It is common following stroke for persons to direct their attention towards and rely heavily on their less affected upper extremity to complete tasks, effectively ignoring their paretic limb. The concept of learned nonuse states that a portion of the post-stroke functional deficit is not directly related to structural damage but rather occurs due to learned suppression of movement [11]. This behavior is acquired and reinforced during the acute phase following neurologic injury when attempts to use the paretic limb result in failure and compensation with the less affected limb is successful. CIMT forces the participant to use the affected upper extremity and is believed to overcome this phenomenon.

It has also been hypothesized that processes of neural plasticity and reorganization form the basis for motor recovery following stroke. The utilization of functional magnetic resonance imaging (fMRI) has facilitated our understanding of how the brain changes in response to rehabilitation techniques. While the cortical area representing the affected upper extremity has been shown to shrink in size following stroke [12], the literature suggests that CIMT is associated with both functional and structural brain reorganization [13].

In a randomized controlled trial (RCT) by Lin et al., the clinical improvements seen in the distributed CIMT group were accompanied by a significant increase in activation of the contralesional hemisphere during movement of the affected and unaffected hand [14]. This suggests that recovery in the affected upper extremity may occur through the establishment of an ipsilateral motor pathway [14]. Interestingly, different fMRI activation patterns were seen between the intervention and control groups of the study by Lin et al., which indicates that the type of cerebral reorganization may in fact be specific to the rehabilitation technique being employed. For example, the control group receiving dose-matched traditional therapy based on neurodevelopmental techniques showed a decrease in ipsilateral sensorimotor cortex activation during performance with the affected hand [14].

In addition, a longitudinal fMRI study by Murayama et al. showed that affected limb movement in post-stroke patients before receiving CIMT was associated with contralateral cerebellar activation on fMRI. When they were reassessed post-CIMT there was a change towards bilateral cerebellar activation. Subsequently, at 3 months post-CIMT there was a trend towards increasing ipsilateral cerebellar activation. Following CIMT, brain activation patterns of post-stroke patients developed to more closely resemble those seen in healthy controls [15].

Transcranial Doppler sonography (TCD) is another way of studying the brain functionally. Interpretation of TCD is based on the assumption that increased blood flow velocity within an artery is an indicator of increased regional brain activity [16]. In a study by Treger et al., post-stroke patients at baseline demonstrated reductions in mean blood flow velocity (MFV) within the middle cerebral artery (MCA) of the affected hemisphere when compared to that of the unaffected hemisphere. In healthy controls, the MFV was similar in both hemispheres at baseline. When performing motor tasks with the non-dominant hand, healthy controls showed a slight increase in MFV in both hemispheres and there was no significant change when the dominant hand was restrained. However, in post-stroke patients, restraint of the less affected upper limb while performing motor tasks with the affected side resulted in near normalization of MFV in the MCA of the affected hemisphere [17].

There is uncertainty regarding which component(s) of the CIMT protocol—restraint, mode of training, or therapy intensity—are most responsible for its therapeutic benefit. Studies comparing groups receiving the same therapy with and without a restraint found that all participants improve from baseline with no significant differences in outcomes between groups [18, 19]. Furthermore, extended use of the restraint after completion of the protocol does not appear to augment treatment outcomes [20]. Overall, restraint use does not appear to be a critical component of the protocol as participants make equal gains with and without it.

Conversely, the beneficial effects of CIMT appear to be more closely related to the mode of therapy as well as therapy intensity. When trying to determine whether the effectiveness of CIMT is attributable to intensity alone, it is important to look at studies with a control group receiving focused therapy of the affected upper extremity matched for intensity and duration. In the literature to date, there appears to be a trend towards non-inferiority of CIMT when compared with alternative high-intensity therapies focusing on the affected paretic limb [21]. Similarly, studies have shown CIMT to have a similar effect on upper extremity motor function when compared with intensity-matched bimanual therapy [22, 23]. Furthermore, a systematic review with meta-analysis performed by Stevenson et al. concluded that CIMT produced superior improvements in indicators of upper limb function in adult stroke survivors when compared with control interventions of equal dose and duration [24].

Traditionally, neurorecovery of the hemiparetic upper extremity is thought to occur predominantly during the first 3 months following stroke [37], though improvement has been shown to continue well beyond this period [38]. However, it is important to note that motor recovery in the upper extremity is notorious for lagging behind that in the lower extremity [39]. Originally, CIMT research focused on the chronic phase post-stroke. However, recent studies turned their attention to the acute and subacute phases.

The literature suggests that CIMT introduction within the first 14 days following stroke is safe. Pilot studies during this period show a trend towards greater improvements in affected limb function and use with CIMT when compared with traditional therapies [40–42]. The VECTORS trial compared traditional therapy with dose-matched mCIMT and high-intensity mCIMT. mCIMT was as effective as the intensity-matched control group during the acute phase following stroke. However, the high-intensity CIMT group showed less improvement in upper extremity function at 90 days; during the acute phase following stroke, there appears to be a threshold in terms of therapy intensity above which there is no added benefit and poorer outcomes may be observed [43].

The EXCITE trial also studied patients in the subacute phase and found that CIMT produced improvements in upper limb function that were both statistically and clinically significant when compared with customary care in patients 3–9 months post-stroke. Furthermore, these benefits were maintained upon reassessment at 1 year [4].

McIntyre and colleagues conducted a systematic review and meta-analysis of the evidence on the use of CIMT among stroke survivors more than 6 months following stroke [38]. They examined 16 RCTs and found that CIMT was associated with a significant benefit in terms of function as measured by the amount of use and quality of movement subscales of the Motor Activity Log, Fugl Meyer Assessment, and Action Research Arm Test. They concluded that for patients during the chronic phase post-stroke, CIMT is a beneficial therapy [38]. Similarly, the EXCITE trial concluded that regardless of whether CIMT was implemented 3–9 months or 15–21 months following stroke, patients reached the same level of affected upper limb function 2 years after their neurologic event [44].

Despite the prerequisites for participation described previously, CIMT has been used successfully in a variety of patient populations who do not fully meet these criteria.

Siebers et al. implemented a 2-week mCIMT protocol in a group of 20 outpatients with spastic hemiplegia and found improvement in functional upper limb use and reduction in spasticity as measured using the Modified Ashworth Scale that were maintained at the 6-month follow-up [45]. Sun et al. published a case study of a male patient 4 years post-stroke with severe flexor spasticity and nonuse of the dominant upper limb that did not meet the minimum motor requirements for participation in CIMT. He received botulinum toxin A injections targeting the elbow, wrist, and finger flexors followed by a 4-week mCIMT protocol and a 5-month home exercise program. He showed improvements in muscle tone as well as function and use of the affected upper limb [46]. Similarly, chronic post-stroke patients with plegic fisted hands at baseline have been shown to benefit from CIMT combined with conventional rehabilitation techniques [47].

In addition, Wu and colleagues demonstrated that a distributed 3-week CIMT protocol was well tolerated by elderly stroke survivors with considerable nonuse of their affected upper limb and resulted in significantly greater improvements in function when compared with traditional rehabilitation [48]. Similarly, Boe et al. determined that while patients with cognitive impairment required extra therapist attention, they still showed significant motor gains following 2 weeks of CIMT [49].