2018 Projects

A BIDS app for de-identification of neuroimaging data

Add a validate_sequences() module to pybids that creates user-friendly summary data frames to check whether a BIDS data set: 1. Contains any files with duplicate content. 2. Is missing files that are expected to exist. 3. Has more files than are expected for a given subject.

How old is your brain? The current project implements Support Vector Regression (SVR) to predict #brain-age from structural MRI and resting-state fMRI separately in neurotypical and clinical populations.

Train a simple convolutional neural network to classify brain CT images has having an intracranial hemorrhage or not. Determine and select usable scans and organize them in the BIDS format

Expanding and creating tutorials for existing Neuropointillist project; adding some code to facilitate Neuropointillist use

A meta-project for examining Neurohackademy by the (wildly oversimplified and non-representative) numbers

Loops are Evil!? Avoid Loops!? We listen or read this affirmation everywhere. The goal is to learn Numpy advanced features and see the benefit/drawback. We decide to implement a denoising algorithm (non local means – 6 loops) without any loop.

NeuroImage is a next-gen web viewer for visualization of neuroimaging data. Or at least it will be, one day. Maybe.

https://nibetaseries.readthedocs.io/en/latest/

Building a pipeline and tutorial for task fMRI analysis from nistats to nilearn

We dockerized PALS with a GUI, refactored PALS with nipype (in progress), incorporated ni-learn for QC images

A nipype workflow for The Virtual Brain

Open-access articles were scraped from the PubMed Central API. Their full-text was searched for keywords (and phrases) indicative of code and data sharing. The prevalence of code and data sharing (and preregistration) was computed as a measure of the “openness” of the research published in that journal (it’s “O-Factor”).

Create a Wikipedia for academic papers with a discussion forum

Fragments a surface mesh into N-equally sized annotation parcels. Besides looking beautiful, such surface parcellation could be used as feature extract for machine learning or functional connectivity approaches.

Interactive brain viewer and scatterplot visualization of univariate and multivariate neuroimage analyses.

Interactive brain viewer and scatterplot visualization of univariate and multivariate neuroimage analyses.

Broadly, examine how neural representations of characters and emotions are shared between observers, change with experience, and track with other aspects of physiology.

U-nets on T1 to T2, and HCP interactive tool