Machine learning classifiers and fMRI: a tutorial overview

By Francisco Pereira, Tom Michael Mitchell, Matthew Botvinick. 2008

1,743 citations recorded by the source corpus.

Interpreting brain image experiments requires analysis of complex, multivariate data. In recent years, one analysis approach that has grown in popularity is the use of machine learning algorithms to train classifiers to decode stimuli, mental states, behaviours and other variables of interest from fMRI data and thereby show the data contain information about them. In this tutorial overview we review some of the key choices faced in using this approach as well as how to derive statistically significant results, illustrating each point from a case study. Furthermore, we show how, in addition to answering the question of 'is there information about a variable of interest' (pattern discrimination), classifiers can be used to tackle other classes of question, namely 'where is the information' (pattern localization) and 'how is that information encoded' (pattern characterization).

Subfields: Cognitive Neuroscience

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

Citations in the corpus

Causal and anti-causal learning in pattern recognition for neuroimaging (2015, background): “As decoding models are becoming ever more popular in the analysis of neuroimaging data [4], the correct interpretation of such models is of considerable importance.”
Revisiting Classifier Two-Sample Tests (2016, background): “Implicitly, binary classiﬁers also perform two-sample tests in algorithms that discriminate data from noise, such as unsupervised-as-supervised learning ( ?”
Comparison of multivariate classifiers and response normalizations for pattern-information fMRI (2010, positive): “Pattern-classifier analysis requires many decisions (Pereira et al., 2009).”
Clinical applications of the functional connectome (2013, positive): “All of these methods must be run inside crossvalidation (CV) to avoid overfitting and optimistic estimates of the ability of the model to generalize (Pereira et al., 2009).”
Using deep learning to investigate the neuroimaging correlates of psychiatric and neurological disorders: Methods and applications (2017, neutral): “functions (kernels) that map the original data into a new feature space where the data become linearly separable (Pereira et al., 2009; Vapnik, 1995).”
A Review of Feature Reduction Techniques in Neuroimaging (2013, methodology): “SVM or RVM) to neuroimaging data (Bray et al. 2009; Cheng et al. 2012; De Martino et al. 2008; Duchesnay et al. 2007; Duchesnay et al. 2004; Duff et al. 2011; Franke et al. 2010; Liu et al. 2012; Lohmann et al. 2007; Mitchell et al. 2004; Mwangi et al. 2012a; Norman et al. 2006; Pereira et al. 2009;”
Multivariate pattern analysis of fMRI: The early beginnings (2012, background): “…of review papers began to appear that presented the principles of MVPA in a clear and accessible manner, further demystifying the basis of this new analytic approach (Haynes and Rees, 2006; Mur et al., 2009; Norman et al., 2006; O'Toole et al., 2007; Pereira et al., 2009; Tong and Pratte, 2012).”
Decoding Dynamic Brain Patterns from Evoked Responses: A Tutorial on Multivariate Pattern Analysis Applied to Time Series Neuroimaging Data (2016, methodology): “1 Common pitfalls The first caveat applies to all studies using classifiers and is well-described in the literature (Kriegeskorte, Lindquist, Nichols, Poldrack, & Vul, 2010; Kriegeskorte, Simmons, Bellgowan, & Baker, 2009; Pereira et al., 2009).”
Neural Evidence for a Distinction Between Short-Term Memory and the Focus of Attention (2011, positive): “Using multivariate pattern analysis (MVPA) of brain activity recorded in event-related fMRI (Pereira, Mitchell, & Botvinick, 2009 ; Haynes & Rees, 2006 ; Norman, Polyn, Detre, & Haxby, 2006 ), we tested the hypothesis that delay period activity reflects the information that is being attended to, ”
Sparse logistic regression for whole brain classification of fMRI data (2010, methodology): “Multivariate pattern recognition (MPR) methods are rapidly becoming a popular tool for analyzing fMRI data (Cox and Savoy, 2003; De Martino et al., 2008; Haynes et al., 2007; Kriegeskorte et al., 2006; Mourao-Miranda et al., 2005; Pereira et al., 2009).”
Human brain mapping: A systematic comparison of parcellation methods for the human cerebral cortex (2017, methodology): “Before the classification step, dimensionality reduction through Principal Component Analysis is performed (Pereira et al., 2009) and the components explaining 100% of the variance (Robinson et al., 2010) in the training data are preserved for both training and testing.”
Confounds in multivariate pattern analysis: Theory and rule representation case study (2013, positive): “We generated single-subject classification accuracy maps using leave-1-run-out cross-validation (e.g., Pereira et al., 2009).”