Week 1: May 19 -May 23
Introductions
RCR Training by Dr. Siadat
Teamwork lecture by Dr. Debnath
Research Procedures lecture by Dr. Hanna
Title IX training by Chad Martinez
Went over Cybersecurity fundamentals modules
Read "Transformer Model for Alzheimer’s Disease Progression Prediction Using Longitudinal Visit Sequences".
Searched for more literature involving graph prediction models for Alzheimer's via PubMed.
Started setting up the REU-2025 website.
Searched for more papers on fMRI and GNNs to gain a better understanding on both.
Read "GNNs in neuroscience: graph convolutional networks for fMRI analysis".
Read "BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis".
Created power point presentation
Week 2: May 26 -May 30
Continued literature review & began to focus on papers addressing the missing data problem
Among several other papers, these were my favorites:
Read "Addressing the missing data challenge in multi-modal datasets for the diagnosis of Alzheimer's disease"
Read "Multimodal Data Fusion of Deep Learning and Dynamic Functional Connectivity Features to Predict Alzheimer’s Disease Progression"
Read "Multi-Modal Fusion and Longitudinal Analysis for Alzheimer’s Disease Classification Using Deep Learning"
Read "A Self-guided Multimodal Approach to Enhancing Graph Representation Learning for Alzheimer’s Diseases"
Developed a pipeline plan for the upcoming weeks
Week 3: June 2 -June 6
Began looking at the fMRI dataset and learning how fMRIPrep works
Downloaded all relevant patients data from a filtered TADPOLE dataset
Learned dcm2niix for converting data from ADNI from DICOM to NIFTI
Created a script to convert the subject folders in ADNI to the BIDS folder structure
Started to preprocess with fMRIPrep
Week 4: June 9 -June 13
Began looking for code from research papers.
Found code from "STNAGNN: Data-driven Spatio-temporal Brain Connectivity beyond FC"
fMRIPrep failed and multiple subjects were corrupted
Wrote a script to check which subjects failed partially and fully
Came up with a new method to run fMRIPrep that runs a set of subjects in parallel instead of all at once.
In case of failure only a small amount of subjects will be corrupted instead of all.
Week 5: June 16 -June 20
Created a FC to graph translation feature
Found in interesting paper with code from "FC-HGNN: A heterogeneous graph neural network based on brain functional connectivity for mental disorder identification"
Reviewed code
Modified model to work for AD Prediction
Was able to get results and continued to further tune the model to make it better
Week 6: June 23 -June 37
Continued tuning FC-HGNN's model
Started working on a model from the paper "BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis"
Modified to work for AD Prediction
Worked on fixing the overfitting problem for all models tested
Focused on learning how all of these models work for of the purpose of creating our own model
Week 7: June 30 -July 4
Began working on our own model
Started looking at more literature papers for the purpose of finding Alzheimer's prediction related code and methods
Looked into "Auto-Metric Graph Neural Network Based on a Meta-Learning Strategy for the Diagnosis of Alzheimer’s Disease"
Worked on implementing their tabular data method into our own model
Week 8: July 7 -July 11
Decided to scrap tabular idea for the scope of the program to focus more on the fMRI features
Continued to build and tune model, adding gcn layers and accuracy features
Began working on poster
Began working on paper
Week 9: July 14 -July 18
Continued building and adding features to our model
Worked on tweaking our model parameters to try and get the best results possible for our model
Implemented features to extract figures from our model for our poster
Submitted poster for printing
Continued working on research paper
Week 10: July 21 -July 25
Attended Mid-SURE
Finished research paper
Finished project deliverables