CaltechAUTHORS
  A Caltech Library Service

Data Augmentation for Deep-Learning-Based Electroencephalography

Lashgari, Elnaz and Liang, Dehua and Maoz, Uri (2020) Data Augmentation for Deep-Learning-Based Electroencephalography. Journal of Neuroscience Methods, 346 . Art. No. 108885. ISSN 0165-0270. https://resolver.caltech.edu/CaltechAUTHORS:20200811-085708957

[img] PDF - Submitted Version
See Usage Policy.

2754Kb

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20200811-085708957

Abstract

Background: Data augmentation (DA) has recently been demonstrated to achieve considerable performance gains for deep learning (DL)—increased accuracy and stability and reduced overfitting. Some electroencephalography (EEG) tasks suffer from low samples-to-features ratio, severely reducing DL effectiveness. DA with DL thus holds transformative promise for EEG processing, possibly like DL revolutionized computer vision, etc. New method: We review trends and approaches to DA for DL in EEG to address: Which DA approaches exist and are common for which EEG tasks? What input features are used? And, what kind of accuracy gain can be expected? Results: DA for DL on EEG begun 5 years ago and is steadily used more. We grouped DA techniques (noise addition, generative adversarial networks, sliding windows, sampling, Fourier transform, recombination of segmentation, and others) and EEG tasks (into seizure detection, sleep stages, motor imagery, mental workload, emotion recognition, motor tasks, and visual tasks). DA efficacy across techniques varied considerably. Noise addition and sliding windows provided the highest accuracy boost; mental workload most benefitted from DA. Sliding window, noise addition, and sampling methods most common for seizure detection, mental workload, and sleep stages, respectively. Comparing with existing methods: Percent of decoding accuracy explained by DA beyond unaugmented accuracy varied between 8% for recombination of segmentation and 36% for noise addition and from 14% for motor imagery to 56% for mental workload—29% on average. Conclusions: DA increasingly used and considerably improved DL decoding accuracy on EEG. Additional publications—if adhering to our reporting guidelines—will facilitate more detailed analysis.


Item Type:Article
Related URLs:
URLURL TypeDescription
https://doi.org/10.1016/j.jneumeth.2020.108885DOIArticle
https://doi.org/10.31219/osf.io/jm2xuDOIDiscussion Paper
ORCID:
AuthorORCID
Lashgari, Elnaz0000-0002-0302-8098
Maoz, Uri0000-0002-7899-1241
Additional Information:© 2020 Elsevier B.V. Received 21 February 2020, Revised 10 July 2020, Accepted 24 July 2020, Available online 31 July 2020. This publication was made possible in part through the support of a joint grant from the John Templeton Foundation and the Fetzer Institute, and Boston Scientific Corporation. The opinions expressed in this publication are those of the author(s) and do not necessarily reflect the views of the John Templeton Foundation or the Fetzer Institute. This publication was also made possible in part by the support of Boston Scientific Investigator-Sponsored Research Program.
Funders:
Funding AgencyGrant Number
John Templeton FoundationUNSPECIFIED
Fetzer InstituteUNSPECIFIED
Boston Scientific CorporationUNSPECIFIED
Subject Keywords:Electroencephalography; Deep learning; Data augmentation; Review
Record Number:CaltechAUTHORS:20200811-085708957
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20200811-085708957
Official Citation:Elnaz Lashgari, Dehua Liang, Uri Maoz, Data augmentation for deep-learning-based electroencephalography, Journal of Neuroscience Methods, Volume 346, 2020, 108885, ISSN 0165-0270, https://doi.org/10.1016/j.jneumeth.2020.108885. (http://www.sciencedirect.com/science/article/pii/S0165027020303083)
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:104903
Collection:CaltechAUTHORS
Deposited By: Tony Diaz
Deposited On:11 Aug 2020 16:44
Last Modified:27 Aug 2020 22:51

Repository Staff Only: item control page