effects of motor training.[102] Clinical trials with amphetamines in hemiparetic stroke patients have been conducted as well, with initial promising results if the drug was combined with physical therapy.[103] In the largest study of amphetamines in stroke recovery, 71 patients were randomized to 10 sessions of physiotherapy coupled with either 10 mg of D-amphetamine or placebo twice weekly beginning 5 to 10 days after stroke. Although there was not any overall difference between the groups, D-amphetamine accelerated the recovery of arm motor function among those with moderate deficit.[104] However, subsequent controlled randomized clinical trials showed mixed results.[105,106]

Future Directions

The next decade will be important for further advances in stroke recovery and rehabilitation. The questions with respect to timing and intensity of rehabilitation will hopefully be answered with the completion of the AVERT trials and other similar trials. One such trial nearing completion is the Interdisciplinary Comprehensive Arm Rehabilitation Evaluation Stroke (ICARE) Trial. This randomized controlled study of upper extremity poststroke recovery compares a program of challenging, intensive, and meaningful practice of tasks to dose equivalent usual customary occupational therapy (30 hours over 10 weeks) as well as usual and customary occupational therapy (estimated to be less than 30 hours).[107] Further advances in technology will address new pharmacological agents, stem cells, brain stimulation, and other novel treatments such as virtual reality.[108] Combining various therapies and interventions may augment the efficacy of individual interventions and could be a new direction of stroke intervention as well.

Biomarkers may also better identify patient candidates and tailor individualized treatments, thereby enabling both prediction and monitoring of patient responses to specific therapeutics. New technology will improve rehabilitation access to patients living in remote or rural areas. For instance, application of telemedicine networks to assess and evaluate the rehabilitation need of patients at home or in rural areas will be an area of further growth. Telemedicine has the potential to improve access and reduce treatment disparities for stroke patients who live in rural areas where rehabilitation centers may not be readily available. Patient-oriented rehabilitation programs for patients with residual deficits but exhausted insurance need to be developed; such programs may be more home-based in nature and focus more on “self-training” or “wellness.” We need therapies and rehabilitation strategies that are not only more effective but are more cost effective and efficient.

The field of stroke rehabilitation has a bright future. Physicians involved in stroke care will have many more tools for stroke recovery and rehabilitation in the coming years.