-
Notifications
You must be signed in to change notification settings - Fork 0
Projects
The goal of the project is to develop sequence-based deep learning models to predict the impact of variation in non coding portions of the genome which shape transcriptional regulation associated with cognitive processes.
With the large number of interactions detected by CNNs, the optimal way to correct for known confounders in brain imagining seems to be within the network itself. This project will focus on experimenting with model architectures on a multi-site dataset.
I would like to adapt mriqc and fmriprep to run with pig anatomical template, and then run resting state functional connectivity analyses. Final goal would be to run a seed-based analysis, determine ROIs for whole brain analyses, and get ROI-ROI correlation matrix from ROIs. Ultimate goal is to test whether we see similar set of canonical networks in pig as human and primate.
The BRAINS AutoWorkup proprietary output structure will better serve the community if it generates BIDS compliant derivative outputs. This project will focus on the efforts necessary to generate BIDS compliant output.
How to organize, normalize, and analyze data from omics level datasets (ie; proteomics, metabolomics, transcriptomics)
NiBetaSeries is BIDS-compatible application that calculates betaseries correlations. In brief, a beta coefficient (i.e. parameter estimate) is calculated for each trial (or event) resulting in a series of betas that can be used to correlate regions of interest with each other. A variety of things could be worked on:
- Least Squares All
- Renaming outputs to match BIDS
- fixing spelling/errors in documentation
- test users
See the issues page for a full listing of things that could be worked on.
- Find data that fit this description, attempt to use the current standard to organize the data the proposed way
- Suggest changes to improve usability