Combining multivariate voxel selection and support vector machines for mapping and classification of fMRI spatial patterns

By Federico De Martino, Giancarlo Valente, Noël Staeren, John Ashburner, Rainer Goebel, Elia Formisano. 2008

542 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Interpreting support vector machine models for multivariate group wise analysis in neuroimaging (2015, neutral): “Machine learning based methods Davatzikos (2004) Cuingnet et al. (2011) De Martino et al. (2008) Fan et al. (2007) Gaonkar and Davatzikos (2012) Xu et al. (2009) Van De Ville et al. (2007) Van De Ville et al. (2004) Mingoia et al. (2012) can offer diagnostics based on neuroimaging data.”
Causal and anti-causal learning in pattern recognition for neuroimaging (2015, background): “permuting or removing Xi from the feature set and testing whether this significantly decreases decoding accuracy [10].”
On the Definition of Signal-To-Noise Ratio and Contrast-To-Noise Ratio for fMRI Data (2013, unknown): “In contrast, an amplitude measurement can be related to the standard deviation of the noise [21–24] and [25],”
Comparison of multivariate classifiers and response normalizations for pattern-information fMRI (2010, unknown): “In the literature, some pattern-information studies used beta estimates to define the response patterns (e.g. Haxby 2001; De Martino et al., 2008) and other studies used t-values (e.g. Martínez- Ramón et al., 2006; Kriegeskorte et al., 2008a).”
Investigating the predictive value of whole-brain structural MR scans in autism: A pattern classification approach (2010, methodology): “There is some evidence that SVM-RFE can produce improvements in prediction accuracy for fMRI data (De Martino et al., 2008), but if used for this purpose, it is necessary to use nested cross-validation to determine the optimal number of features for each LOO-CV fold.”
The Decoding Toolbox (TDT): a versatile software package for multivariate analyses of functional imaging data (2015, methodology): “Recursive feature elimination is another popular method (De Martino et al., 2008) and is sometimes referred to as a wrapper method, although the ﬁnal feature set and classiﬁ-cation can depend on the previous steps.”
Model sparsity and brain pattern interpretation of classification models in neuroimaging (2012, methodology): “We considered two strategies for obtaining sparsity—logistic regression with the ENET penalty and SVM based recursive feature elimination (RFE) [27,57].”
Deconvolving BOLD activation in event-related designs for multivoxel pattern classification analyses (2011, methodology): “One solution to this problem is simply to use a slow event-related design (De Martino et al., 2008).”
Multi-Site Diagnostic Classification of Schizophrenia Using Discriminant Deep Learning with Functional Connectivity MRI (2018, methodology): “In the RFE-SVM and RFE-LDA classi-ﬁ cation frameworks, we used nine folds as the training dataset and one fold as the testing dataset, and the training dataset was further partitioned in nine splits both in the RFE-SVM and RFE-LDA nested cycles (Martino et al., 2008).”
Disease state prediction from resting state functional connectivity (2009, methodology): “Possible reasons for the discrepancy are that the previous study (22) applied feature selection to a dataset with much higher dimensionality.”
A Primer on Pattern-Based Approaches to fMRI: Principles, Pitfalls, and Perspectives. (2015, methodology): “A different approach is to use an algorithm that automatically selects only a subset of voxels for the classification (De Martino et al., 2008).”
Decoding brain states from fMRI connectivity graphs (2011, methodology): “This also allows us to use the large variety of existing pattern recognition methods that operate on vector spaces.”
PRoNTo: Pattern Recognition for Neuroimaging Toolbox (2013, methodology): “Future versions of PRoNTo are likely to comprise methods such as Recursive Feature Elimination (RFE, De Martino et al., 2008) and the searchlight approach.”
Does feature selection improve classification accuracy? Impact of sample size and feature selection on classification using anatomical magnetic resonance images (2012, methodology): “It is also a popular feature selection method for fMRI data (Craddock et al., 2009; De Martino et al., 2008; Marquand et al., 2010).”