Classifying the valence of autobiographical memories from fMRI data

Alex Frid; Larry M. Manevitz; Norberto Eiji Nawa

doi:10.1007/s10472-020-09705-3

Classifying the valence of autobiographical memories from fMRI data

Alex Frid, Larry M. Manevitz, Norberto Eiji Nawa

Department of Computer Science

Research output: Contribution to journal › Article › peer-review

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.

Original language	English
Pages (from-to)	1261-1274
Number of pages	14
Journal	Annals of Mathematics and Artificial Intelligence
Volume	88
Issue number	11-12
DOIs	https://doi.org/10.1007/s10472-020-09705-3
State	Published - 1 Dec 2020

Bibliographical note

Funding Information:
The CiNet Research Institute hosted LM (twice) and AF (once) to work with NEN on this work. We thank its staff and especially Director Takahisa Taguchi for their support and encouragement of this work. This work was performed under an MOU between CiNet and the University of Haifa. Some of the computations were performed under an equipment grant of NVIDIA corporation to the Neurocomputation Laboratory at the Cesarea Research Institute, University of Haifa. NEN was partially supported by the Japan Society for the Promotion of Science under a Grant-in-Aid for Scientific Research (JPS KAKENHI grant number JP17K00220).

Funding Information:
The CiNet Research Institute hosted LM (twice) and AF (once) to work with NEN on this work. We thank its staff and especially Director Takahisa Taguchi for their support and encouragement of this work. This work was performed under an MOU between CiNet and the University of Haifa. Some of the computations were performed under an equipment grant of NVIDIA corporation to the Neurocomputation Laboratory at the Cesarea Research Institute, University of Haifa. NEN was partially supported by the Japan Society for the Promotion of Science under a Grant-in-Aid for Scientific Research (JPS KAKENHI grant number JP17K00220).

Publisher Copyright:
© 2020, Springer Nature Switzerland AG.

Keywords

Analysis of cognitive processes
Autobiographical memories
Classification
Feature selection
Machine learning

ASJC Scopus subject areas

Artificial Intelligence
Applied Mathematics

Access to Document

10.1007/s10472-020-09705-3

Cite this

@article{b62663f7c6ec4ebbb28aa6fbf90f786d,

title = "Classifying the valence of autobiographical memories from fMRI data",

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, Feature selection, Machine learning",

author = "Alex Frid and Manevitz, {Larry M.} and Nawa, {Norberto Eiji}",

note = "Funding Information: The CiNet Research Institute hosted LM (twice) and AF (once) to work with NEN on this work. We thank its staff and especially Director Takahisa Taguchi for their support and encouragement of this work. This work was performed under an MOU between CiNet and the University of Haifa. Some of the computations were performed under an equipment grant of NVIDIA corporation to the Neurocomputation Laboratory at the Cesarea Research Institute, University of Haifa. NEN was partially supported by the Japan Society for the Promotion of Science under a Grant-in-Aid for Scientific Research (JPS KAKENHI grant number JP17K00220). Funding Information: The CiNet Research Institute hosted LM (twice) and AF (once) to work with NEN on this work. We thank its staff and especially Director Takahisa Taguchi for their support and encouragement of this work. This work was performed under an MOU between CiNet and the University of Haifa. Some of the computations were performed under an equipment grant of NVIDIA corporation to the Neurocomputation Laboratory at the Cesarea Research Institute, University of Haifa. NEN was partially supported by the Japan Society for the Promotion of Science under a Grant-in-Aid for Scientific Research (JPS KAKENHI grant number JP17K00220). Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.",

year = "2020",

month = dec,

day = "1",

doi = "10.1007/s10472-020-09705-3",

language = "English",

volume = "88",

pages = "1261--1274",

journal = "Annals of Mathematics and Artificial Intelligence",

issn = "1012-2443",

publisher = "Springer Netherlands",

number = "11-12",

}

TY - JOUR

T1 - Classifying the valence of autobiographical memories from fMRI data

AU - Frid, Alex

AU - Manevitz, Larry M.

AU - Nawa, Norberto Eiji

N1 - Funding Information: The CiNet Research Institute hosted LM (twice) and AF (once) to work with NEN on this work. We thank its staff and especially Director Takahisa Taguchi for their support and encouragement of this work. This work was performed under an MOU between CiNet and the University of Haifa. Some of the computations were performed under an equipment grant of NVIDIA corporation to the Neurocomputation Laboratory at the Cesarea Research Institute, University of Haifa. NEN was partially supported by the Japan Society for the Promotion of Science under a Grant-in-Aid for Scientific Research (JPS KAKENHI grant number JP17K00220). Funding Information: The CiNet Research Institute hosted LM (twice) and AF (once) to work with NEN on this work. We thank its staff and especially Director Takahisa Taguchi for their support and encouragement of this work. This work was performed under an MOU between CiNet and the University of Haifa. Some of the computations were performed under an equipment grant of NVIDIA corporation to the Neurocomputation Laboratory at the Cesarea Research Institute, University of Haifa. NEN was partially supported by the Japan Society for the Promotion of Science under a Grant-in-Aid for Scientific Research (JPS KAKENHI grant number JP17K00220). Publisher Copyright: © 2020, Springer Nature Switzerland AG.

PY - 2020/12/1

Y1 - 2020/12/1

N2 - 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.

AB - 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.

KW - Analysis of cognitive processes

KW - Autobiographical memories

KW - Classification

KW - Feature selection

KW - Machine learning

UR - http://www.scopus.com/inward/record.url?scp=85088994860&partnerID=8YFLogxK

U2 - 10.1007/s10472-020-09705-3

DO - 10.1007/s10472-020-09705-3

M3 - Article

AN - SCOPUS:85088994860

SN - 1012-2443

VL - 88

SP - 1261

EP - 1274

JO - Annals of Mathematics and Artificial Intelligence

JF - Annals of Mathematics and Artificial Intelligence

IS - 11-12

ER -

Classifying the valence of autobiographical memories from fMRI data

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this