Direct Brain Interface Laboratory

The UM-DBI Laboratory’s current work focuses on the development of electroencephalogram (EEG)-based brain-computer interfaces (BCIs) into practical clinical tools for use by people with physical impairments. Barriers to clinical use include signal processing challenges, selection of tasks for BCI operation, interactions between BCIs and different conditions causing impairments, and technical support issues to troubleshoot in-home BCI use. UM-DBI Laboratory studies will address many of these areas.

SCH: New Statistical Learning Methods for Brain-Computer Interfaces

Grant funding (Grant #: 2123777 ) was received from the National Science Foundation (NSF) as part of the funding opportunity NSF 21-530 Smart Health and Biomedical Research in the Era of Artificial Intelligence and Advanced Data Science with Dr. Jian Kang from the Department of Biostatistics as the principal investigator.

Brain-computer interfaces (BCIs) are an emerging communication and computer access option for people with severe physical impairments, who are in a “locked-in” state due to an acquired or congenital disability. One of the most successful non-invasive BCIs for communication is the P300-BCI design, named after the brain activity that is called the P300 event-related potential (ERP). This design presents stimuli (flashing groups of keys) in an on-screen keyboard. The electroencephalogram (EEG) after each stimulus is then classified according to whether it contains the ERPs that are produced only for the stimulus the user has selected as their target. This design has been used for in-home communication by people with amyotrophic lateral sclerosis. This ERP-based BCI design can be calibrated for an individual in a single session. However, the BCI still takes time to calibrate to a user and the selection speed is slow. The calibration process has been especially challenging for people without other communication methods and for children, who have limited attention spans.

• This project will create new statistical methods that:

  1. Reduce the time required to calibrate the BCI for an individual user.

  2. Reduce the calibration effort for individual users by leveraging prior knowledge from other BCI users.

  3. Improve the selection speed of the BCI through dynamic adjustments to the patterns of stimuli.

This contribution is significant because the proposed methods will substantially improve the classification process and communication speed. The research outcome of this project will also provide new insights for a better understanding of brain functions and neurobiology of thinking, valuable information, and experience for the future design of the BCI system. The PIs plan to involve undergraduate and graduate students in the project, educate them on state-of-the-art research in BCI and statistical machine learning, organize summer training workshops, and develop free software.

The UM-DBI Laboratory will support the development of the new statistical methods by providing data from BCI experiments, participating in the development and testing of the statistical methods, assisting with the on-line implementation of the new methods and testing the new methods with research participants.

Real World Testing of a Brain-Computer Interface to Operate a Commercial Augmentative and Alternative Communication System

Grant funding (Grant#: 90IFDV0002) was received from the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR) in the Administration for Community Living.

For the most vulnerable individuals who cannot otherwise access augmentative and alternative communication (AAC) devices, access through brain computer interfaces (BCIs) offers the opportunity to obtain AAC’s vital quality-of-life benefits. However, little evidence exists on the features, clinical services and resources needed to effectively deliver an AAC-BCI.

The University of Michigan has partnered with the University of Pittsburgh, the Prentke Romich Company (PRC) and the ICAN Talk Clinic, as well as patients and caregivers, to meet this need.

• Goal: The project goal is development of an AAC-BCI device with innovative resources to improve the delivery and quality of AAC services for individuals and their families dealing with minimal movement due to amyotrophic lateral sclerosis (ALS), brainstem strokes, severe cerebral palsy, or traumatic brain injury (TBI).

• Objectives: The objectives are to: 1) test an AAC-BCI prototype that advances the effectiveness of current BCI communication; 2) improve the procedures and tools for comprehensive assessment to provide clinical evidence to support AAC-BCI funding; and 3) improve in-home training and treatment necessary for successful daily communication using an AAC-BCI.

• Outcomes: Anticipated outcomes include: 1) greater AAC access options for stakeholders (individuals, families, and practitioners); 2) improved tools for practitioners to compare access methods and recommend an AAC-BCI; 3) improved AAC-BCI in-home training resources for stakeholders; 4) improved clinical evidence for practitioners to support treatment decisions; and 5) available outcome data to advocate for billing codes and funding of AAC-BCI.

• Products: The expected products are an AAC-BCI prototype incorporating a commercial high-efficiency AAC device, dry electrode technology, assessment protocols, in-home training materials, and a language sample repository for data sharing.

Compatibility Between Brain-Computer Interface and High Efficiency Augmentative and Alternative Communication Systems

Grant funding (Grant #: R42DC015142 ) was received as a subaward from the Prentke Romich Company’s SBIR grant award from the National Institute on Deafness and other Communication Disorders (NIDCD).

Brain-computer interfaces (BCIs) enable text production for people who cannot move but include only simple communication interfaces and are not widely used. Augmentative and alternative communication (AAC) systems are widely used and give efficient and precise communication, but require movement, preventing use by people with severe impairments, such as advanced amyotrophic lateral sclerosis or severe cerebral palsy. Phase I and Phase II STTR efforts created a prototype commercial-grade AAC-BCI peripheral for the Prentke Romich Company’s (PRC’s) extensive language application and communication system product line.

In accordance with PAR-20-130, the SBIR/STTR Commercialization Readiness Pilot (CRP) Program for Technical Assistance and Late Stage Development, the overall objective of this award is to complete late-stage development of the AAC-BCI, optimize the design for long-term use by target users, finalize regulatory and reimbursement pathways, evaluate train-the-trainer resources, and conduct a small clinical trial of in-home product use to plan future, larger clinical trials. This object will be accomplished by the following specific aims:

• Aim 1: Optimize the AAC-BCI design with regard to long-term use with target user populations.

Iterative cycles of laboratory testing and software development will optimize the AAC-BCI. Optimization will include an automatic standby mode which the user can independently exit when communication is desired.

• Aim 2: Optimize AAC-BCI design for regulatory and reimbursement considerations.

Risk Analysis, laboratory testing, iterative engineering refinements and documentation will ensure that the AAC-BCI qualifies for medical device reimbursement by the Centers for Medicaid and Medicare Services (CMS), meets FDA Class II (Exempt) Medical Device requirements, and meets corresponding international regulations.

• Aim 3: Evaluate dosage and delivery of AAC-BCI training resources to achieve practitioner competence for clinical services.

AAC-BCI users, support persons and interprofessional practitioners will evaluate training and support resources. PRC consultants and clinicians will participant in trainings that vary in dosage and types, evaluate their experience and satisfaction, and complete a competency task checklist, brief interview, and online survey.

• Aim 4: Clinical trial of in-home AAC-BCI use by people with specific target user characteristics.

PRC consultants and clinicians from Aim 3 trainings will identify potential AAC-BCI users and support people, train them on in-home AAC-BCI set-up and communication and follow monthly to track use and satisfaction.

The innovation of this work is merging BCI access into PRC’s high-efficiency AAC language production designs that match individual ability, need, and preference. The significance is the extension of quality-of-life communication with CMS reimbursement approval to those with the most severe physical impairments, providing an AAC-BCI from an established company with a clinical support network to provide services across the lifespan.

BCIs for Cognitive Testing

Many tests of cognitive ability rely on accurate and precise movements or speech to generate results, creating a false dependence between movement and speech and perceived mental abilities. BCI may allow administration of cognitive tests without requiring physical movement.

BCIs for Efficient Communication

P300 BCIs are functional, but provide only very, very slow communication. Using such BCIs in conjunction with commercial augmentative and alternative communication systems and/or language models could greatly improve the efficiency of BCI communication.

Hold-Release Functionality for a P300 BCI

We developed a novel P300 BCI functionality in which activation and deactivation (hold-release) of a P300 BCI speller can be separately controlled. This allows for control of the duration of activations, faster response time for the deactivations and a more analog-like control than using the traditional P300 BCI method. Dr. Ramses Alcaide, a graduate of the lab, leads a start-up company, Neurable, that is commercializing this groundbreaking BCI functionality for computer games and virtual reality.

No-Control Performance

An important aspect of all interfaces is operation only when the user intends to interact with the interface. BCIs, however, are always in contact with the user. Therefore, a BCI must determine from the user's brain activity whether an activation is intended. Identifying periods in which the user does not intend to use the interface (no-control periods), can improve overall BCI performance accuracy by eliminating errors that may occur due to user distraction or fatigue.

Identifying and Accommodating User Variability

P300 BCIs generally rely on a fixed model of the brain activity related to the target stimulus to identify the user's desired target. However, variations in brain activity may occur naturally due to fluctuations in user attention or mental workload. We have created a classifier based latency estimation (CBLE) method and identified a relationship between variance in the latency of the P300 signals and BCI performance. Accommodation of this variability could improve BCI performance.