2007-2010
[Holy Grail]
Model interoperability and reuse has been called the Holy Grail of biosimulation and promises to make the development and management of complex biological models significantly more efficient. I believe this technology will provide a quantum leap for the biosimulation field, allowing modelers to create more complex, sophisticated, powerful models for use in drug discovery, clinical applications and biological research.
As a doctoral student from 2006-2010, I worked with members of the UW faculty on an ontology-based, semantic simulation model description format called SemSim. Casting models in this format provides a first step towards modular recombination of models and general model interoperability. For my dissertation work I developed and tested a Java-based software tool called SemGen which helps automate the creation, composition and decomposition of SemSim models. I demonstrated SemGen’s utility as a multi-scale, multi-domain model composition tool by using the software to successfully perform two demonstrations of modular modeling. You can download my dissertation here.
As a grad student I also managed the Virtual Human Research Seminar, a cross-cutting colloquium that addressed topics like biosimulation, clinical informatics, bioinformatics, systems biology, ontology, and digital visualization from September 2006 through December 2007.
2005-2007
[El Dorado]
After the Virtual Soldier Project folded in 2005, I focused on creating a model-based method for estimating patient-specific hemodynamics during life-threatening hemorrhage. This research was largely in the spirit of the Soldier Project in that the goal was to leverage computational power and provide a more effective triage tool for mass casualty scenarios. The method I devised would provide first responders with additional information about “what was happening inside the patient.”
The research culminated in a paper that was published in the journal Cardiovascular Engineering (click here to get the pdf). It details my method for automatically tuning a cardiovascular model to match “normal” (baseline) patient-specific hemodynamics, along with a model that estimates their cardiac output and a host of other hemodynamics from heart rate and blood pressure data recorded externally.
This estimation method falls within a family of techniques aimed at deriving cardiac output from blood pressure measurements. Donald A. McDonald, in his classic textbook Blood Flow in Arteries, wrote “the discovery of a technically simple and non-traumatic way of estimating the output of the heart per beat is something of an El Dorado,” (quoted in Kouchoukos et al. 1970) and many researchers have devised such methods. However, as far as I know, none are in widespread clinical use. Time will tell if anyone can make use of my method, which although not technically simple, does not require an invasive blood pressure measurement.
The cardiac output estimation model described in the paper can be viewed and downloaded for free here.
I became a graduate student in 2006, and most of this work on estimating hemodynamics during hemorrhage was completed prior to that, but I continued to edit the manuscript and pursue its publication during my first school year. I would like to continue this research thread in the future.
Also in 2006 I collaborated with Roy Kerckhoffs (University of California San Diego) on coupling his finite-element ventricular model with my closed-loop lumped parameter circulatory system. This research resulted in a novel, closed-loop multiscale model of the cardiovsacular system. Roy wrote a great paper about it which was published in the Annals of Biomedical Engineering in January 2007.
2003-2005
[Virtual Soldier]
In the fall of 2003 I was hired as a Research Scientist in Jim Bassingthwaighte’s lab at the University of Washington. Jim’s lab had recently been funded to participate in a new DARPA program called the Virtual Soldier Project. The goal of the Project was to radically advance combat triage care by providing field medics with a more accurate means of assessing a soldier’s survivability following injury.
The idea went something like this. A detailed electronic medical record for an individual soldier would be created and stored on a small, dogtag-shaped device called a P-Tag. Included in this record would be a complete 3D CT scan of the soldier detailing all of their internal anatomy. Also included would be a mathematical model simulating that soldier’s physiology (blood pressure, heart rate, etc.). If that soldier is wounded in combat and must be triaged, the field medic can take the P-Tag off the soldier, plug it into their hand-held PDA-type computer, tell the computer the extent of the injuries, and the computer would return a prognosis and an estimated time to death. The estimate would help the medic triage soldiers more effectively.
I worked on creating the physiology simulation models developed as part of this Project, many of which are free for download at Physiome.org. They are largely based on previously published works by various titans in the simulation field.
In the eyes of most of the participants, the Soldier Project was a success, although Phase II was never funded. Even though it was shown that the time to death predictions using our system were more accurate than standard combat care techniques, in DARPA’s eyes the difference was not great enough to warrant further funding.
The international research team brought together for the Project worked quite well together, and consisted of some truly excellent scientists and physicians. I was privileged to have the chance to meet and work directly with these elite professionals, and my experience on the Project gave me the confidence and motivation to further pursue my own research in the field of medical simulation.
1999-2002
[Brain Lab]
I graduated with my BS in Biology from Case Western Reserve University (CWRU) in the summer of 1999. My first job out of college was as a research assistant/lab manager in Dr. Joseph C. LaManna’s lab in the CWRU Department of Anatomy. I spent three years in the LaManna lab and learned many ins and outs about the academic biomedical research world. Aside from my managerial duties I was eventually able to help with studies on hypoxia, hypercapnia or ischemia in the brain and conduct some research myself. My own research focused on locating chemosensitive areas in the rat brainstem using the pH-sensitive dye Neutral Red. I developed a new lab protocol for reconstructing 3-D psuedocolor volumes of intracelllular pH in the rat brainstem using digital image processing. I was a co-author on the paper in which the results were published.
I left the LaManna lab in December 2002 when I decided to move to Seattle. My three years in this research group were formative. By putting me in touch with Dr. Jim Bassingthwaighte at the University of Washington, Dr. LaManna helped lead me down my current research path.