Neurofeedback Using the Phenotype and Z-Score Modalities

A question-and-answer piece in which two developers explain when each of their approaches fits: Tom Collura for live z-score training and Jay Gunkelman for EEG phenotypes. Cynthia Kerson put the questions to them after a conference discussion. Collura says any QEEG-based training needs a complaint, a matching EEG abnormality and a reason to expect that fixing it will help. He warns that some deviations, such as excess SMR or certain hypocoherences, may be strengths that should not be trained away. He sees z-scores as most useful for training many connectivity measures at once. Gunkelman explains that phenotypes are a small set of EEG patterns, read from the raw EEG, that predict response to medication or neurofeedback. He argues that healthy brains do not sit at the group mean and that training to the mean can remove useful traits. He calls z-score training promising but unproven and advises informed consent where proven methods exist, as in epilepsy. It is an exchange of expert views, not a study.