Decoding Cognitive States from fMRI Data Using Single Hidden-Layer Feedforward Neural Networks

The development of functional magnetic resonance imaging (fMRI) offers promising approaches in the study of human brain function. It dramatically improves an ability to collect large amount of data about brain activity in human subjects performing tasks. Analysis of fMRI is essential for successful detection of cognitive states. This paper presents the use of single hidden-layer feedforward neural networks (SLFNs) to decode cognitive states from fMRI data. The SLFNs are trained by an improved extreme learning machine (ELM) which is named as regularized least-squares ELM (RLS-ELM). Experimental results show that the proposed method can give better performance compared to the Gaussian Naive Bayes (GNB) classifier that is known as one of the best classifiers for decoding cognitive states.