Automated Identification Of Neural Correlates Of Continous Variables
This automated supervised feature selection method aims to identify features that best correlate with a continuous independent variable. It is suitable for use in identifying good features from the EEG (or other neuro-imaging method) that change with respect to 3 or more control classes.
Further details of the operation of the feature selection method can be found in the paper. Which may be obtained here.
If you don’t have access to the above link a pre-print (authors self-archive) version of the paper can be found here.
The code can be downloaded from here.
The feature selection algorithm runs in Matlab. It does not rely on any other external toolboxes for its operation.
Please cite the following paper when using the feature selection method, or a modified version of the feature selection method, in your research.
Daly I, Hwang F, Kirke A, et al. (2014) Automated identification of neural correlates of continuous variables. J Neurosci Methods. doi: 10.1016/j.jneumeth.2014.12.012