A Hybrid Brain-Computer Interface for Long-Term Use by Persons with Severe Motor Deficit: Towards Development of Personalized Algorithms

University Of Rhode Island, Kingston RI

Investigators

Abstract

The inability to communicate is a major consequence of severe motor disabilities such as amyotrophic lateral sclerosis (ALS), also known as Lou Gehrig's disease. ALS is the most common adult-onset motor neuron disease. Within the past three decades, Brain-Computer Interface (BCI) technologies have emerged as an alternative communication channel for persons with neuromuscular disorders. However, these communication systems typically require adequate eye-gaze control, which is difficult for those lacking voluntary muscle control. In addition, most current BCIs are inadequate for long-term use due to significant day-to-day variation in performance. To address these shortcomings, this project will combine two technologies to monitor the electrical properties of the neurons and the blood flow near the neurons. This combined system will adapt as the disease progresses and improve communication for persons with ALS. If successful, the system may be generalized to other types of motor control loss, including minimally conscious and vegetative states. The technologies developed will be integrated into a university curriculum that educates undergraduate and graduate students about the field. Furthermore, this research will be broadly disseminated through k-12 curriculum design and will provide training opportunities for women and under-represented minorities. Outreach activities include workshops: "Personalized Algorithms in BCI Systems" and "Hybrid BCI for Bedside and Home Users," poster presentations to disseminate novel outcomes from this study among students and researchers and development of a K-12 curriculum, "Engineering the Brain," designed to promote BCI learning opportunities among middle school students. The project focuses on designing Brain Computer Interfaces (BCIs) based on functional imaging of the human brain using a combination of two non-invasive techniques: electroencephalography (EEG-to measure electrical activity) and functional near-infrared spectroscopy (fNIRS-to measure hemodynamic activity). Algorithms will be developed such that the system considers both the neuropsychological status and environmental factors and can be implemented for in-home, long-term care of people with no motor control. The system will be tested on persons with ALS (amyotrophic lateral sclerosis), the most common adult-onset motor neuron disease. The project's hypothesis is that the incorporation of hemodynamic activities into conventional EEG-based BCI permits superior learning from brain states, provides unique features of the user's intent, and allows potentially game-changing solutions to provide end-users with a new level of communication autonomy. To test this hypothesis, the Research Plan is organized under two objectives. The FIRST Objective is to explore the associations of BCI performance variations and design a multimodal augmented predictive platform to improve the robustness of the BCI systems for nonverbal patients who have residual motor ability to control their eye-gaze. The two aims of the objective are to explore the associations of ALS-BCI performance variation with both functional brain changes in ALS and environmental noise and to establish a set of predictive electro-vascular features best representative of users' BCI performance, and accordingly correct for unrelated activities in future BCI experiments. Expected objective outcomes include (i) longitudinal assessment of brain pattern changes, (ii) establishing a set of predictive electro-vascular features, (iii) optimizing subject-specific factors essential for successful BCI performance, (iv) employing appropriate correction strategies to minimize unwanted activities and (v) maximizing adaptability with users' needs and environmental factors while minimizing inter-subject variabilities. The Second Objective is to develop an autonomous hybrid BCI for non-communicative persons without any residual motor control (preferably participants in the first objective who have progressed to the locked-in stage.) The system developed will have seven Degrees of Freedom (DOF) that can produce seven different commands for an external device based on the users' ability to modulate their brain signals through motor imagery (5 tasks), mental arithmetic (1 task) and rest. The expected objective outcome is the introduction of a hybrid BCI that (i) is fully autonomous-signal-based, (ii) employs personalized techniques to enhance system performance, (iii) has enhanced information transfer rates relative to single modal EEG and fNIRS and (iv) can be conveniently set up at users' bedsides for long-term use. This project is jointly funded by the Disabilities and Rehabilitation Engineering (DARE) program and the Established Program to Stimulate Competitive Research (EPSCoR). This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.