Our lab focuses on using quantitative methods, including machine learning, applied to MRI-based measures of brain connectivity networks to understand the mysteries of the human brain. One major interest of our lab is in uncovering connectome-based mechanisms of impairment and recovery after neurological injury or disease, including traumatic brain injury, multiple sclerosis and stroke. If we can understand brain-behavior relationships, we may be able to develop accurate diagnostics, prognostics and individualized therapeutics that can boost recovery after neurological disease or injury.
Some of the current projects we are working on are highlighted below.
Connectomics and Multiple Sclerosis
Disability evolution is highly heterogeneous in multiple sclerosis patients, making it extremely difficult to make short- and long- term predictions of disability. Traditional imaging is not strongly correlated with clinical symptoms, thus resulting in what is known as the clinico-radiological paradox in multiple sclerosis. We are interested in understanding how structural and functional connectomes may be used to better characterize the relationship between the changes in the brain and clinical symptoms, and identifying the regional neurophysiological correlates of disability.
Prediction of Recovery in Chronic Stroke
Neurologists are often faced with the difficult challenge of knowing whether a stroke patient will benefit for a specific treatment. We are interested in developing models using machine learning methods to predict motor function and recovery in chronic stroke patients. Our models show promise when it come to quantitative predictions of post-stroke recovery metrics. We also seek to understand the demographic and neuroimaging features most important in predicting of worsening symptoms, which can then be used to develop personalized therapeutics plans for chronic stroke patients.
Longitudinal Network Properties in Pontine Stroke
Longitudinal changes in functional connectivity in areas of the cortex structurally connected to pontine lesions correlates with motor recovery. We are interested in understanding how functional connectivity patterns in areas that are either structurally connected to or isolated from the lesion change over time. By elucidating longitudinal network properties in areas connected to focal brain lesions, we can then understand how these changes are related to measures of patients’ motor abilities.
Sex and the Brain
Sex differences exist in risk factors, prevalence, clinical diagnosis, symptomatology, disease trajectory, and treatment of many diseases. A thorough understanding of the sex differences that exist in the brains of healthy individuals is crucial for the study of neurological illnesses that exhibit differences between males and females. We are interested in understanding sex differences in neural structure and function in both health and disease using machine learning techniques applied to multi-modal neuroimaging data.
Neural Correlates of Cognition
Neural properties such as anatomical thickness, connectivity, and activation patterns have been linked to behavioural and cognitive function in both humans and animal models. We are interested in identifying structural and functional neural correlates of cognitive function. By understanding the structural and functional connections most important for various cognitive abilities, we can develop therapeutic techniques to enhance cognitive recovery following neurological trauma and/or disease.