Journal article icon

Journal article

Standardized low-resolution brain electromagnetic tomography does not improve EEG Alzheimer's disease assessment

Abstract:
Quantitative EEG has been shown to reflect neurodegenerative processes in Alzheimer's disease (AD) and may provide non-invasive and widely available biomarkers to enhance the objectivization of disease assessment. To address EEG's major drawback - its low spatial resolution - many studies have employed 3D source localization. However, none have investigated whether this complex mapping into 3D space actually adds value over standard surface derivation. In fact, we found no prior study - in any disease - that quantitatively compared the results of a 3D source localization method with those achieved by surface derivation. We analyzed data from one of the largest prospective AD EEG studies ever conducted (four study centers, 188 patients, 100 female). Thousands of distinct quantitative EEG markers of slowing, complexity, and functional connectivity were computed and regressed against disease severity, with rigorous control for multiple testing. We found highly significant associations between quantitative EEG markers and disease severity. However, standardized low-resolution electromagnetic tomography (sLORETA), a widely used 3D source localization method, did not improve results. Furthermore, a surface derivation marker (auto-mutual information of the left hemisphere during the eyes-closed condition) was the best performing marker across our entire sample. While our findings strongly support that quantitative EEG markers reflect neurodegenerative processes in AD, they do not demonstrate additional benefit from sLORETA. Importantly, our results are specific to AD and sLORETA. Therefore, they should not be generalized to other neurological or psychiatric disorders or to other 3D source localization methods without further validation. Finally, these findings do not diminish the value of 3D source localization for visual EEG inspection.
Publication status:
Published
Peer review status:
Peer reviewed

Actions

Access Document

Publisher copy:
10.1016/j.neuroimage.2025.121144

Authors

More by this author
Institution:
University of Oxford
Research group:
Machine Learning Research Group
Role:
Author
More by this author
Role:
Author
ORCID:
0000-0001-9727-7580


More from this funder
Funder identifier:
https://ror.org/028jc0449
Grant:
827462


Publisher:
Elsevier
Journal:
NeuroImage More from this journal
Volume:
310
Article number:
121144
Place of publication:
United States
Publication date:
2025-03-14
Acceptance date:
2025-03-13
DOI:
EISSN:
1095-9572
ISSN:
1053-8119
Pmid:
40090555


Language:
English
Keywords:
Pubs id:
2096817
Local pid:
pubs:2096817
Deposit date:
2025-04-15
ARK identifier:

Terms of use


Views and Downloads






If you are the owner of this record, you can report an update to it here: Report update to this record

TO TOP