Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
mrtools:tutorialsclassify [2015/10/13 00:38]
steeve
mrtools:tutorialsclassify [2015/10/28 00:30] (current)
justin
Line 3: Line 3:
 This tutorial explains classification analysis with fMRI data using mrTools. Classification analysis can be used to examine distributed patterns of activity - often subtle responses that are hard to examine with other methods. For example, the responses to different orientations which might be distributed across the cortex (see: [[http://​www.jneurosci.org/​content/​31/​13/​4792.long|Freeman et al]] and [[http://​www.cell.com/​neuron/​abstract/​S0896-6273(06)00586-1|Sasaki et al]] for how they are distributed). [[http://​www.nature.com/​neuro/​journal/​v8/​n5/​full/​nn1444.html|Kamitani and Tong]] showed that you could use classification techniques to read-out the orientation representation. That is, they build classifiers on a subset of data collected as subjects viewed gratings of different orientations and then tested that classifier with a left-out set of data. They were able to show that they could correctly determine what stimulus orientation subjects had been looking at, just by looking at their fMRI measured brain activity. Here we will go through what you need to know to do these types of analysis. The tutorial has one section on basics about classification (not necessarily about fMRI) and another section which uses actual fMRI to do a very simple classification analysis (was a stimulus presented above or below the fixation point?​). ​ This tutorial explains classification analysis with fMRI data using mrTools. Classification analysis can be used to examine distributed patterns of activity - often subtle responses that are hard to examine with other methods. For example, the responses to different orientations which might be distributed across the cortex (see: [[http://​www.jneurosci.org/​content/​31/​13/​4792.long|Freeman et al]] and [[http://​www.cell.com/​neuron/​abstract/​S0896-6273(06)00586-1|Sasaki et al]] for how they are distributed). [[http://​www.nature.com/​neuro/​journal/​v8/​n5/​full/​nn1444.html|Kamitani and Tong]] showed that you could use classification techniques to read-out the orientation representation. That is, they build classifiers on a subset of data collected as subjects viewed gratings of different orientations and then tested that classifier with a left-out set of data. They were able to show that they could correctly determine what stimulus orientation subjects had been looking at, just by looking at their fMRI measured brain activity. Here we will go through what you need to know to do these types of analysis. The tutorial has one section on basics about classification (not necessarily about fMRI) and another section which uses actual fMRI to do a very simple classification analysis (was a stimulus presented above or below the fixation point?​). ​
  
-You can download ​the data for the second part here. Its a big file, so it may take some time to download.+To run the tutorial you will need to have some matlab files, that you get by downloading ​the following ​file [[http://​gru.stanford.edu/​pub/​classification.tar.gz|classification.tar.gz]]. After downloadingyou should be able to either double-click the file to make it into a directory, or from the command line do:
  
-You can download the tutorial files from:+  gunzip classification.tar.gz 
 +  tar xfv classification.tar
  
-[[http://​gru.stanford.edu/​pub/​classificationTutorial.tar.gz|classificationTutorial.tar.gz]]+And then add it to your path in matlab.
  
-Note that this is a rather ​big file, approximately 430MB.+  >> addpath(genpath('​~/​directoryWhereYouDownloaded/​classification'​));​ 
 + 
 +This is all you need to do for the first part of the tutorial.  
 + 
 +If you want to try the second part of the tutorial where you test some of the ideas on real data, then you can download the datahere. Its a big file, so it may take some time to download (approximately 430MB). 
 + 
 +[[http://​gru.stanford.edu/​pub/​classificationTutorial.tar.gz|classificationTutorial.tar.gz]]
  
 The files are provided as a tar/zip file. In Mac OS X you should be able to just click to open the files in the Finder. Otherwise, you can go to a terminal and do: The files are provided as a tar/zip file. In Mac OS X you should be able to just click to open the files in the Finder. Otherwise, you can go to a terminal and do:
Line 15: Line 22:
   gunzip classificationTutorial.tar.gz   gunzip classificationTutorial.tar.gz
   tar xvf classificationTutorial.tar   tar xvf classificationTutorial.tar
- 
  
 You should familiarize yourself with the basics of data analysis in mrTools by doing the [[:​mrTools:​tutorialsRetinotopy|retinotopy tutorial]] and the [[:​mrTools:​tutorialsEventRelated|event related tutorial]] before going through this tutorial. You should familiarize yourself with the basics of data analysis in mrTools by doing the [[:​mrTools:​tutorialsRetinotopy|retinotopy tutorial]] and the [[:​mrTools:​tutorialsEventRelated|event related tutorial]] before going through this tutorial.
- 
-Finally, you will need to download and put the following matlab functions into your path: 
- 
-  svn checkout http://​gru.stanford.edu/​svn/​matlab/​classification classification 
  
 Some documentation for these functions can be found [[:​mrtools:​classificationtools|here]]. Some documentation for these functions can be found [[:​mrtools:​classificationtools|here]].