What are the detections based on? How were the algorithms created?
A common question is wanting to know how the normative database was built. That is, what are the EEG readings compared to?
Our detection suites are based on several different experimental datasets collected from volunteers in our own experiments.
For example, our Frustration detection was conducted with over 30 mixed gender subjects who wore Emotiv headsets, ECG, respiration band, GSR and sphygmomanometer equipment during defined activities. They were accompanied by a psychologist, filmed and self-reported before and after the event and baseline readings for all equipment were taken before and after.
The experiments consisted of a series of online games which were designed to elicit frustration, for example, a PacMan style game where controls randomly failed or reversed with increasing frequency, and a FPS game where the subject was gradually overwhelmed and their equipment was unreliable.
This is all proprietary data and unpublished, but the experiments were conducted under the Ethics board at a local University.
Additionally, some of our detections do not rely on prior data, they collect tagged data and train new signatures against each individual’s EEG data (our mental command suite works in this way).