|
Dr. Shanechi recently published a paper titled "A Real-Time Brain-Machine Interface Combining Motor Target and Trajectory Intent Using an Optimal Feedback Control Design." In this conversation, she discusses her approach, challenges, and the value of interdisciplinary collaboration.
NNx: How does your approach to BMIs differ from what other labs are doing?
Dr. Shanechi: First, this real-time BMI not only decodes the trajectory of movement from the neural activity during execution, but also decodes the intended target of movement from the neural activity prior to movement execution. We show that decoding both target and trajectory improves the BMI performance in real time. Second, this work models the BMI as an optimal feedback-control system. In this model of BMI, the primate controls the movement based on the visual feedback of the state of the BMI, the target or goal of movement, and its internal forward model for movement. We show that this real-time optimal feedback control modeling improves the BMI performance.
NNx: What are the main obstacles that need to be overcome in order for BMIs to be a large-scale, viable treatment for humans?
Dr. Shanechi: For BMIs to be truly clinically viable, improvement is needed in the reliability and safety of the multi-electrode implants such that they can last for decades and provide reliable signals. Moreover, advanced mathematical decoding and control algorithms need to be developed to improve the BMI performance and allow subjects to use BMIs to perform more complex three dimensional tasks.
NNx: Your paper has many authors. How important do you think multidisciplinary collaboration is in order to produce impact-worthy research?
Dr. Shanechi: Interdisciplinary collaborations are key to the success of the BMIs field as improvements are needed on multiple fronts to make them clinically viable. These include improvements in the mathematical decoding models and control algorithms, multi-electrode array design, prosthetic arm design, neurophysiology of motor control, to name a few. Our work was the result of collaborations across three departments at MIT and Harvard.
NNx: How important is the ability to record with multiple recording sites for the development of BMIs? In your perfect world, how many sites would you like to be able to record from simultaneously to produce an optimal BMI?
Dr. Shanechi: Reliability of the recording devices are key for the clinical viability of BMIs. A BMI that can record reliably from multiple cortical areas can allow the decoding of both higher-level goal information and lower level kinematic information. For example in our paper we decode both the high level goal or target of movement and the low level kinematics of the movement.
 If you have a question for Dr. Shanechi, comment in Science Update
|
