Comparative Study of SVM Methods Combined with Voxel Selection for Object Category Classification on fMRI Data
2011

Comparative Study of SVM Methods for fMRI Data Classification

Sample size: 14 publication 10 minutes Evidence: moderate

Author Information

Author(s): Song Sutao, Zhan Zhichao, Long Zhiying, Zhang Jiacai, Yao Li

Primary Institution: Beijing Normal University, China

Hypothesis

This study aims to investigate the overall performance of linear and RBF SVM for fMRI classification together with voxel selection schemes.

Conclusion

The study found that RBF SVM performed better than linear SVM in low-dimensional feature spaces, while linear SVM outperformed RBF SVM in high-dimensional spaces.

Supporting Evidence

  • RBF SVM outperformed linear SVM significantly in low-dimensional feature spaces.
  • Linear SVM performed better than RBF SVM in high-dimensional feature spaces.
  • Voxel selection methods had a significant impact on classification accuracy.

Takeaway

Researchers compared two methods to classify brain images and found one method worked better when using fewer features, while the other was better with more features.

Methodology

The study employed six voxel selection methods and compared the performance of linear and RBF SVM classifiers on fMRI data.

Limitations

The study did not explore the impact of deactivation voxels on classification performance.

Participant Demographics

14 healthy college participants (6 males and 8 females).

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0017191

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication