Neural plasticity and treatment across the lifespan for motor deficits in cerebral palsy



    1. Baltimore VA Medical Center Geriatric Research, Education and Clinical Center, Baltimore, MD, USA; Department of Neurology, University of Maryland School of Medicine, Baltimore, MD, USA.
    Search for more papers by this author

    The author declares no conflicts of interest.

George F Wittenberg at Baltimore VA Medical Center Geriatric Research, Education and Clinical Center, 10 N Greene St (BT/18/QR) 21201 Baltimore, MD, USA. E-mail


The past decade of research in neuroscience and stroke rehabilitation has demonstrated that the adult brain is capable of recovery through physiological processes (often called ‘plasticity’). Some of the recovery is spontaneous and some is a result of experience, including interventions such as physical therapy, which probably enhance or activate changes in brain structure and function. There is virtually no literature on physiological changes in the brains of children or adults with cerebral palsy (CP) after an intervention. It is unclear whether the principles of plasticity that have been deduced from animal models of stroke might also apply to children and adults with CP. But children with CP should have the potential to respond to experience in a similar way to adults, with the additional potential of regulation of neuronal development in response to injury. This article describes mechanisms of plasticity and a rehabilitation strategy to preserve the substrates for motor control in CP and then to apply later therapies for more refinement of motor control.


Brain-derived neurotrophic factor


Long-term depression


Long-term potentiation


Transcranial magnetic stimulation

Recent developments in neurorehabilitation have been stimulated by information on neuronal recovery processes and their modulation by various physical and pharmacological interventions. A growing body of evidence demonstrates that the human brain is capable of recovery after an injury through the ability of neurons and other brain cells to alter their structure and function (plasticity) in response to a variety of external and internal pressures, including behavioral training.1 From decades of basic neuroscience research, including animal models of stroke, neuroscientists have determined many principles of activity-dependent plasticity that may mediate recovery and rehabilitation outcomes in the damaged brain.

Because much of the basic research has come from animal models of stroke, a critical question is how generalizable these principles are to treating other neurological disorders resulting from brain injury. The brain injury resulting in cerebral palsy (CP) occurs early in neurodevelopment, whereas stroke generally occurs late in adult life. Thus, there is a difference not only in age at time of injury but also in the timing of the injury relative to treatment.

There is every reason to believe that plasticity is greater in the developing brain than in the mature brain; however, plasticity should be present in adults with CP just as it is in adults with stroke. If the processes underlying neuroplasticity were better understood, they might be enhanced and accessed by therapies administered throughout the lifespan. These might include physical therapy, such as task-specific training; pharmacological treatments, such as neuronal growth factors; and electrical methods, probably in combination.

Mechanisms of Neural Plasticity

In neuroscience research, the term plasticity has long been used to describe changes in organization in the nervous system; the details of these changes have not been fully described.2 A great deal is known about the rules governing central nervous system synaptic plasticity from work in young adult animal models.

The archetypal model is long-term potentiation (LTP) as described in the hippocampus,3 in which properly timed pre- and postsynaptic activity lead to specific strengthening of synaptic efficacy. Hippocampal LTP has been studied in detail, although its theoretical underpinnings remain controversial. It appears to occur by a mechanism in which presynaptic release of glutamate onto postsynaptic NMDA receptors (a subtype of glutamate receptors that are enabled by membrane depolarization) leads to a cascade of events that result in increased synaptic efficacy. LTP is sensitive to timing of pre- and postsynaptic events, so that it requires that the postsynaptic neuron be active close to the time of the presynaptic release of glutamate. But early and late changes that support LTP appear to be distinct, with the former mainly postsynaptic and the latter presynaptic.4 Thus, while glutamate and NMDA receptors appear to be critical to LTP, there are many other mechanisms, transmitters, and mediators.

Long-term potentiation has been shown to occur in areas other than the hippocampus,5,6 although less is known about mechanisms in those other areas. Long-term depression (LTD) is a related phenomenon that serves to reduce synaptic efficacy, generally when the timing of synaptic activity is unrelated to the occurrence of action potentials in the postsynaptic neuron. The existence of such counterbalancing processes appears to be important to the prevention of runaway changes in synaptic efficacy, which could result in seizures and other types of pathological activity. LTP and LTD are thought to be closely related to the learning and memory functions of the brain, and thus serve as models for experience-dependent changes that could also function during recovery from stroke.7 A glutamate-dependent, LTD-like motor plasticity in mice8 appears to be dependent on an endocannabinoid receptor for its expression.

Besides LTP and LTD, there are other mechanisms, described as homeostatic, that act to maintain mean neuronal firing rates near a criterion value. Homeostatic mechanisms may also be important in supporting recovery because the firing rates of the neurons not damaged by a stroke can be changed by loss of input of neurons that are damaged. This loss of input may be so severe that the remaining inputs can no longer modulate the firing rates of their postsynaptic targets (in a phenomenon known as diaschisis.) Homeostatic mechanisms may thus restore function by increasing excitability of those undamaged neurons and allowing them to respond to their remaining inputs. Homeostatic mechanisms may be important in development, although they may also be detrimental after injury, such as when spinal-reflex strength increases after stroke or spinal cord injury, resulting in the familiar clinical phenomenon of spasticity.

Neuronal mechanisms that promote stability in average firing rate are called homeostatic and include synaptic scaling, a non-specific change in synaptic strength, and modulation of neuronal excitability.9,10Synaptic scaling is partly mediated by activity-dependent release of brain-derived neurotrophic factor (BDNF) and thus is a mechanism that ties together experience with a trophic factor. BDNF is released and transferred to postsynaptic neurons in an activity-dependent manner.11 BDNF is globally affected by exercise,12 suggesting a link between non-specific exercise regimens and task-specific training.

Restoration of normal function after the occurrence of brain lesions may, therefore, depend, at least in part, on activity-dependent changes in excitability and synaptic strength. The known activity-dependent mechanisms include synaptic scaling, changes in excitability, timing-dependent mechanisms such as LTP and LTD, and shorter-term activity-dependent mechanisms not discussed here, but including facilitation and habituation. There are known signaling processes that mediate these mechanisms such as glutamate, BDNF in a forward direction, and endocannabinoids and nitric oxide retrograde. These processes and signaling molecules are targets for pharmacologic interventions.

Neuron Migration and Neurite Outgrowth

Neurons are born without neurites (i.e. dendrites and axons). Neurons may migrate soon after birth, and often protrude neurites as part of the guidance process by which they reach their final destination. The process by which neurons selectively connect to other neurons is highly complex, with evidence for surface-based and diffusible cues. This process may involve competition among multiple neurons as they attempt to innervate a target, so that damage to one set of neurons may confer a competitive advantage to another. In this manner, early damage may lead to preservation of connections that otherwise would have been lost to competition.

There are two areas in which plasticity is likely to be quite different in children with CP than in adults with stroke: control of neurite outgrowth and neuronal survival. This is because there is developmental regulation of these processes, with a much greater potential for plasticity in the younger brain. The adult brain normally is a non-permissive environment for neurite outgrowth owing to glycoproteins with inhibitory properties.13 The embryonic brain is characterized by a complex patterning of cell movements and development of connections; elucidation of these mechanisms and relationships has been among the greatest challenges to biological investigations of the past century.

One of the specific types of developmental plasticity that occurs in CP is the maintenance of axonal projections that would normally be retracted.14,15 The corticospinal tract in the adult normally crosses almost completely in the pyramidal decussation, while in the neonate there is a large uncrossed component. This uncrossed component is likely to represent axon collaterals to the normally crossing axons, explaining the presence of mirror movements in individuals with hemiplegic CP.16 Maturation to the adult phenotype is complete by age 13,17 suggesting that there is a fairly long period when interventions could take advantage of developmentally regulated mechanisms for plasticity.

Regulation of Neuron Number

Programmed cell death, or apoptosis, is another mechanism that shapes the developing nervous system and controls the number of neurons in the neonatal brain.18 In theory, loss of, or damage to, neurons in one part of the brain could result in downregulation of programmed cell death. In the adult, neurogenesis has been shown to occur after ischemic injury.19 But such responses in neuron number are empirically inadequate for restoring neurons and function in the damaged area. This may be because there are limitations to neuronal migration into the affected area, but it may also stem from the fact that the loss of trophic factors may increase, rather than decrease, programmed cell death outside of the lesioned area. Still, recognition that apoptosis could be modulated after injury to the brain means that research in this area has the potential to reduce neurological deficits.

Comparison of Plasticity in Stroke and CP

Recovery of motor function after stroke is not fully understood. There are differences in functional activity of the brain after a hemiparetic stroke20 and in responses to transcranial magnetic stimulation (TMS).21 Physical methods of restoring motor function also appear to affect both functional activity and TMS responses.22,23 Besides physical methods, non-invasive methods of stimulating the brain appear to have promise in promoting recovery.24–26 Numerous studies have demonstrated changes in brain physiology after a variety of rehabilitation methods in stroke; there are no such changes, to my knowledge, in CP.

From a theoretical standpoint, there should be more plasticity in the fetal and infant brain than in the adult brain, while the mechanisms for activity-dependent plasticity should apply across brain ages. Therefore, there are opportunities to intervene across the lifespan in CP. The empirical question that needs to be answered is what interventions are effective in CP, and whether it matters when they are applied. On the basis of current knowledge of development and plasticity, it would make sense early on to use therapies that could preserve the potential substrates for motor control, while later therapies could provide more refinement of motor control. Two therapies first used in stroke, constraint-induced movement therapy27 and robot-assisted movement practice,28 have been applied to children with promising results. But there is no reason to assume that the window for intervention cannot extend to adults with CP, just as many chronic stroke patients have been successfully treated years after onset.

Researchers have made efforts to assess the brain substrate for movement in children with CP and to use that knowledge to guide the application of therapy.29 Such studies suggest that therapies designed to restore the most normal type of motor system will be most successful. That means that early interventions should be directed toward improving survival of the crossed corticospinal tract axons, while later therapies should address the imbalance between intact and damaged motor representations.