Personalizing Rehabilitation and Brain Stimulation

By Ela Plow, PhD, PT; Vishwanath Sankarasubramanian, PhD; David Cunningham, MS; Kelsey Potter-Baker, PhD; Ken Sakaie, PhD; and Andre Machado, MD, PhD

Personalized medicine is an emerging medical model emphasizing interventions that are biologically or genetically tailored to maximize outcomes upon initial use, without the need for trial and error.

The great challenge facing neurologic rehabilitation, however, is the lack of information about which clinical or diagnostic characteristics are appropriate for deriving tailored therapies. A project called the BRAIN initiative (Brain Research through Advancing Innovative NeurotechnologiesSM), recently launched by the federal government in conjunction with the National Institutes of Health, promises to yield the first interactive map of the neural circuitry of the human brain. This type of mapping may provide the key to tailored rehabilitation treatments.

In our work at Cleveland Clinic, we are leveraging the concepts of the BRAIN initiative to develop personalized rehabilitation care programs. This work is based on our creation of an interactive map of the diseased or damaged brain.

Confronting variance in brain plasticity

Despite the fact that stroke is the most common and well-studied condition leading to persistent disability, the field of stroke rehabilitation is plagued with generic, non-specific therapies. The best example of this scattershot approach is seen with noninvasive brain stimulation.

Although the technology was initially considered promising as a means to increase adaptive neuroplasticity, the latest clinical trials have failed to demonstrate a consistent improvement in outcomes. Noninvasive brain stimulation using magnetic fields or direct current ultimately may prove helpful, but its current indiscriminate use ‒ driven by a one-size-fits-all approach based on the assumption of a generic substrate for brain plasticity ‒ is likely to produce inconsistent outcomes, given differences in the nature and extent of stroke-related disability among individual patients.

Rather than discounting the potential of brain stimulation to dramatically maximize and accelerate outcomes of rehabilitation in stroke, we operate on a conceptual framework based on tailoring the stimulation to an individual’s neurological characteristics.

Guided by our empirical understanding that mechanisms of neuroplasticity actually vary from patient to patient, we use advanced imaging neurotechnologies to investigate individual characteristics that generate such variance. These techniques include functional MRI to illustrate brain perfusion and function during real-time limb movement; diffusion tensor imaging to visualize the structure of white matter pathways devoted to moving the paralyzed limb; and transcranial magnetic stimulation to map the physiology of pathways and cortices.

Through these imaging innovations, we are able to determine which substrates remain spared in the damaged brain, how they interact with other regions and how they contribute to the potential to move paralyzed limbs. We process this information in concert with information regarding the patient’s pattern and severity of impairment, collected using validated clinical scales that track deficit and recovery.

Next, we process these neural and clinical characteristics via a hypothesis-driven decision tree, with the goal of developing treatment interventions unique to the patient’s pathology. The goal is to identify whether substrates on the injured or damaged side of the brain are spared adequately to be entrained with rehabilitation of the paretic upper limb, or if they are disrupted to such a degree that it would be best to rely on compensatory therapies such as those involving use of the less-affected side.

Considering that rehabilitation interventions are patient specific and customized to individual impairment and etiology, it is necessary to similarly customize stimulation therapies. After all, stimulation therapies seek to facilitate processes adopted in recovery with rehabilitation. Brain stimulation customized to a patient’s pathology will likely offer the most consistent boost to paired rehabilitative therapy.

In a clinical trial and two other clinical studies under way at Cleveland Clinic, we are attempting to understand how clinical and neural characteristics predict what substrates likely are inherent to a patient’s expression of plasticity. We are testing stimulation of such candidate substrates against the traditional approaches and validating whether stimulation of the “patient-specific” substrate is most effective for recovery.

Tailoring stimulation for a range of deficits

If this is successful in stroke, there is potential across other disease states. Our framework could be translated to tailor stimulation in conditions such as pain, vision loss and depression, and to potentiate therapies in brain injury, cerebral palsy and multiple sclerosis, where generic approaches have rendered rehabilitation arduous, varyingly effective and poorly funded.

In a much larger context, using a simple, noninvasive, inexpensive treatment tailored to what NIH Director Francis Collins, MD, PhD, has termed “brain types,” we challenge the long-standing assumption of generic plasticity.

With recent, drastic Medicare cuts for reimbursement of therapies, neurologic rehabilitative practice demands more effective, precise and accelerated outcomes. Brain stimulation guided by characteristics that maximize individual patients’ mechanisms of neuroplasticity would enable promising, consistent and prompt gains in neurologic rehabilitation.

Figure 1. Advanced diffusion tensor imaging methods help reconstruct all surviving pathways in the lesioned areas of the stroke-affected hemisphere. Differences between integrity of pathways in the affected vs. the unaffected hemisphere serve as an important baseline characteristic to predict levels of recovery and the types of brain stimulation therapies that can be employed. The images at top right represent pivot points used for the analysis ‒ internal capsule and motor cortices. Pathways are reconstructed between these nodes to form the analyses.

Figure 2. Similarly to Figure 1, this image illustrates the ability to reconstruct all surviving pathways in the lesioned areas of the stroke affected hemisphere. The image also demonstrates the ability to map key transcallosal pathways that collect bilateral motor cortices. Survival and physiology of these pathways are key to dictating chronic stroke recovery.

Figure 3. This image illustrates the ability to interface two mapping methodologies. Cuboid cells in blue and red represent points on the surface of the brain that are mapped with a neurophysiologic technique (transcranial magnetic stimulation, or TMS); red cells represent sites on the surface that were responsive to TMS, i.e., sites that are able to elicit neurophysiologic responses in the corresponding muscles of the hand. Pathways emerging from these sites have been reconstructed with diffusion tensor imaging (DTI). A combined TMS-DTI approach demonstrates regions of the brain that offer structurally and neurophysiologically sound pathways.

Figure 4. This image shows sites on the surface of the brain that are mapped with transcranial magnetic stimulation.

Dr. Plow is an assistant staff member in Cleveland Clinic Lerner Research Institute’s Department of Biomedical Engineering and an assistant professor of medicine at Cleveland Clinic’s Lerner College of Medicine. Drs. Sankarasubramanian and Potter-Baker are postdoctoral fellows and Mr. Cunningham is a doctoral student in Dr. Plow’s lab.

Dr. Sakaie is an assistant staff member of Cleveland Clinic’s Department of Diagnostic Radiology and of the Mellen Center, and an assistant professor of radiology at the Lerner College of Medicine. Dr. Machado is Director of the Center for Neurological Restoration in Cleveland Clinic’s Neurological Institute, and an associate professor of surgery at the Lerner College of Medicine.