MEDiC: Mitigating EEG Data Scarcity Via Class-Conditioned Diffusion Model

Gulshan Sharma; Abhinav Dhall; Ramanathan Subramanian

MEDiC: Mitigating EEG Data Scarcity Via Class-Conditioned Diffusion Model

Gulshan Sharma
, Abhinav Dhall
, Ramanathan Subramanian

Information Systems

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

Abstract

Learning with a small-scale Electroencephalography (EEG) dataset is a non-trivial task. On the other hand, collecting a large-scale EEG dataset is equally challenging due to subject availability and procedure sophistication constraints. Data augmentation offers a potential solution to address the shortage of data; however, traditional augmentation techniques are inefficient for EEG data. In this paper, we propose MEDiC, a class-conditioned Denoising Diffusion Probabilistic Model (DDPM) based approach to generate synthetic EEG embeddings. We perform experiments on a publicly accessible dataset. Empirical findings indicate that MEDiC efficiently generates synthetic EEG embeddings, which can serve as effective proxies to original EEG data.

Original language	English
Title of host publication	Deep Generative Models for Health Workshop NeurIPS 2023
Editors	Vincent Fortuin, Sonia Laguna, Laura Manduchi, Stephan Mandt, Emanuele Palumbo, Melanie F. Pradier, Julia Vogt
Publisher	MIT Press
Pages	1-7
Number of pages	7
Publication status	Published - 27 Oct 2023
Event	The 37th Annual Conference on Neural Information Processing Systems (NeurIPS 2023) - Ernest N. Morial Convention Center, New Orleans, United States Duration: 10 Dec 2023 → 16 Dec 2023

Conference

Conference	The 37th Annual Conference on Neural Information Processing Systems (NeurIPS 2023)
Abbreviated title	NeurIPS 2023
Country/Territory	United States
City	New Orleans
Period	10/12/23 → 16/12/23

Access to Document

https://openreview.net/forum?id=0aeDKGhlToLicence: Unspecified

Cite this

@inproceedings{1fae37b10f2b4177897edddd93514044,

title = "MEDiC: Mitigating EEG Data Scarcity Via Class-Conditioned Diffusion Model",

abstract = "Learning with a small-scale Electroencephalography (EEG) dataset is a non-trivial task. On the other hand, collecting a large-scale EEG dataset is equally challenging due to subject availability and procedure sophistication constraints. Data augmentation offers a potential solution to address the shortage of data; however, traditional augmentation techniques are inefficient for EEG data. In this paper, we propose MEDiC, a class-conditioned Denoising Diffusion Probabilistic Model (DDPM) based approach to generate synthetic EEG embeddings. We perform experiments on a publicly accessible dataset. Empirical findings indicate that MEDiC efficiently generates synthetic EEG embeddings, which can serve as effective proxies to original EEG data.",

keywords = "Synthetic EEG data, Class-Conditioned Diffusion Model, Alzheimer{\textquoteright}s disease, Frontotemporal Dementia, EEG Augmentation",

author = "Gulshan Sharma and Abhinav Dhall and Ramanathan Subramanian",

year = "2023",

month = oct,

day = "27",

language = "English",

pages = "1--7",

editor = "Fortuin, \{Vincent \} and Laguna, \{Sonia \} and Manduchi, \{Laura \} and Mandt, \{Stephan \} and Palumbo, \{Emanuele \} and \{ F. Pradier\}, Melanie and Vogt, \{Julia \}",

booktitle = "Deep Generative Models for Health Workshop NeurIPS 2023",

publisher = "MIT Press",

note = "The 37th Annual Conference on Neural Information Processing Systems (NeurIPS 2023) , NeurIPS 2023 ; Conference date: 10-12-2023 Through 16-12-2023",

}

Sharma, G, Dhall, A & Subramanian, R 2023, MEDiC: Mitigating EEG Data Scarcity Via Class-Conditioned Diffusion Model. in V Fortuin, S Laguna, L Manduchi, S Mandt, E Palumbo, M F. Pradier & J Vogt (eds), Deep Generative Models for Health Workshop NeurIPS 2023. MIT Press, pp. 1-7, The 37th Annual Conference on Neural Information Processing Systems (NeurIPS 2023) , New Orleans, Louisiana, United States, 10/12/23. <https://openreview.net/forum?id=0aeDKGhlTo>

MEDiC: Mitigating EEG Data Scarcity Via Class-Conditioned Diffusion Model. / Sharma, Gulshan; Dhall, Abhinav; Subramanian, Ramanathan.
Deep Generative Models for Health Workshop NeurIPS 2023. ed. / Vincent Fortuin; Sonia Laguna; Laura Manduchi; Stephan Mandt; Emanuele Palumbo; Melanie F. Pradier; Julia Vogt. MIT Press, 2023. p. 1-7.

Research output: A Conference proceeding or a Chapter in Book › Conference contribution › peer-review

TY - GEN

T1 - MEDiC

T2 - The 37th Annual Conference on Neural Information Processing Systems (NeurIPS 2023)

AU - Sharma, Gulshan

AU - Dhall, Abhinav

AU - Subramanian, Ramanathan

PY - 2023/10/27

Y1 - 2023/10/27

N2 - Learning with a small-scale Electroencephalography (EEG) dataset is a non-trivial task. On the other hand, collecting a large-scale EEG dataset is equally challenging due to subject availability and procedure sophistication constraints. Data augmentation offers a potential solution to address the shortage of data; however, traditional augmentation techniques are inefficient for EEG data. In this paper, we propose MEDiC, a class-conditioned Denoising Diffusion Probabilistic Model (DDPM) based approach to generate synthetic EEG embeddings. We perform experiments on a publicly accessible dataset. Empirical findings indicate that MEDiC efficiently generates synthetic EEG embeddings, which can serve as effective proxies to original EEG data.

AB - Learning with a small-scale Electroencephalography (EEG) dataset is a non-trivial task. On the other hand, collecting a large-scale EEG dataset is equally challenging due to subject availability and procedure sophistication constraints. Data augmentation offers a potential solution to address the shortage of data; however, traditional augmentation techniques are inefficient for EEG data. In this paper, we propose MEDiC, a class-conditioned Denoising Diffusion Probabilistic Model (DDPM) based approach to generate synthetic EEG embeddings. We perform experiments on a publicly accessible dataset. Empirical findings indicate that MEDiC efficiently generates synthetic EEG embeddings, which can serve as effective proxies to original EEG data.

KW - Synthetic EEG data

KW - Class-Conditioned Diffusion Model

KW - Alzheimer’s disease

KW - Frontotemporal Dementia

KW - EEG Augmentation

UR - https://neurips.cc/virtual/2023/workshop/66495

UR - https://blog.neurips.cc/category/2023-conference/

M3 - Conference contribution

SP - 1

EP - 7

BT - Deep Generative Models for Health Workshop NeurIPS 2023

A2 - Fortuin, Vincent

A2 - Laguna, Sonia

A2 - Manduchi, Laura

A2 - Mandt, Stephan

A2 - Palumbo, Emanuele

A2 - F. Pradier, Melanie

A2 - Vogt, Julia

PB - MIT Press

Y2 - 10 December 2023 through 16 December 2023

ER -

MEDiC: Mitigating EEG Data Scarcity Via Class-Conditioned Diffusion Model

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this