1. Does your data meet the requirements (anonymization is mandatory)? Information on the availability of informed consent and ethics approval must be declared. Yes. Data are anonymized. Written informed consent was given by all participants. Ethics approval for this study was provided through the IRB of Charité University Medicine, Berlin. 2. Who owns the copyright of the data? Is your data set freely available, or are there any restrictions pertaining to the sharing of the data? Do you license your data under one of the commonly used open licenses for data, as offered for instance by Creative Commons or Open Data Commons? The copyright is owned by the Neurotechnology Group at TU Berlin. We intend to make the data publicly available under the Creative Commons Attribution Non-Commercial No Derivatives license (CC BY-NC-ND 4.0) 3. Does the data come with a detailed description on the experimental paradigm, recording setup, and other relevant metadata? The description is as follows. The data are those that have been presented in Haufe S, Treder M S, Gugler M F, Sagebaum M, Curio G and Blankertz B 2011 EEG potentials predict upcoming emergency brakings during simulated driving J. Neural Eng. 8 056001 There were 18 subjects: VPae, VPbba, VPgab, VPgag, VPgam, VPja, VPbad, VPdx, VPgac, VPgah, VPih, VPsaj, VPbax, VPgaa, VPgae, VPgal, VPii, VPsal . Their task was to drive a virtual car using the steering wheel and gas/brake pedals, and to tightly follow a computer-controlled lead vehicle. This lead vehicle occasionally decelerated abruptly. The driver was instructed to perform immediate emergency braking in these situations in order to avoid a crash. Three blocks (45 min each) of driving were conducted with rest periods of 10–15 min in between. EEG data were acquired from 59 scalp sites and 2 bi-polar electroocular (EOG) sites (nose reference) using Ag/AgCl electrodes mounted on a cap (Easycap, Germany). EMG data were recorded using a bipolar montage at the tibialis anterior muscle and the knee of the right leg. The EEG and EMG signals were amplified and digitized using BrainAmp hardware (BrainProducts, Germany). Technical and behavioural markers such as stimulus onset times, brake and gas pedal deflection, acceleration of the lead vehicle and the driver’s own vehicle, as well as the distance between vehicles, were provided by the TORCS software. Braking response times were defined based on the first noticeable (above noise-level) braking pedal deflection after an induced braking manoeuvre. The EEG data were lowpass-filtered (tenth-order causal Chebychev type II filter) at 45 Hz. The EMG data were bandpass-filtered between 15 and 90 Hz (sixth-order causal Elliptic filter) with an additional notch filter (second-order digital) at 50 Hz for removing line noise, and rectified. Physiological and technical channels were synchronized and (causally) down-/ upsampled to a common sampling rate of 200 Hz. The concatenated data of all three blocks yielded one multivariate multi-modal time series per subject, which is provided here along with information about the timings of the braking onsets of the lead vehicle and driver. 4. Is your data in a format that can be readily accessed with standard tools (we strongly recommend to use the .mat format)? Yes. The list of files is VPae.mat VPbba.mat VPgab.mat VPgag.mat VPgam.mat VPja.mat VPbad.mat VPdx.mat VPgac.mat VPgah.mat VPih.mat VPsaj.mat VPbax.mat VPgaa.mat VPgae.mat VPgal.mat VPii.mat VPsal.mat In each file, there is a cnt structure containing the data, an mnt structure defining electrode positions, and an mrk structure containing the braking event. cnt.clab is the electrode names 'EOGv', and 'EOGh' are vertical and horizontal electrooculogram 'EMGf' is foot EMG 'lead_gas' and 'lead_brake' are gas and brake pedal deflections of the lead vehicle 'dist_to_lead' is the distance between driver and lead vehicle 'wheel_X' and 'wheel_Y' are steering wheel deflections of the driver 'gas' and 'brake' are gas and brake pedal deflections of the driver all other channels refer to scalp EEG cnt.fs is the sampling rate cnt.x is the continuous multivariate data mnt.pos3D and .x and .y contain 2D and 3D electrode positions for plotting mrk.classNames are the names of the different types of events The order of events is typically like this 1. car_brake: lead vehicle starts to brake 2. car_hold: lead vehicle stops braking and stays at slow speed 3. react_emg: the subject starts to brake. The onset is defined here through the EMG 4. the lead vehicle starts to accelerate again 5. car_collision: denotes collisions between driver and lead vehicle, happened very rarely mrk.time is the timestamp for each event in milliseconds mrk.y is a binary matrix telling which of the five types each event is mrk.events.react is the reaction time (the time between car_brake and react_emg) for each of the react_emg markers