Solomon Diamond ’97 Th’98 engineers a novel approach to how the brain works.
By Adrienne Mongan
The mystery of how the brain works has long been on the human mind. Imaging technologies have opened the brain to medical and scientific study as never before. But according to one of Thayer School’s newest professors, Solomon Diamond ’97 Th’98, there’s so much more that could be done with the data these technologies produce. Coupling his engineering background with medicine, he is using systems engineering to unite disparate medical data into a working model of brain physiology. If he succeeds, the results may change the way we understand the brain and treat brain diseases and disorders. In the following interview, Diamond explains how he’s engineering a new approach to the workings of the brain.
How did you become interested in studying human brain function?
My interest started while I was a student at Thayer School working on my B.E. project. I designed the IBEX, an exercising machine for bedridden elderly, and after graduation I continued working on it at Synergy Innovations in Lebanon, N.H. While I was conducting clinical trials, I ended up working with a stroke patient. This experience affected me deeply. I was stirred by the devastating impact of stroke on motor function. Consequently, I studied the biomechanics and neurophysiology of stroke for my doctorate at Harvard’s School of Engineering and Applied Sciences. I was particularly interested in investigating the recovery of motor function after stroke and began to focus on the uses of neuroimaging, which measures what is going on in the brain. For my dissertation, I used neuroimaging to examine the effects of hypnosis on motor function and cortical activation in chronic stroke patients.
What are your motivations for modeling brain function?
My three motivations are to improve our understanding of human brain function, advance neuroimaging to improve the diagnosis of neurological diseases, and ultimately customize treatments to each patient’s brain physiology.
What are you trying to achieve by developing these models?
I’m working to help transform large-scale descriptive data into concise results that can be applied to understand the way a complex system functions. For example, clinicians who order tests such as magnetic resonance imaging (MRI), electroencephalogram (EEG), and CT scans understand that the information gathered is underutilized because the interpretation is done manually. I’m trying to add a layer of data synthesis to these important tests to predict the interactions of vascular and neuron activity in the brain. The way to achieve this is through predictive modeling of brain physiology.
For those unfamiliar, what is predictive modeling?
Predictive modeling is the process by which a model is created or chosen to predict the future behavior of a system based on its current state. I’m using mathematical models of neurophysiology combined with neuroimaging data to provide a more comprehensive picture of human brain function. Comparing model predictions with measured data provides a way to test assumptions and hypotheses.
For instance, a human physiology book may contain a detailed description of the vessels that supply blood to the visual cortex of the brain. However, these descriptions do not tell you in a quantitative manner how to predict, in this example of the eye, how the blood vessels will respond to a stimulus. In contrast, if you take an engineering book, a description of a physical system will be accompanied by a series of governing equations that predict how the system will respond to a new stimulus. So combining engineering with medicine provides powerful new tools for interpreting medical imaging data.
So you’re trying to close the gap between the physiology and engineering books?
Yes, essentially that is what I am working to achieve.
What brain functions do you study?
An area of brain physiology that I am particularly interested in is cerebral autoregulation, the ability of the brain to maintain approximately constant blood flow despite changing arterial blood pressure. For example, if you quickly move from a sitting to a standing position, the blood flow to your brain is maintained by a system involving cerebral arteries, which immediately dilate to maintain steady flow to the brain despite the momentary drop in arterial pressure. Failure of this autoregulatory system could result in unconsciousness, brain damage, or death. This system is always in operation, but its function can be damaged by neurodegenerative diseases and external mechanical insults, such as a traumatic brain injury (TBI). Although it remains unproven, failure of cerebral autoregulation resulting from TBI may further impair the recovery process following a mechanical insult. Therefore, greater understanding of cerebral autoregulation might not only enhance our understanding of healthy brain function, but also increase our ability to treat patients with a condition such as TBI.
Traumatic brain injury is unfortunately a common condition among service men and women returning from Iraq, affecting approximately 10% of these veterans. TBI can lead to temporary and potentially permanent deficits in cognitive, physical, and psychosocial function. While the body has a remarkable ability for recovery, impairments in cerebral autoregulation after trauma may limit brain recovery.
Cerebral autoregulation has previously been studied by measuring blood flow and blood pressure with techniques such as transcranial doppler ultrasound, which provides a localized measurement of cerebral blood flow.
My hypothesis is that an autoregulatory deficit may vary in different parts of the brain, depending on the vascular circuit and physical properties. Rather than treat localized measurements of cerebral blood flow as unrelated quantities, my models incorporate the major features of the vascular anatomy and relate the physical properties of the vessels to the dynamic autoregulatory response. I plan to use MRI imaging of the vascular structure and blood flow to tune the model to an individual patient. The result is a concise quantitative assessment of that patient’s cerebral circulation and physiology. This is how my models could help to interpret neuroimaging data.
An advantage of working with models is the ability to predict the system’s response to new situations. In a healthy individual you would expect a robust recovery of blood flow to a sudden drop in blood pressure, and this should bear out in the tuned model prediction without submitting the individual to such a potentially risky test. If a patient’s model predicts a poor recovery to a simulated test, then the model will also give clues about where and why there is a problem.
Given the interdisciplinary nature of your research, who are you collaborating with?
I’m working with the psychiatry, neurology, physiology, and radiology departments at Dartmouth Medical School, and the psychological and brain sciences department and Neukom Institute for Computational Science at Dartmouth. We’ve formed the Dartmouth Brain Imaging Group and meet monthly to apply our individual research to ongoing clinical studies. We look to each other for new insights, and we plan to write research grants together.
Are people elsewhere doing similar work?
The notion of applying an engineering approach to functional neuroimaging is relatively new. Thayer School is uniquely positioned to be a leader in this area because of its inherently interdisciplinary structure.
What challenges have you faced in your research?
A major challenge continues to be capturing the salient features that are needed to understand the data and observations without becoming too complex and rendering the model useless.
What is the timeframe for your work?
The good thing is that my research can be scaled up or down in terms of complexity. I’m predicting that it will take three to five years to create models to understand cerebral autoregulation.
What are the potential clinical applications for your research?
Ultimately, I hope that my research helps in the diagnosis and monitoring of treatment for a variety of brain diseases, especially Alzheimer’s, stroke, and traumatic brain injury. Other outcomes include creating individualized treatments and exploring how brain physiology changes during normal aging.
— Adrienne Mongan is a contributing editor at Dartmouth Engineer.