Consumer-grade EEG-based Eye Tracking

TL;DR

This work introduces a publicly accessible dataset for EEG-based eye tracking using consumer-grade hardware, aimed at benchmarking the feasibility of reconstructing gaze from EEG signals. It comprises 113 participants across 116 sessions (11h45m total) recorded with a Muse S2 headset and webcam across four gaze paradigms, with synchronized EEG, gaze, and stimulus data and a two-step preprocessing pipeline (Kalman smoother SARIMA imputation and 0.5–40 Hz filtering). The data are provided in CSV and XDF formats, include train/test splits per paradigm, and are complemented by open-source code to reproduce preprocessing and analyses, with validation of data quality and explicit known-issue reporting. The dataset enables realistic benchmarking of EEG-ET methods on affordable hardware and invites time-series and functional data analysis approaches to improve gaze reconstruction under consumer-level constraints.

Abstract

Electroencephalography-based eye tracking (EEG-ET) leverages eye movement artifacts in EEG signals as an alternative to camera-based tracking. While EEG-ET offers advantages such as robustness in low-light conditions and better integration with brain-computer interfaces, its development lags behind traditional methods, particularly in consumer-grade settings. To support research in this area, we present a dataset comprising simultaneous EEG and eye-tracking recordings from 113 participants across 116 sessions, amounting to 11 hours and 45 minutes of recordings. Data was collected using a consumer-grade EEG headset and webcam-based eye tracking, capturing eye movements under four experimental paradigms with varying complexity. The dataset enables the evaluation of EEG-ET methods across different gaze conditions and serves as a benchmark for assessing feasibility with affordable hardware. Data preprocessing includes handling of missing values and filtering to enhance usability. In addition to the dataset, code for data preprocessing and analysis is available to support reproducibility and further research.

Figures (8)

• Figure 1: Histogram of the age distribution of the participants. The left y-axis shows the number of participants in each age group, the right y-axis shows the proportion of participants in each age group.
• Figure 2: Left: Positions of electrodes of the Muse S 2 Headband in the 10-20 system. Right: Recording setup of the experiments.
• Figure 3: Left: Grid points in the "saccades" experiments. Displayed as blue diamonds (level-$1$ experiment), black dots (level-$2$ experiment) and black diamond (both). Right: Exemplary trajectory of the target that was presented to the participants. It was created using the parameterized function \ref{['eq:dataset:curve']}.
• Figure 4: Stimuli presentation. Top: Presentation of the stimulus in the level-2 smooth experiment. Bottom: Presentation of the stimulus in the level-2 saccades experiment.
• Figure 5: Exemplary imputation of missing values for recording "P$045$_$01$ level-$1$-saccades". Electrode AF7 did not contain missing values in this segment and is not displayed.