EEG-based decoding of error-related brain activity in a real-world driving task

Zhang, H; Chavarriaga, R; Khaliliardali, Z; Gheorghe, L; Iturrate, I; Millán, J d R · 2015 · Crossref

DOI: 10.1088/1741-2560/12/6/066028

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the feasibility of using electroencephalography (EEG) to decode error-related brain activity during real-world driving, aiming to integrate brain-computer interfaces (BCI) into driving assistant systems. While previous research explored BCI applications in simulated environments, this work extends the paradigm to actual vehicle operation. The primary objective was to determine if a system could detect when a driver’s intended direction conflicted with a suggested turning direction provided by an assistant system, thereby inferring the driver’s true intention through the detection of error-related potentials (ERPs). The researchers conducted experiments in two settings: a car simulator with 22 participants and a real Infiniti FX30 on a closed track with 8 participants. In both scenarios, drivers followed static signs while a visual cue indicated a suggested turning direction before reaching an intersection. This cue matched the driver’s intended path 70% of the time. EEG signals were recorded from 64 channels and preprocessed using common average reference and bandpass filtering. Classification was performed using Linear Discriminant Analysis (LDA) on features extracted from 41 central electrodes, specifically targeting the 0.2 to 0.7-second window after the cue onset. The study evaluated both offline classification accuracy and online, closed-loop performance, where the system provided visual feedback in the real-car condition upon detecting an error potential. The results demonstrated that error-related brain activity could be reliably decoded in both environments. Offline classification yielded an average accuracy of 0.698 ± 0.065 in the simulator and 0.682 ± 0.059 in the real car, with performance significantly exceeding chance levels in all cases. Event-related potential analysis revealed consistent frontal negative deflections and theta-band power increases associated with error trials. Online experiments showed equivalent performance to offline tests, with accuracy improving across runs in the simulator as the classifier updated. In the real car, online decoding maintained robust performance, confirming the stability of error-related signals across different days and conditions. Control analyses confirmed that classification performance was driven by neural activity rather than eye movements or artifacts. The study concludes that decoding error-related potentials is feasible in real-world driving scenarios, marking the first online study of its kind in an actual vehicle. These findings support the potential for BCI systems to monitor driver intentions and verify alignment with automated assistant suggestions. The authors suggest that further improvements in machine learning algorithms could enhance system reliability, paving the way for integration into intelligent vehicle safety systems.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• in vehicle coaching
• human error taxonomy

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: tool software
• Theoretical Contribution: computational model