A New Approach to Brain Stimulation for Post-Stroke Motor Recovery

By Andre Machado, MD, PhD

Despite significant improvements in acute care aimed at reducing tissue loss from stroke, there are still hundreds of thousands of patients in the United States alone who become permanently disabled by stroke.

To date, post-acute care consists mostly of physical and occupational therapy. While most patients experience at least some recovery, approximately half of patients who suffer a stroke require long-term assistance for activities of daily living. Given current limitations, new technologies are needed to facilitate post-stroke rehabilitation. These may include electrical or magnetic stimulation, stem cell therapies or other emerging technologies.

A new approach to targeting stimulation

In the past decade, major research efforts were dedicated to investigating the effects of cortical stimulation on post-stroke recovery. Despite success in proof-of-principle animal models and demonstration of safety in small patient cohorts, a large randomized study failed to show improvements beyond those achieved with rehabilitation alone.

We have proposed a different approach for applying electrical stimulation to the brain in an effort to augment recovery. Instead of aiming the electrical field at the cortical surface around the stroke, we have chosen to apply deep brain stimulation (DBS) to natural neural pathways that terminate in the perilesional cortex. This approach holds several hypothetical advantages:

• The effects of stimulation will be delivered to cells in specific cortical layers, irrespective of their location within a gyrus. This is an advantage over epidural cortical stimulation, which is likely to reach neurons at the crown of the gyri more efficiently than those in the depth of the sulci. In addition, cortical stimulation can have opposite effects depending on the orientation of neurons in relation to the electrodes, as shown by Holsheimer and colleagues, a limitation that is obviated if the effects of stimulation are delivered trans-synaptically.
• Depending on the targeted pathway, the effects of DBS can be carried to perilesional areas medial, lateral, anterior and posterior to the infarct core. This would only be possible with very large cortical grids, depending on infarct size.
• The technology for DBS is already very well developed. This is true also for the tools and techniques of implantation.
• DBS can directly target neural pathways associated with post-stroke reorganization at specific “nodes.” This allows for more energy-efficient chronic stimulation compared with large cortical matrices.

Selecting the pathway

Several neural pathways could be candidates for chronic DBS. Our choice was to target a long-loop neural network directly involved in post-stroke changes. The rationale for targeting the cerebello-thalamocortical pathway has been discussed in detail previously. Briefly, we based our rationale on the following:

• The well-known massive connections between the cerebrum and contralateral cerebellum and vice-versa
• The functional co-dependence of the cerebral and contralateral cerebellar hemispheres, illustrated by the dramatic metabolic changes of crossed-cerebellar diaschisis
• That patients suffering from reduced cerebellar output have contralateral cerebral hypoexcitability
• That patients with reduced cerebellar function due to crossed-cerebellar diaschisis have worse post-stroke motor outcomes than patients not affected by crossed-cerebellar diaschisis, as shown by Takasawa and colleagues
• The wide projection of the dentatothalamocortical pathway to the contralateral hemisphere via multiple thalamic nuclei

Hence, stimulation can affect cortical areas spared after large-convexity infarctions. In a very simplistic way, one could explain our approach as reversed or upside-down crossed-cerebellar diaschisis.

Proof of principle and current evidence

Our work began with several open questions, including:

• How should we choose the stimulation parameters that would best enhance cortical excitability?
• Would increased cortical excitability correlate with motor recovery?
• When should stimulation be initiated after stroke?
• Should stimulation be paired with motor training?

Stimulation parameters.To date, DBS therapies for conditions ranging from Parkinson disease and essential tremor to obsessive-compulsive disorder have used high-frequency (> 100 Hz) stimulation. High-frequency DBS is thought to have reversible, lesion-like effects. This is illustrated by the similar results obtained with Vim thalamotomy or Vim DBS in patients with essential tremor or Parkinson disease. However, the goal of our novel therapy is to enhance rather than diminish the cerebellar output.

To address the question of stimulation frequency, we used a rodent model of intracortical motor evoked potentials to index cortical excitability. These animals were also implanted with contralateral deep cerebellar electrodes. When cerebellar stimulation was set to high frequency, we observed a net reduction of cortical excitability, in agreement with current DBS models. We then tested the effects of various frequencies from 10 to 50 Hz. We found that frequencies between 20 and 50 Hz were associated with increments in cortical excitability but that maximal and sustained results were obtained within the beta frequency range.

First proof of principle. We also tested the effects of chronic deep cerebellar stimulation at various frequencies in a rodent model of middle cerebral artery (MCA) infarctions. Stroke was induced by microsurgical ligation of the MCA and temporary occlusion/reperfusion of the common carotid arteries. A nonsignificant trend for greater recovery was observed in animals treated with 50-Hz stimulation but not 10-Hz stimulation. However, animals treated with 20-Hz stimulation (i.e., within the beta band) presented with significantly greater recovery than did animals receiving sham treatment, that is, implanted with DBS leads and connected to stimulation apparatus but not stimulated. In this first test, animals were stimulated during 12 hours of their awake, freely moving cycle but, due to technical limitations, were disconnected from the stimulation system at the time of motor training and testing.

Paired treatment.In a recent evaluation, we changed our outcome measure for the rodent model to the pasta-matrix task. In this assessment, rats reach for pieces of pasta through a small slot, promoting use of the paw weakened by cerebral ischemia. A key advantage is that animals do not have to be disconnected from their stimulation devices during motor training/testing. Animals were divided into two groups: chronic stimulation at 30 Hz during the wake cycle, and sham treatment. In this re-evaluation we found significant improvements in motor function associated with chronic stimulation paired with motor training, indexed with the new task, relative to sham. In fact, treated animals returned to their pre-stroke baseline motor performance. We also found that chronic stimulation was associated with increased expression of synaptophysin in the perilesional cortex. Because synaptophysin is a protein present in presynaptic vesicles, this suggests that stimulation promoted recovery of function and perilesional synaptogenesis. Although these results hint at the possibility that recovery was mediated by greater synaptogenesis, causality has not been proven.

Next steps: Combining techniques to test new stimulation paradigms

Our group still has several challenges ahead, including achievement of human translation. Next steps in the laboratory include further evaluation of possible mechanisms associated with motor recovery and testing of new, complex stimulation paradigms.

We are currently using a combination of techniques to address these questions, in collaboration with the laboratory of Grahame Kidd, PhD, and Bruce Trapp, PhD, in Cleveland Clinic’s Department of Neurosciences. Further experiments should address the question of optimal timing and duration of chronic stimulation.

Dr. Machado is a neurosurgeon and Director of Cleveland Clinic’s Center for Neurological Restoration.