###
Zscore training

Z Score Training

The Z Scores are based on brain dx datatbase.

Z Score training is entirely new way to work. Imagine a protocol that simply says “normalize coherence in all bands”. The “thresholding” is handled by the trainee’s age, eyes-open or closed condition, and the sensor sites used. We can design many protocols simply based on normalizing, or increasing, or decreasing, any quantities we like (there are 72 Z scores available). It will take some getting used to, but it is surely an empirically based, and sound method. Theta/beta ratio training is just one of the examples built into our latest release. Regarding the problems with ratio-based training, those that relate to numerator and denominator problems do not exist in this method. We are not computing metrics that can “blow up” when signals are small. Indeed, using Z scores provides a smooth normalization, that will, at the extremes, only to to 4 or 5 sigma. So this method deserves a new look, relative to ratio training (there are 10 such ratios in the software) We do not seeing this all as obviating the need for a QEEG where a need exists, but Z score training is an efficient and valid way of getting the same information in a real-time form, for appropriate needs. We personally do not see this as affecting the current role of the QEEG in any significant way. It does not replace the capabilities or information provided by a QEEG. What it provides is an additional, consistent, way of seeing what is happening. It also provides 1/30 second resolution, instead of 1-minute resolution. We are looking at the instantaneous value of a normative variable, and interpreting it in terms of the population mean. So we have not yet really articulated the meaning of “If this were your average value, then you would be in X% of the population”. We assume, and find empirically, that using Z scores as training variables is useful. How we interpret the scores and build protocols is new territory. But, since it does provide instantaneous measures of relative and absolute amplitude, coherence, asymmetry, phase, and ratios, it should be of value during training, so that you have a multidimensional handle on the EEG criteria you are training. I agree, when doing any 2-channel work in particular, the “way to go” seems to be to acquire both channels, then perform metrics and derived computations, to emulate bipolar as well as more complex protocol arrangements More significantly, it provides an entirely new conceptual framework for designing protocols. It amplifies the value of the QEEG, and the QEEG significantly informs the use of Z score training. Importantly, Z scores greatly simplify the process of biofeedback by reducing disparate measures such as power, coherence, phase delays and ratios to a single metric, i.e., the metric of the Z score. One no longer has to wonder whether to increase or decrease coherence or phase, etc. in a give electrode pair for a particular age or frequency, because the Z score simplifies this process by removing the guess work. We think these advancements are here to stay.