Done and To Do
- Set up a meeting with Jaan about clinicalBERT. Largely addresses unstructured data such as physician notes included as part of a patient’s EHR.
- Got a general understanding of the BHS data that we will have access to from CHOP. Ran also sent along a dictionary that gives a bit more detail. However, he confirmed with me this week that we should wait until the next meeting to discuss data access. Based on the information that I have (from the paper and the dictionary), I can put together a project proposal that we can discuss in the group meeting.
- Results from predictive modeling of hbn phenotypic data (including feature importance) are updated here. Next steps on this front is to identify a core (uncorrelated) set of questionnaires that have predictive power. This has potential to be of practical use.
- Didn’t get around to starting on naturalistic data from HBN - that will a task for next week.
- All research assistants have started on their projects and are set up on OpenMind
- I want to adopt some of the methodology from this paper, which I summarize here to the hbn data, particularly the application of ANNs to both cognitive profiles and morphological and connectivity data