Classifying the valence of autobiographical memories from fMRI data
- PDF / 837,800 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 96 Downloads / 183 Views
Classifying the valence of autobiographical memories from fMRI data Alex Frid 1,2
& Larry M. Manevitz
1,3
& Norberto Eiji Nawa
4,5
# Springer Nature Switzerland AG 2020
Abstract
We show that fMRI analysis using machine learning tools are sufficient to distinguish valence (i.e., positive or negative) of freely retrieved autobiographical memories in a cross-participant setting. Our methodology uses feature selection (ReliefF) in combination with boosting methods, both applied directly to data represented in voxel space. In previous work using the same data set, Nawa and Ando showed that whole-brain based classification could achieve above-chance classification accuracy only when both training and testing data came from the same individual. In a cross-participant setting, classification results were not statistically significant. Additionally, on average the classification accuracy obtained when using ReliefF is substantially higher than previous results - 81% for the within-participant classification, and 62% for the cross-participant classification. Furthermore, since features are defined in voxel space, it is possible to show brain maps indicating the regions of that are most relevant in determining the results of the classification. Interestingly, the voxels that were selected using the proposed computational pipeline seem to be consistent with current neurophysiological theories regarding the brain regions actively involved in autobiographical memory processes. Keywords Analysis of cognitive processes . Autobiographical memories . Classification . Machine learning . Feature selection
* Alex Frid [email protected]
1
Neurocomputation Laboratory, Department of Computer Science, University of Haifa, Haifa, Israel
2
The Laboratory of Clinical Neurophysiology, Technion Faculty of Medicine, Haifa, Israel
3
Department of Computer Science, Ariel University, Ariel, Israel
4
Center for Information and Neural Networks (CiNet), National Institute of Information and Communications Technology (NICT), Tokyo, Japan
5
Graduate School of Frontiers Biosciences, Osaka University, Suita, Japan
Frid A. et al.
1 Introduction Memory encoding and retrieval are arguably two of the most complex cognitive processes performed by humans [1, 2]. Study of this process is a central concern of psychology and memory researchers. A technological window on cognitive activities in general, and memory in particular, is the use of neuroimaging techniques to help elucidate the neurophysiological basis underlying such memory processes. The combination of neuroimaging technology with machine learning techniques [3–6] has opened a promising front in the past decade or so. A common approach has been to attempt to identify cognitive processes, states or disorders from neuroimaging data using various types of machine learning techniques. Although progress has been made in the field, the subtleties of memory processes pose a considerably more challenging task, as compared to tasks involving, for instance, the perception of visual stimuli.
Data Loading...
