EEG-GAN: Generative adversarial networks for electroencephalograhic (EEG) brain signals

EEG-GAN: Generative adversarial networks for electroencephalograhic (EEG) brain signals. Hartmann, K. G., Schirrmeister, R. T., & Ball, T. June, 2018. arXiv:1806.01875 [cs, eess, q-bio, stat] version: 1

Paper abstract bibtex

Generative adversarial networks (GANs) are recently highly successful in generative applications involving images and start being applied to time series data. Here we describe EEG-GAN as a framework to generate electroencephalographic (EEG) brain signals. We introduce a modification to the improved training of Wasserstein GANs to stabilize training and investigate a range of architectural choices critical for time series generation (most notably up- and down-sampling). For evaluation we consider and compare different metrics such as Inception score, Frechet inception distance and sliced Wasserstein distance, together showing that our EEG-GAN framework generated naturalistic EEG examples. It thus opens up a range of new generative application scenarios in the neuroscientific and neurological context, such as data augmentation in brain-computer interfacing tasks, EEG super-sampling, or restoration of corrupted data segments. The possibility to generate signals of a certain class and/or with specific properties may also open a new avenue for research into the underlying structure of brain signals.

@misc{hartmann_eeg-gan_2018-2,
	title = {{EEG}-{GAN}: {Generative} adversarial networks for electroencephalograhic ({EEG}) brain signals},
	shorttitle = {{EEG}-{GAN}},
	url = {http://arxiv.org/abs/1806.01875},
	abstract = {Generative adversarial networks (GANs) are recently highly successful in generative applications involving images and start being applied to time series data. Here we describe EEG-GAN as a framework to generate electroencephalographic (EEG) brain signals. We introduce a modification to the improved training of Wasserstein GANs to stabilize training and investigate a range of architectural choices critical for time series generation (most notably up- and down-sampling). For evaluation we consider and compare different metrics such as Inception score, Frechet inception distance and sliced Wasserstein distance, together showing that our EEG-GAN framework generated naturalistic EEG examples. It thus opens up a range of new generative application scenarios in the neuroscientific and neurological context, such as data augmentation in brain-computer interfacing tasks, EEG super-sampling, or restoration of corrupted data segments. The possibility to generate signals of a certain class and/or with specific properties may also open a new avenue for research into the underlying structure of brain signals.},
	urldate = {2023-01-15},
	publisher = {arXiv},
	author = {Hartmann, Kay Gregor and Schirrmeister, Robin Tibor and Ball, Tonio},
	month = jun,
	year = {2018},
	note = {arXiv:1806.01875 [cs, eess, q-bio, stat]
version: 1},
	keywords = {Computer Science - Machine Learning, Electrical Engineering and Systems Science - Signal Processing, Quantitative Biology - Neurons and Cognition, Statistics - Machine Learning},
}

Downloads: 0

{"_id":"6PjTbDH2nxenFEARY","bibbaseid":"hartmann-schirrmeister-ball-eeggangenerativeadversarialnetworksforelectroencephalograhiceegbrainsignals-2018","author_short":["Hartmann, K. G.","Schirrmeister, R. T.","Ball, T."],"bibdata":{"bibtype":"misc","type":"misc","title":"EEG-GAN: Generative adversarial networks for electroencephalograhic (EEG) brain signals","shorttitle":"EEG-GAN","url":"http://arxiv.org/abs/1806.01875","abstract":"Generative adversarial networks (GANs) are recently highly successful in generative applications involving images and start being applied to time series data. Here we describe EEG-GAN as a framework to generate electroencephalographic (EEG) brain signals. We introduce a modification to the improved training of Wasserstein GANs to stabilize training and investigate a range of architectural choices critical for time series generation (most notably up- and down-sampling). For evaluation we consider and compare different metrics such as Inception score, Frechet inception distance and sliced Wasserstein distance, together showing that our EEG-GAN framework generated naturalistic EEG examples. It thus opens up a range of new generative application scenarios in the neuroscientific and neurological context, such as data augmentation in brain-computer interfacing tasks, EEG super-sampling, or restoration of corrupted data segments. The possibility to generate signals of a certain class and/or with specific properties may also open a new avenue for research into the underlying structure of brain signals.","urldate":"2023-01-15","publisher":"arXiv","author":[{"propositions":[],"lastnames":["Hartmann"],"firstnames":["Kay","Gregor"],"suffixes":[]},{"propositions":[],"lastnames":["Schirrmeister"],"firstnames":["Robin","Tibor"],"suffixes":[]},{"propositions":[],"lastnames":["Ball"],"firstnames":["Tonio"],"suffixes":[]}],"month":"June","year":"2018","note":"arXiv:1806.01875 [cs, eess, q-bio, stat] version: 1","keywords":"Computer Science - Machine Learning, Electrical Engineering and Systems Science - Signal Processing, Quantitative Biology - Neurons and Cognition, Statistics - Machine Learning","bibtex":"@misc{hartmann_eeg-gan_2018-2,\n\ttitle = {{EEG}-{GAN}: {Generative} adversarial networks for electroencephalograhic ({EEG}) brain signals},\n\tshorttitle = {{EEG}-{GAN}},\n\turl = {http://arxiv.org/abs/1806.01875},\n\tabstract = {Generative adversarial networks (GANs) are recently highly successful in generative applications involving images and start being applied to time series data. Here we describe EEG-GAN as a framework to generate electroencephalographic (EEG) brain signals. We introduce a modification to the improved training of Wasserstein GANs to stabilize training and investigate a range of architectural choices critical for time series generation (most notably up- and down-sampling). For evaluation we consider and compare different metrics such as Inception score, Frechet inception distance and sliced Wasserstein distance, together showing that our EEG-GAN framework generated naturalistic EEG examples. It thus opens up a range of new generative application scenarios in the neuroscientific and neurological context, such as data augmentation in brain-computer interfacing tasks, EEG super-sampling, or restoration of corrupted data segments. The possibility to generate signals of a certain class and/or with specific properties may also open a new avenue for research into the underlying structure of brain signals.},\n\turldate = {2023-01-15},\n\tpublisher = {arXiv},\n\tauthor = {Hartmann, Kay Gregor and Schirrmeister, Robin Tibor and Ball, Tonio},\n\tmonth = jun,\n\tyear = {2018},\n\tnote = {arXiv:1806.01875 [cs, eess, q-bio, stat]\nversion: 1},\n\tkeywords = {Computer Science - Machine Learning, Electrical Engineering and Systems Science - Signal Processing, Quantitative Biology - Neurons and Cognition, Statistics - Machine Learning},\n}\n\n","author_short":["Hartmann, K. G.","Schirrmeister, R. T.","Ball, T."],"key":"hartmann_eeg-gan_2018-2","id":"hartmann_eeg-gan_2018-2","bibbaseid":"hartmann-schirrmeister-ball-eeggangenerativeadversarialnetworksforelectroencephalograhiceegbrainsignals-2018","role":"author","urls":{"Paper":"http://arxiv.org/abs/1806.01875"},"keyword":["Computer Science - Machine Learning","Electrical Engineering and Systems Science - Signal Processing","Quantitative Biology - Neurons and Cognition","Statistics - Machine Learning"],"metadata":{"authorlinks":{}},"html":""},"bibtype":"misc","biburl":"https://bibbase.org/zotero/kpiebaareh","dataSources":["NX7P25CsTGD9b2Wgx"],"keywords":["computer science - machine learning","electrical engineering and systems science - signal processing","quantitative biology - neurons and cognition","statistics - machine learning"],"search_terms":["eeg","gan","generative","adversarial","networks","electroencephalograhic","eeg","brain","signals","hartmann","schirrmeister","ball"],"title":"EEG-GAN: Generative adversarial networks for electroencephalograhic (EEG) brain signals","year":2018}