As the field of neurofeedback evolves, we define and apply new metrics that have value in understanding as well as training the brain.
No topic has garnered more interest or controversy than that of coherence training. Coherence addresses the connectivity in the brain, not just the tone or level of excitation. It reflects the content of the messages in a manner that amplitude training cannot.
Classical or "pure" coherence reflects the stability of the timing between two brain sites. If the EEG peaks and valleys have a constant phase separation, they will have a high coherence. This reflects the amount of information shared between the sites.
Phase measures the actual separation between two sites. It reflects the speed of information transfer between the sites, or between the two sites and other "third party" sites. For example, the left and right occipital cortices communicate with sensory thalamic nuclei including the lateral geniculate nucleus and the superior colliculus. As a result of symmetrical thalamocortical reverberations, the occipital lobes often exhibit synchronous (symmetrical) bilateral alpha waves. So although the left and right occipital lobes do not communicate directly with each other, they share a common connection that still serves to bring them into synchrony.
Why is connectivity training important? Because it reflects the relative amount of processing that the brain is devoting to particular activities, in contrast to others. For example, if the thalamus is engaged in a reverberation with the cortex, then other mechanisms having to do with recall, processing, and other internalized operations are limited. Similarly, coherence is relevant to language, planning, and other higher brain activities. It is also of value in cases of injury or trauma, in which brain connectivity is compromised.
In addition to "pure" coherence, other metrics are of value in assessing and training brain connectivity. The BrainMaster system includes a "similarity" metric that is related to coherence, but is sensitive to the actual phase separation of the signals, as well as to their relative size. This metric is maximized when the signals are lined up in phase, and are of similar size, and has particular value for "synchrony" training.
Another such measure is the Spectral Correlation Coefficient (SCC) implemented by David Joffe in Lexicor's BioLex software. This reflects how similar two signals are in their spectral energy signature. In other words, if in a particular frequency band, the energy distributions in the Fast Fourier Transform (FFT) have similar shape, then the SCC will be large. This was developed as a convenient way to assess how similar two signals are, without regard to their phase relationship or timing. It only looks at how they are similar in terms of the "bars" in the FFT amplitude spectrum.
Another related measure is the "comodulation" originally defined by Drs. Barry Sterman and David Kaiser, implemented in their SKIL software. This looks at a particular frequency band, and watches how the amplitudes of the energy in each band vary over time, in comparison to each other. Simply put, if the signals wax and wane together, they will have a high comodulation. Like SCC, comodulation does not look at the relative timing (phase) of the signals themselves at all, it only looks at the amplitudes of the signals.
Both SCC and comodulation have been found of clinical and research value, in that they provide important indicators of brain functioning in a variety of scenarios. For example, Dr. Kirtley Thornton has studied and trained the SCC metric in children with learning disabilities, and has created a database of normative data as well as a set of protocols useful for neurofeedback training.
David Kaiser has noted that brain connectivity has a "Goldilocks" aspect. It can be too high, or too low in any given pair of sites.
Connectivity training is thus oriented toward getting connectivity "just right," hence optimized.