IS 22. Coupling of motor imagination and nervous system stimulation to induce cortical plasticity

Abstract

We recently developed a novel technique for inducing plasticity in the human motor cortex by combining the physiologically generated signal when a person imagines a simple dorsiflexion task with the peripheral stimulation of the nerve that innervates the muscle involved in the task ( Mrachacz-Kersting et al., 2012 ). The subject activates the relevant brain areas via imagination and is provided with the expected afferent feedback via the single peripheral electrical stimulation to the target nerve. This protocol induced significant plasticity only when the afferent volley was timed to arrive during the peak negativity (PN) of the movement-related cortical potential (MRCP) generated in the corresponding brain area during the imagined task. The changes were specific to the target muscle, long lasting and rapidly evolving. These changes could be induced independent of if the subject was cued to start the movement or self-selected when to perform ( [Mrachacz-Kersting et al., 2012] , [Niazi et al., 2012] ). Only 50 repetitions of this combined stimulation paradigm are required to have an effect that outlasts the intervention. In a recent study with our collaborator Professor Kostic of the Department of Neurology, University of Belgrade, Serbia, we applied this intervention in a group of 13 chronic stroke subjects and evaluated neural plasticity and functional changes as quantified by the 10 m walk test and a foot tapping task. Patients attended five separate sessions. In the first and last session clinical measures and the MRCP during 50 attempted dorsiflexion movements were collected. In sessions two to four they were exposed to the intervention as described above. Across all subjects the motor evoked potentials (MEPs), quantified prior to and following the interventions increased significantly (80% on average). Patients were able to walk faster (on average 8% in 10 m walk test) and improved foot tapping frequency by 18%. Interestingly a finger tapping task showed no changes, thus further supporting the specificity of this intervention. Our initial results using the novel BCI interface leave a number of open issues. Firstly, the MRCP generated during task execution or imagination has several generators that differ depending on whether the task is cue based or self-paced ( Lu et al., 2012 ), yet our results show that our intervention is successful in both paradigms ( [Mrachacz-Kersting et al., 2012] , [Niazi et al., 2012] ). In an ongoing study we are randomly exposing healthy subjects to either paradigm following a selective blocking of either PMd or SMA to further enhance our understanding of the generators of the MRCP. Secondly, we have shown in the past that we can differentiate between ballistic versus slow movements as well as between low force and high force movements ( [do Nascimento and Farina, 2008] , [Gu et al., 2009] ), also from single trial MRCPs. The implications for the development of a brain-driven electrical stimulator that will provide the exact amount of afferent feedback necessary during an imagined movement are substantial. It will allow a more engaging approach to the rehabilitation of affected function that is tailored to the exact need of the patient. Current work aims at refining our BCI to allow this. The goal is to ask subjects to imagine a dorsiflexion consisting of different movement parameters (force and speed) and to provide the afferent feedback that would have been generated had the movement been performed rather than imagined. We alter both the rate and intensity of the peripheral nerve stimulation to meet this demand. Preliminary results show that cortical plasticity as assessed by changes in MEP size is further enhanced when the afferent feedback is matched to the characteristics of the imagined movement.