An attempt to perform multivariate pattern analysis (MVPA) with neuroimaging data using PRoNTo software.
In this attempt, the classification challenge is to differentiate 2 choices (i. gambles that were accepted and ii. those that were rejected) in a risky decision-making task. See the results presented on the poster below. The task is based on the second fMRI experiment reported here (Supplementary Materials: Methods): https://www.biorxiv.org/content/10.1101/473975v2
The analysis pipeline in Pronto can be mainly separated into 4 steps (The 'scripts' folder includes some sample scripts that are modified from those I used in my analysis).
Step1. (specify) Data & Design
Step2. Feature selection
Step3. Specify Model (choice which algorithm and cross-validation scheme, etc. I used multi-kernel learning here and a customised cross-validation model)
Step4. Run Model
For further details related to the analysis/pipeline and underlying algorithms applied, refer to Pronto's documentation (http://www.mlnl.cs.ucl.ac.uk/pronto/prtdocs.html).
Required software: PRoNTo software (download from http://www.mlnl.cs.ucl.ac.uk/pronto/prtsoftware.html) & Matlab.
The version used in my analysis: v2.1.
