question about fMRI min dreading data (dataTrnS1)

I have downloaded “dataset vim-1 (visual imaging - 1)”.I want to use this data set for classification methods.

As I understood from the dataTrnS1, there are 1750 distinct images presented to subject at training stage and their corresponding voxel responses are recorded. So without any classification we have 1750 voxel patterns.  how can I define distinct categories or voxel response patterns as different classes? Then I can label different and limit class number to a group of patterns at dataTrnS1. Can I use the “presented image types” as labels to the classes?

regards,

Mary

Hi Mary, 

We have provided the stimuli along with the data set, and the stimuli were not chosen to belong to any particular categories. Thus it's up to you to determine what categories you think exist in the stimulus set if you want to do any kind of classification using this data. A word of warning, though - there are better data sets out there to use if you are just learning how to do machine learning. 

Mark