Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Are EEG Sequences Time Series? EEG Classification with Time Series Models and Joint Subject Training

Johannes Burchert, Thorben Werner, Vijaya Krishna Yalavarthi, Diego Coello de Portugal, Maximilian Stubbemann, Lars Schmidt-Thieme

TL;DR

This work reframes EEG classification as a time-series problem and benchmarks standard time-series classifiers against domain-specific EEG models. It introduces three joint-training strategies to inject static subject information via subject embeddings (CIC, CEC, SE) and evaluates performance on MI, SSVEP, and ERN datasets, showing that joint TSC models can match or surpass EEG models in several cases. Subject-conditional training yields top results on SSVEP and strong performance on ERN, highlighting the cross-domain value of time-series methods for EEG analysis. The findings motivate broader use of TSC baselines in EEG and point to future directions such as foundation-model-style EEG learning and scalable pretraining.

Abstract

As with most other data domains, EEG data analysis relies on rich domain-specific preprocessing. Beyond such preprocessing, machine learners would hope to deal with such data as with any other time series data. For EEG classification many models have been developed with layer types and architectures we typically do not see in time series classification. Furthermore, typically separate models for each individual subject are learned, not one model for all of them. In this paper, we systematically study the differences between EEG classification models and generic time series classification models. We describe three different model setups to deal with EEG data from different subjects, subject-specific models (most EEG literature), subject-agnostic models and subject-conditional models. In experiments on three datasets, we demonstrate that off-the-shelf time series classification models trained per subject perform close to EEG classification models, but that do not quite reach the performance of domain-specific modeling. Additionally, we combine time-series models with subject embeddings to train one joint subject-conditional classifier on all subjects. The resulting models are competitive with dedicated EEG models in 2 out of 3 datasets, even outperforming all EEG methods on one of them.

Are EEG Sequences Time Series? EEG Classification with Time Series Models and Joint Subject Training

TL;DR

This work reframes EEG classification as a time-series problem and benchmarks standard time-series classifiers against domain-specific EEG models. It introduces three joint-training strategies to inject static subject information via subject embeddings (CIC, CEC, SE) and evaluates performance on MI, SSVEP, and ERN datasets, showing that joint TSC models can match or surpass EEG models in several cases. Subject-conditional training yields top results on SSVEP and strong performance on ERN, highlighting the cross-domain value of time-series methods for EEG analysis. The findings motivate broader use of TSC baselines in EEG and point to future directions such as foundation-model-style EEG learning and scalable pretraining.

Abstract

As with most other data domains, EEG data analysis relies on rich domain-specific preprocessing. Beyond such preprocessing, machine learners would hope to deal with such data as with any other time series data. For EEG classification many models have been developed with layer types and architectures we typically do not see in time series classification. Furthermore, typically separate models for each individual subject are learned, not one model for all of them. In this paper, we systematically study the differences between EEG classification models and generic time series classification models. We describe three different model setups to deal with EEG data from different subjects, subject-specific models (most EEG literature), subject-agnostic models and subject-conditional models. In experiments on three datasets, we demonstrate that off-the-shelf time series classification models trained per subject perform close to EEG classification models, but that do not quite reach the performance of domain-specific modeling. Additionally, we combine time-series models with subject embeddings to train one joint subject-conditional classifier on all subjects. The resulting models are competitive with dedicated EEG models in 2 out of 3 datasets, even outperforming all EEG methods on one of them.
Paper Structure (20 sections, 5 equations, 1 figure, 8 tables)

This paper contains 20 sections, 5 equations, 1 figure, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. EEG Classification
  4. Time Series Classification
  5. Static and time-independent features
  6. Understanding EEG data as Time Series with Static Attributes
  7. EEG Classification
  8. Time Series Classification
  9. Joint Training for Time Series with Static Attributes
  10. Constant Indicator Channels
  11. Constant Embedding Channels
  12. Separate Embedding
  13. Experiments
  14. Datasets
  15. Experimental Setup
  16. ...and 5 more sections

Figures (1)

  • Figure 1: In the figure above we demonstrate the three proposed methods of using the subject information where is (a) Constant Indicator Channels, (b) the Constant Embedding Channels, and (c) Separate Embedding. Here $\oplus$ denotes the concatenation of two tensors across the channel dimension and for (b) $d$ represents the embedding size. We have set the subject Id to 2 as an example.