2. Decoding and MVPA: predicting from brain images

Decoding consists in predicting external variables such as behavioral or phenotypic variables from brain image. It can be useful for diagnostic of prognosis, or to probe the information content of brain activity images.

These are Supervised learning tasks, focused on predicting an output value.

• 2.1. A decoding tutorial
  • 2.1.1. Data loading and preparation
    • 2.1.1.1. The Haxby 2001 experiment
    • 2.1.1.2. Loading the data into Python
  • 2.1.2. Performing the decoding analysis
    • 2.1.2.1. The prediction engine
      • 2.1.2.1.1. An estimator object
      • 2.1.2.1.2. Applying it to data: fit (train) and predict (test)
    • 2.1.2.2. Measuring prediction performance
      • 2.1.2.2.1. Cross-validation
      • 2.1.2.2.2. Choosing a good cross-validation strategy
      • 2.1.2.2.3. Choice of the prediction accuracy measure
      • 2.1.2.2.4. Measuring the chance level
    • 2.1.2.3. Visualizing the decoder’s weights
  • 2.1.3. Decoding without a mask: Anova-SVM
    • 2.1.3.1. Dimension reduction with feature selection
    • 2.1.3.2. Visualizing the results
  • 2.1.4. Going further with scikit-learn
    • 2.1.4.1. Changing the prediction engine
    • 2.1.4.2. Changing the feature selection
• 2.2. Choosing the right predictive model
  • 2.2.1. Predictions: regression, classification and multi-class
    • 2.2.1.1. Regression
    • 2.2.1.2. Classification: two classes or multi-class
  • 2.2.2. Setting estimator parameters
  • 2.2.3. Different linear models
• 2.3. Decoding on simulated data
  • 2.3.1. Simple NeuroImaging-like simulations
  • 2.3.2. Running various estimators
• 2.4. Searchlight : finding voxels containing information
  • 2.4.1. Principle of the Searchlight
  • 2.4.2. Preparing the data
    • 2.4.2.1. Loading
    • 2.4.2.2. Reshaping the data
    • 2.4.2.3. Masking
  • 2.4.3. Third Step: Setting up the searchlight
    • 2.4.3.1. Classifier
    • 2.4.3.2. Score function
    • 2.4.3.3. Cross validation
  • 2.4.4. Running Searchlight
  • 2.4.5. Visualization
    • 2.4.5.1. Searchlight
    • 2.4.5.2. Comparing to massively univariate analysis: F_score or SPM