Week 59

02/05 - 02/09:

Done and To Do

• Re-ran models for reading impairment, specifically new language models. Updated clustering approach to identify subtypes
  • To-do: write-up new results and send to Ola
• Re-ran models for ADHD and ASD analysis – new participant filtering, and new stratification of train/test data - now using more comprehensive set of metrics to evaluate test data.
  • To-do: next week, update manuscript with new results
• Met with Aaron to discuss geospatial analysis of CHOP data, zipcode seems to be an important predictor of suicide attempt. Therefore, I want to do a geospatial analysis of Philadelphia zipcodes to better understand the population makeup of strongly predictive zipcodes as well as information about crime, socioeconomic status, and other environmental variables. I’m envisioning a python function that takes in a given zipcode and returns summary information about the population (gathered from census data). We already have template scripts but I have not yet had time to test it.
• Met with Sheeba and Katerina (co-presenters at OHBM symposium on developing cerebellum) to discuss potential collaborations. We’re compiling a list of available developmental datasets and possible research questions here
• Met with Alina to discuss HBN data – added her to a github repo that Kseniia, Kai, and I worked on last summer to characterize semantic similarity of HBN phenotypic data. Offered to help her think through project ideas + run some simple ML analyses on HBN and ABCD data.