A Multimodal EEG–Source–Vehicle Dynamics Dataset for Road-Rage-Related Affect Recognition in Real-Road Traffic Conflicts

The data were collected in a real vehicle environment, where participants experienced two types of traffic-conflict induction scenarios, including longitudinal threat and lateral intrusion, under the control of a professional safety driver. The dataset contains synchronized 32-channel scalp EEG signals, source-space brain-region time series reconstructed from scalp EEG, vehicle dynamics signals, and four-level subjective road-rage-related affect labels.

The vehicle dynamics signals include vehicle speed, longitudinal acceleration, lateral acceleration, and yaw rate. All samples were temporally aligned using unified timestamps and assigned a four-level affective label, where 0 denotes no road rage, 1 denotes mild irritation, 2 denotes moderate anger, and 3 denotes severe anger.To protect participant privacy, the dataset has been anonymized and does not contain participant names, contact information, facial images, in-cabin videos, or other personally identifiable information.