Ben Weppner / Dr. Hooman Kamel - Week 6
Week 6: 07/10/23-07/14/23
This week, I joined Dr. Judy Ch'ang in the neuro ICU at WCMC. I saw a variety of patients, including those with intracranial aneurysms, small vessel disease, stroke, and cerebral edema. It was interesting to see the improvements in the medical images for patients who had undergone surgery, especially considering how serious the conditions they are suffering from are.
On Tuesday during our weekly meeting, we had a guest presentation by Dr. George Shih regarding the potential impact of large language models in healthcare. Dr. Shih discussed the evolution of large language models, such as ChatGPT, and how they could potentially be used in healthcare. For example, generative AI could produce synthetic patient data that could then be used in research. However, he also warned about the pitfalls of these models, such as hallucinations, that may pose an ethical problem if applied to healthcare.
On Friday, I attended a mentorship lecture by Dr. Elizabeth Hecht. She discussed common traits that make a good mentor and mentee, as well as some common misconceptions about what the relationship between the mentor and mentee may be like. She described the advantages of having a network of mentors instead of a single mentor, and how communicating what the mentee is looking for out of the relationship is important for finding the correct mentor for the mentee.
For research, I worked on extracting arterial input functions (AIFs) from perfusion-weighted images, and then fitting the curves to get a representation of the AIFs as a mathematical equation. This involved using a gamma variate function, which is commonly used for modeling AIFs in dynamic susceptibility contrast (DSC) MRI. I used a variation of the gamma variate function that takes into account the second pass of the tracer through the vasculature in order to more accurately model the AIF.
Comments
Post a Comment