The Application of kNN and SVM in the Decoding of fMRI Data

DOI

10.2991/aiie-16.2016.78How to use a DOI?

Keywords

functional magnetic resonance imaging; support vector machine; k-nearest neighbor

Abstract

For the decoding analysis of functional magnetic resonance imaging (fMRI) data, the appropriate method for feature selection and classification algorithm was a core issue. Given the high dimensionality of fMRI data in whole brain, the localization of regions of interest (ROI) usually was used to select voxels relevant to task, which was based on the physiological evidence. K-nearest neighbor (kNN) and linear support vector machine (SVM) were performed to classify the fMRI data. The result indicated that the method of ROI could indeed select voxels involved in the recognition task. The accuracy of ROI relevant to task was significantly higher than chance level and different stimuli could be decoded successfully. Additionally, by the comparison of kNN and SVM, the performance of SVM was better than that of kNN on the whole.

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Fangyuan Ma
AU  - Junhai Xu
PY  - 2016/11
DA  - 2016/11
TI  - The Application of kNN and SVM in the Decoding of fMRI Data
BT  - Proceedings of the 2016 2nd International Conference on Artificial Intelligence and Industrial Engineering (AIIE 2016)
PB  - Atlantis Press
SP  - 343
EP  - 345
SN  - 1951-6851
UR  - https://doi.org/10.2991/aiie-16.2016.78
DO  - 10.2991/aiie-16.2016.78
ID  - Ma2016/11
ER  -