Holistic Systems Approach in Investigating Brain Injury Mechanisms
Multidomain Assessments of Sports Brain Injury ‘Dose-Response’ Relationship
The human brain is a complex system. As such, how the brain gets injured and what potential outcomes would arise from such injuries is also not a trivial question to answer. Traditionally, studies have tried to correlate head trauma input with binary concussion or no concussion outcomes, or have used single-domain assessments (e.g., one neurocognitive test) to evaluate the brain injury outcome. Such study designs may not fully capture the complexity of the system. In a collaborative project, we work with researchers from diverse backgrounds to assess sports brain injury using a more holistic approach, with the goal of creating a more systems-level model of the complex ‘biomechanical dose – brain response’ relationship in brain injury.
Head Acceleration Exposure Measurements Across Diverse Sports Populations
Athletes in contact and collision sports can sustain repeated and excessive head accelerations that may lead to acute or long-term brain injury. We have collaborated with UBC Women’s Soccer, Men’s Ice Hockey, Women’s Ice Hockey, and Women’s Rugby varsity teams in studies to measure complete longitudinal head acceleration exposure. Our focus on more diverse sports and on women’s sports populations stem from the need to increase the diversity and inclusion of head impact exposure data in literature, such that findings can be applied to help manage brain injury risks across a wider range of populations. Besides exposure measurements, another research goal is to develop reliable and rigorous methodologies for on-field, real-world head impact exposure measurements, such as our multi-stage ice hockey head acceleration exposure screening process illustrated below. We show here that through each step of the multi-stage screening process, we effectively remove instrumented mouthguard sensor recordings that likely arise from spurious events instead of true head acceleration events, reducing potential false positive findings.
Multiscale Modeling of Brain Deformations using Field Data
While head kinematics parameters such as the linear acceleration and angular acceleration of the head have traditionally been used to estimate brain injury risk, we recognize the complex material behaviour of the brain and have collaborated with experts in brain computational models such as Dr. Songbai Ji’s group from the Worcester Polytechnic Institute (Lab Website) to estimate brain deformation parameters at multiple scales. Such parameters may provide more direct estimate of brain injury risk and enable more insight into the interplay between the mechanics and brain structural / functional outcomes. The video below shows an example concussion impact simulation of brain axonal strain using the Worcester Head Injury Model (WHIM, link to publication). We are also actively working with Dr. Ji’s group in developing deep learning models of axonal injury (recent publication).
Immediate Neurophysiological Effects of Head Impacts
Our lab is developing sensitive and quantitative methods to evaluate the immediate electrophysiological changes after sports head impacts. Our preliminary results have shown that even the mildest sports head impacts could cause transient, subtle electroencephalogram (EEG) changes across the brain, lasting on the order of seconds. Higher severity impacts or higher frequency impacts may lead to higher levels of physiological changes requiring longer recovery time. This is a first investigation to quantify immediate physiological effects of impacts that are traditionally thought to be mild and non-injurious, which could shed light on the mechanisms of brain changes resulting from mild repetitive sports head impact exposure.
Technology Development for Out-of-Clinic Brain Health Monitoring
Validation of Wearable Head Acceleration Sensors for On-Field Measurements
Wearable head impact sensors leverage miniature, low-cost microelectromechanical system (MEMS) inertial measurement units (IMUs) embedded into head-mounted equipment such as helmets and mouthguards to measure up to six degrees of freedom of linear and angular head kinematics. Specifically, the goal of these sensors is to capture skull motion parameters, which can then be used to infer brain injury risks. Despite the promise of these wearable sensors in collecting human data, sensor accuracy limitations arise due to inherent sensor errors, imperfect head-sensor coupling, and complexity of real-world impact conditions. Our group has led recent work in standardizing sensor validation methods (publication), quantification of potential biases in standard trigger methods (publication), and on-field skull-coupling of instrumented mouthguard sensors (publication). The figure below illustrates that based on current linear acceleration triggering methods, head accelerations reaching 30g at the centre of gravity may not always trigger event recording, which could pose substantial biases in field data collection.
Computer Vision Methods for Sports Head Impact Monitoring
Recognizing that with limitations in device impact detection performance and challenges in full compliance, it is unlikely for a purely sensor-based approach to capture full impact exposure, my group has also developed computer vision methods to use sports video for more comprehensive impact exposure capture, starting with soccer (publication). Since sports videos are generally already available and/or more easily obtained with lower resource requirements, further work on computer vision methods for impact and contact detection will contribute more feasible methods for complete longitudinal exposure estimates.
Mobile Brain and Body Imaging (MoBI) Systems
Mobile brain/body imaging (MoBI) is an emerging research approach that leverages technological advancements in wearable technologies for brain and body sensing in dynamic environments. Despite some pilot work in this area, there is a substantial gap between currently available technologies and actual out-of-the-laboratory data collection, due to technological limitations, usability issues, data quality concerns, complexity of real-world stimuli, limited sensing modality, and high individual variability. We are interested in developing and applying MoBI technology to study brain processes and brain health in everyday, natural scenarios. We have recently quantified variability in electroencephalogram (EEG) measurements of brain activity in common everyday activity scenarios (publication) and also done work to compare EEG electrode performance during ambulatory activities (publication).
Sensing Sleep Using Wearable Sensors
Sleep is necessary for human function. Sleep disturbances and disorders are associated with a multitude of neuropsychiatric and neurodegenerative conditions. Investigating these disturbances has historically relied on polysomnography (PSG), a comprehensive technology that leverages multiple sensors to assess multiple systems at once. However, this approach can be expensive, uncomfortable for participants, and time consuming. Wearable technologies, such as wearable electroencephalography (EEG) and actigraphy, are promising supplements to PSG. In particular, the emergence of wearable EEG provides the unique opportunity to directly investigate the brain over multiple nights in the comfort of participant’s own homes. We use these wearable EEG technologies to collect brain activity data during sleep periods. We then apply existing and novel signal processing methods to quantify sleep quality. In addition to wearable EEG, we use a 3-axis inertial measurement unit (IMU) to monitor activity levels longitudinally over multiple days. One of our current projects using these technologies is to investigate how shift work impacts sleep quality in a nurse population.
