Skip to content
No description, website, or topics provided.
Branch: master
Clone or download
Latest commit 920a9aa Jun 7, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
cpp Update README.md Jun 7, 2019
images resized image Sep 24, 2018
matlab Update cpm.m Jun 3, 2019
python add python pipeline May 26, 2019
README.md Update README.md Jan 22, 2019

README.md

Connectome-based Predictive Modelling

Connectome-based Predictive Modeling (CPM) is a data-driven protocol for developing predictive models of brain-behavior relationships from connectivity data using cross-validation. The protocol includes the following steps: 1) feature selection, 2) feature summarization, 3) model building, and 4) assessment of prediction significance.

In this repository are MATLAB scripts which can be used to carry out CPM. Currently the protocol resembles the methodology in Shen et al. [1], and Finn et al. [2]. More features and flexibility will be added in the near future. One can exectute the code via the following:

  1. Using the link in the top right corner of this page, download the repository as a zip, and unzip in the desired folder, or clone it to your local computer using the following command:
git clone https://github.com/YaleMRRC/CPM.git.
  1. In MATLAB add the path to your script or in your workspace:
addpath('localpath/CPM/matlab/')
  1. Load in your explanatory data, which can be two or three dimensional (in the case of a series of connectivity matrices), along with the data you wish to model (perhaps a behavioural score).
  2. Run the following:
[y_predict, performance] = cpm_main(x,y)

This will run split half cross validation, and use a p threshold of 0.01 to select features. These default values can be changed, for example:

[y_predict, performance] = cpm_main(x,y,'pthresh',0.05,'kfolds',10)

For further questions please raise an issue here.

References

[1] Shen X., Finn E.S., Scheinost D., Rosenberg M.D., Chun M.M., Papademetris X., Constable R.T. "Using connectome-based predictive modeling to predict individual behavior from brain connectivity.", Nature Protocols 12.3 (2017): 506.

[2] Finn E.S., Shen X., Scheinost D., Rosenberg M.D., Huang J., Chun M.M., Papademetris X., Constable R.T., et al. "Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity.", Nature Neuroscience 18.11 (2015): 1664.

You can’t perform that action at this time.