Connectome-based Predictive Modeling with Missing Data

Our manuscript is available on bioRxiv https://www.biorxiv.org/content/10.1101/2023.06.09.544392v1.full.pdf and published at Imaging Neuroscience https://direct.mit.edu/imag/article/doi/10.1162/imag_a_00071/118937

Method

Predictive modeling framework for handling missing data. Missing connectomes are rescued using 1) Task average replacement, 2) Mean imputation, 3) Constant value imputation, 4) Robust Matrix Completion, or 5) Nearest Neighbors imputation. Missing phenotypic measures are imputed with auxiliary variables using 1) Predictive mean matching (PMM), 2) ImputePCA, 3) MissForest, or 4) Mean imputation. The phenotypic and connectivity data are then used in standardized predictive models such as ridge regression or support vector machine. The black squares represent missing data.

Dependencies

Missing Connectome (Python)

• numpy
• scikit-learn

Missing Phenotyes (R)

• glmnet
• cvTools
• missMDA
• missForest
• mice
• VIM
• plyr

Data

The dataset should include both connectome data and phenotypical data. The data should be organized in the same structure as demonstrated in the following scripts of the experiments.

Experiment

You could set up your experiment like the example code below and run:

Experiment 1: CPM with missing connectomes (python)

python Expe_coms_behav.py -sd 10

Experiment 2: CPM with missing connectomes and phenotypes (R)

Rscript Experiment_com_behav.R 10

Reference

Qinghao Liang, Rongtao Jiang, Brendan D. Adkinson, Matthew Rosenblatt, Saloni Mehta, Maya L. Foster, Siyuan Dong, Chenyu You, Sahand Negahban, Harrison H. Zhou, Joseph Chang, Dustin Scheinost; Rescuing missing data in connectome-based predictive modeling. Imaging Neuroscience 2024; doi: https://doi.org/10.1162/imag_a_00071

Please cite our paper if you find this repository useful:

@article{10.1162/imag_a_00071,
    author = {Liang, Qinghao and Jiang, Rongtao and Adkinson, Brendan D. and Rosenblatt, Matthew and Mehta, Saloni and Foster, Maya L. and Dong, Siyuan and You, Chenyu and Negahban, Sahand and Zhou, Harrison H. and Chang, Joseph and Scheinost, Dustin},
    title = "{Rescuing missing data in connectome-based predictive modeling}",
    journal = {Imaging Neuroscience},
    year = {2024}
}