Brain Machine Interface for Controlling Neuroprosthesis for Reaching and Grasping

点击翻译按钮获取中文摘要

Neuroprostheses are devices that use bursts of low energy electrical pulses to generate muscle contractions. If the pulses are properly sequenced they can generate movements such as reaching, grasping and walking in individuals who are paralyzed. Since 2001, our team invested considerable amounts of intellectual capital to develop neuroprostheses for reaching and grasping. What differentiates our technology from other similar systems is that our neuroprostheses are not used as permanent orthoses patients need to use at all times to reach and grasp. Instead, we use neuroprosthesis as a short-term therapeutic intervention that helps stroke and spinal cord injury patients “reprogram” their central nervous system, and relearn how to reach and grasp objects on their own. We call this new way of using neuroprstheses the Functional Electrical Stimulation Therapy (FEST). Following FEST and after the neuroprosthesis is removed, the patients are able to reach and grasp on their own without help from the device. Brain Machine Interfaces (BMIs) are systems that use electrical signals recorded from the brain to control external devices such as computers and robotic arms. Given that they do not require any movement (i.e. muscle contraction) to enable the system control, the BMI systems hold enormous potential to assist profoundly disabled individuals such as those with severe stroke and spinal cord injury, to perform tasks they are unable to carry out independently. In 2015 our team discovered that if the FEST is controlled by a BMI system, which is triggered by brain signals associated with hand opening and closing, that such therapy generates stronger and more beneficial neuroplastic changes in the brain as compared to conventional FEST training alone. Therefore, the objective of this research program is to develop a variety of BMI technologies suitable for integration with FEST and to create first clinically viable BMI+FEST intervention. In order to control FEST with a BMI system, the BMI has to be able to determine: (i) what reaching and grasping movement(s) the patient intends to perform; (ii) the speed with which the patient would like these tasks to be performed; and (iii) initial and end positions of the arm during task execution. Ideally, this should be a non-invasive, electroencephalogram (EEG) -based BMI system. The specific objectives of this research program are: Task-1: Develop a BMI system able to detect the hand posture (e.g. pinch and lateral grasps) using EEG recordings. Task-2: Develop a BMI system able to detect the velocity and position of the hand in 3D space using EEG recordings. Task-3: Develop a self-contained hardware and software platform that incorporates BMI systems from Tasks 1 and 2. Task-4: Explore use of microelectrode recording and operant conditioning of single neurons as a means to generate control signals to be used to control FEST.