PI: Nenadic, Zoran and Do, An H. Proposal Number: 1134575
Problem Statement. Brain-computer interface (BCI) is a technology that enables direct brain control of external devices, without generating any motor outputs. BCI technology's greatest potential lies in the field of neuro-rehabilitation, with the ultimate goal of integrating BCI with limb prostheses or functional electrical stimulation (FES) of muscles to restore intuitive and natural movements to individuals with severe paralysis. Current state-of-the-art BCIs use micro-electrode arrays, implanted in the cortex, to acquire neuronal signals as the source of BCI control. However, application of this technology to human neuro-rehabilitation has been limited by the bio-incompatibility of the implant. Consequently, subdural electrocorticogram (ECoG) electrodes emerged as a promising BCI signal acquisition platform. Preliminary studies suggest that ECoG electrodes yield signals whose long-term stability properties are superior to those of microelectrode arrays, while providing a comparable amount of useful motor-related information. However, whether this information is sufficient for BCI control of an upper extremity prosthesis so as to perform goal-oriented tasks useful for human daily activities and justify the risks of electrode implantation surgery, has not been established. Research Plan. The primary goal of this pilot study is to assess the feasibility of ECoG-based BCI for control of arm prostheses in a small population of epilepsy patients who are undergoing ECoG electrodes implantation for epilepsy surgery evaluation. They will perform a series of executed and imagined elementary upper extremity movements while their ECoG, limb trajectories, and electromyogram (EMG) data will be recorded. This data will then be analyzed, and a predictive model to perform real-time decoding of upper extremity trajectories from ECoG signals will be derived. This model will then be incorporated into a BCI system, whose performance will be tested using real-time online control of an anthropomorphic robotic arm (a stand-in for an upper extremity prosthesis). Based on subjects? ability to achieve BCI control of the robot to perform elementary and goal-oriented motor tasks, the feasibility of ECoG-based BCIs for control of arm prostheses control will be assessed. Novelty: 1) The integration of an anthropomorphic robotic arm with an ECoG-based BCI has never been realized, and so its feasibility remains untested. 2) The study will introduce novel experimental paradigms in that ECoG will be recorded in response to both executed and kinesthetic imagery of upper extremity movements. 3) Unlike prior studies, our approach will record EMG of muscles involved in movements. The ability to predict EMG parameters from ECoG will be useful in future studies that aspire to integrate BCIs with upper extremity FES devices. 4) The study will contribute to understanding of the incompletely understood brain plasticity and human-computer co-adaptation processes associated with real-time online BCI control of upper extremity prostheses. 5) The project will lead to the development of a novel class of algorithms for statistical analysis and real-time decoding of ECoG signals. Intellectual Merit. The proposed study will develop novel, state-of-the-art, adaptive algorithms for analysis and real-time decoding of ECoG signals. These methods will avoid unnecessary assumptions and ad hoc strategies, commonly used in this field, and will therefore enable a systematic way of analyzing highdimensional, statistically sparse, nonstationary ECoG signals. They may also be applicable to a wide class spatio-temporal biomedical signals, and perhaps other types of statistical data. Aside from BCI applications, the ECoG, limb trajectory, and EMG data collected in the proposed study will contribute significantly to development of human motor control theory. Broader Impacts. The activities of this study will promote the education, scientific literacy and lifelong learning in undergraduate, graduate, and medical students. Specifically, elements of the study will be integrated into the teaching and mentoring curricula. Undergraduate and graduate students will participate in the proposed research and educational plans and their findings will be broadly disseminated. This will foster the development of their leadership, interdisciplinary, and research skills. The proposed activities will also broaden the participation of underrepresented groups in engineering and science. The investigators will promote college education and the pursuit of engineering/science careers in minority K-12 students by developing educational activities such as presentations, demonstrations, and exhibits. Additionally, the investigators will actively participate in the professional development of K-12 math and science teachers in high-need school districts, with the goal of improving their retention rates and leadership skills.
