Through electrodes on the scalp, a device measures electrical impulses in the brain, amplifies them and then records them. These impulses are divided into different types of brain waves.  For example, in order to concentrate on a task, parts of the brain must produce more high-frequency beta waves. To relax, the brain must produce more low-frequency theta waves.  Using a program similar to a computer game (only without a joystick), people learn to control the video display by achieving the mental state that produces increases in the desired brain wave. The use of neurofeedback for psychiatric problems depends on recent understanding about these diseases - scientists understand better the electrical and chemical components of mental illness, creating opportunities for neurofeedback. Children with attention deficit hyperactivity disorder (ADHD) use neurofeedback games to reduce theta waves and increase beta waves, increasing their attentiveness. Joel Lubar, Ph.D., a psychologist at the University of Tennessee, Knoxville, who originated neurofeedback treatment for ADHD in the 1970s, says neurofeedback can produce some of the same brain wave changes as drugs used to treat the disorder.

Neurofeedback is direct training of brain function, by which the brain learns to function more efficiently. We observe the brain in action from moment to moment. We show that information back to the person and we reward the brain for changing its own activity to more appropriate patterns. This is a gradual learning process. It applies to any aspect of brain function that we can measure. Neurofeedback is also called EEG biofeedback, because it is based on electrical brain activity, the electroencephalogram, or EEG. Neurofeedback is training in self-regulation. It is simply biofeedback applied to the brain directly. Self-regulation is a necessary part of good brain function. Self-regulation training allows the system (the central nervous system) to function better.

Neurofeedback is being used as an addition to many psychotherapists' practices as a boost to a sagging 'talk-therapy' practice. Since in most states there are no licensing requirements, therapists who have little knowledge of physiology or computer technology can use the equipment with little oversight. The client can be hooked up and the computer program does the training with little effort on the therapist's part. It adds a high-tech effect to the session and makes the whole procedure seem more credible.

In the Gravelle case in Huron, Ohio, 11 adoptive children were given thousands of hours of neurofeedback training at a great cost to the county with seemingly little improvement in their conditions. The criticism is that neurotherapy was given to them for an array of different problems simply because the funds were so easily available.