Brain-Computer Interface Race

Collaboration Possibilities: Brain Products

Race Course

Pilots will be equipped with brain-computer interfaces (BCIs) that will enable them to control an avatar in a racing game played on computers. The game will be either horse racing or car racing. Special visual and auditory effects will be used to make the game exciting also for the audience. The image will be displayed on large screens in the stadium. The game will have the following characteristics:

  • Avatars will have their own base velocity even in the absence of BCI input; this is to ensure that teams will not get stuck at the start.
  • Avatars will race along a track with obstacles such as trenches, rocks, low walls, bushes etc. Hitting an obstacle will cause the avatar to slow down temporarily. There will be power-ups such as food or gasoline cans that will temporarily speed up the avatar.
  • The primary BCI input signal will be left/right steering, which will allow obstacles to be avoided and power-ups to be picked up. This input will most likely be discrete. For example, a single received command will move the avatar left/right by a predefined distance.
  • As the delay between the decision and a measurable signal can be up to a few seconds, the time window of detection of the signal should be slow-paced enough to be controllable via BCI, while the game animations should still appear fast, entertaining with good action effects.
  • Funny animations will be added to indicate absence of meaningful BCI input. For example, if the game receives no input from one pilot, a car avatar may start to smoke while a horse avatar may start to snort.
  • A secondary BCI input (‘action signal’) will be added for teams that can offer two independent BCI input channels. The use of this channel will not be absolutely necessary, but it will give the pilot an in-game advantage (e.g. ability to make horse jump).
  • If feasible, a third BCI input (‘action signal’) will be added for teams that can offer three independent BCI input channels. Such a signal could be to decelerate the avatar in a curve to avoid too much time loss due to drifting or sliding along the side fence.
  • There will be a clear finishing line at the end.
  • The duration of the race should be about 5-10 min.
  • The audience will be able to see the overall race course, and single viewers (ego-) perspectives. Perhaps images can be switched (similar to the broadcasting of a Formula 1 race).

 

To design the game, the event website will be open for 6 months to collect developers’ preferences regarding game type and signal-command mapping. A prototype of the game will then be provided on the website as soon as possible so that teams can train for the competition.

 

Game Rules

  • The pilots will start simultaneously.
  • Pilots will not be allowed to intentionally generate artifacts (e.g. eye movement, muscle activation) that could affect the BCI.
  • Pilots will have to keep their eyes open and to look at the animation screen during the race.
  • To prove function and prevent cheating, at least one training session will have to be performed with the BCI prior to the official race. During this session, independent judges will check the actions of all pilots and will have access to all hardware and software used by the teams. They may also request that the pilots e.g. intentionally generate artifacts to see how the BCI is affected.
  • During the event, judges will monitor the behavior of the BCIs (e.g. check, if they apply movement artifacts to get an advantage) and pilots and may impose penalties ranging from avatar slowdown to disqualification.

board_def_BCI_colors
D’Arc. Studio Associates Architects

 

Inclusion Criteria for the Pilots

Participating pilots will be recruited by the teams themselves. It will be each team’s responsibility to ensure that their pilot will be eligible and, for instance, not vulnerable to cybersickness, epilepsy and other problems. Pilots must fulfill the following inclusion criteria

  • Pilots must have severe cervical spinal cord injury that lead to a paralysis with complete loss of motor function below the neck (equivalent to e.g. ASIA A or B for spinal cord injury, see www.sci-info-pages.com/levels.html). The pilots will have to give testimony about the completeness of their lesion. A stimulator-based test will be developed and applied to confirm completeness of the lesion.
  • Pilots must have sufficient cognitive and communication ability to understand the competition, to give informed consent and to communicate independently.
  • No specific diagnosis (e.g. stroke, SCI, ALS) is excluded by default, pilots will be judged on a case-by-case basis with regard to motor and cognitive impairment.
  • Pilots must be at least 18 years old.

Inclusion Criteria for the Technology

The following criteria apply for the BCI hardware:

  • The device must be worn by the pilot.
  • While the primary envisioned BCI type is electroencephalography (EEG), other brain activity measurements such as functional near infrared spectroscopy will be also permitted as long as they will primarily measure brain activity.
  • It will be the team’s responsibility to prove the device’s technical function and safety prior to the race by submitting photographs with longshots and clear close-up views, videos, certification of local boards, risk analysis etc.
  • Electrodes can be wired or wireless, and the BCI amplifier can be powered by any means as long as safety requirements will be fulfilled.
  • Skin-piercing electrodes will not be permitted; Other than that, the choice of electrode type, cap and gel (if any) will be up to the pilot’s team.
  • EOG electrodes will have to be used; if not for artifact rejection, then for judges to check presence of eye artifacts. A similar rule may be added for muscle artifacts.
  • The pilots will have to watch the race screen during the race, and no additional (artificial) visual or audio stimulation may be given by the BCI. [Thus, steady state visually evoked potentials (SSVEPs, P300 etc.) will not be permitted unless they will be elicited from the race animation provided by the organizers and not from an additional display. This is, because we should show that BCIs can work outside the lab, not requiring additional display technology needed to generate additional brain responses.]

The following criteria apply for the BCI software:

  • Artifact removal will be crucial. Teams will have to sign in advance that they will remove muscle, eye and other artifacts. Before the competition, teams will have to send a description of the artifact removal procedure and examples of the signals to be checked by external judges who are experts on signal processing.
  • An alternative option for artifact removal will be to standardize the procedure: each BCI device will send raw data to a central server, which will preprocess it and send it back within 50 ms. This could, however, be difficult, especially if signals other than EEG will be used.
  • Ocular control will not be permitted.
  • Once artifacts will be removed, any signal feature and classification procedure can be used in the BCI as long as it primarily will reflect brain activity. As with artifact removal, teams will have to send the description of the inference process to be checked by external judges before the competition.
  • The output signal will have a standardized format (e.g. XML) that can be directly connected to the race. It will be sent to the race using a standard network protocol such as UDP or TCP/IP.

Leave a Reply