How is EEG biofeedback done well? Biofeedback is information. And the evidence is that more information is better. One "quality factor" for EEG biofeedback appears to be the information density: how much information is being given to the brain in a given amount of time. Joe Kamiya has always lamented the fact that the field was launched with "on-off", "go/no-go" feedback (binary feedback). He has long urged that proportional feedback be provided. In this manner, the feedback can "track" the signal with a "better/worse" implementation. How well the feedback signal accomplishes this is a matter of two quantities: 1) lag of the feedback signal with respect to the real-time EEG, and 2) information density (signal update rate). It is not enough to have one without the other.

In our implementation, bench-marking shows that a feedback signal lag time on the order of 120 msec, and a message update rate of forty times per second, are manifestly sufficient to give good feedback by the above-mentioned criteria. The client is usually able to "relate" to the signal being presented and to "feel" that it belongs to him and that he has some control over it. We also provide a stimulus-rich training condition by minimizing the inhibit periods, which are limited (by threshold) to about 20% of the time. These techniques have enabled us to work successfully with two-and-one-half year-old children, with deaf three year-olds, and with persons having severe cognitive disabilities (traumatic brain injury and stroke