I am halfway through my internship, and yet I feel like I have so much still left to explore and accomplish. This is my first time in a research environment that is not directly relevant to my main focus of computer science, however computer science plays an integral role in the research I’m currently a part of. Our center studies the neurological consequences of neuropathic pain through quantitative analysis of the sciatic nerve, and the technical side makes it possible to conduct our research. This reinforces the idea of how my areas of interest computer engineering can be applied to enhancing and that sometimes it’s just a matter of creativity.
One of the ways that this experience has impacted me is the way I approach computer science related problems. In contrast to the methodical approach to solving a problem taught in classes, I am learning that not every factor can be put into a simple sequence of steps. Rather, research is complex and it is difficult to consider all the factors that can affect our approach – unlike controlled scenarios in school when multiple factors are ignored for simplicity. There is no set algorithm to solve something, especially in the biomedical field. While research can be fun and serve as an outlet for creativity, it can also be quite frustrating when you have to work on the same thing for long periods of time often just trying to correct a mistake.
Most of the work I’m doing so far pertains to diffusion tensor imaging. One of the things that I have had the opportunity to learn about is the different algorithms out there in implementing diffusion tensor imaging. Often with MRI scans in the lower extremities, many artifacts can produce lower image quality can make MRI scans more difficult to analyze. For example moving blood vessels might create a strip of noise and blur out the image in certain areas, which is especially true for axial weighted T2 images. One way I have learned to get around this issue is to use diffusion weighted imaging.
Water molecules in the body go through random motion, and by applying a special diffusion from encoding gradients, the MRI can now be sensitive towards this motion. This is known as a diffusion weighted image. When we apply the diffusion gradients, it is necessary to calculate a diffusion tensor to each pixel in the image. After extensive calculations, you get color coded maps that describe the diffusion anisotropy which provides a better idea of the different nerves in the body. This technique is used to grasp a better understanding of the white matter tracts in the brain, and in the study that I’m a part of this method is being applied to see if it is an efficient neuroimaging technique for the lower extremities.
The coolest part of my internship is that nobody has really attempted to improve neuroimaging techniques to capture the nerves below the spinal chord. While this is exciting, there are unique challenges I face since very little research exists for me to draw on where someone has attempted a similar approach. Overall, this experience has introduced me what research in academia actually entails – both the advantages and disadvantages that research poses – and has introduced me to new ways to think about how to use technology to develop novel approaches to solve problems.
To learn more about the research: