Variability

Our brains are essentially powerful systems for producing and distributing energy. Work done with computer programs that emulate neural networks has shown that they have a tendency to run away with themselves. It is the brain’s control circuits which allow it to focus on useful functions.

EEG Variability

One of the measures of the EEG that helps to identify areas where training may be helpful is variability. When we see specific frequencies or areas in the brain which produce signals that jump from very low to very high, especially without a task to perform, this suggests that those areas are poorly controlled. As such, they are likely to be involved with brain functions which are not working well because the brain’s brakes are not functional. Variability represents a way of identifying inefficiency in the brain’s energy economy.

However, in looking at variability, we may also find areas or frequencies where the level of control is excessive. In such cases the brain’s brakes are clamped down tight, and it is not able to perform its functions effectively.

Measures of variability

When we measure EEG, we usually gather 256 samples every second.  When we talk about EEG in a period of 1 minute, we are talking about more than 15,000 individual measurements combined into a few numbers.  Most often we talk about the mean (or average) of all those values when we say, for example, that theta amplitude is 23u (microvolts), that’s useful, but it’s not completely descriptive.

To say that the brain has high levels of theta (rather sleepy and internal) or low levels of theta (blocked off from its sub-conscious processes) or a level more in the mid-range can be helpful, and the mean may help us to do this.

However, we can imagine 3 separate brains with the same mean of 23u with very different capabilities.  One might produce all 15,000 measures between 22 and 24u, though that’s not very likely. Another with the same average could be ranging from 18-28u.  A third could have the same average but include values from 13-33u.  Although all three of these brains would show the same average level of theta, we would expect them to operate quite differently.

The third brain, when it was at 33u would be clearly dominated by slow activity and out of contact with its environment; a bit later, when it was producing only 13u, theta levels might be nearly turned off and the individual very focused and intent.  In other words, the brain would be alternating between completely “off” in terms of external awareness and strongly “on”, going in and out of contact.

Our first example would be blocked off from either the strongly internal to the strongly external state by the rigidity of its control system.

We need something besides the simple average to more fully describe how each brain works.

Variance and Standard Deviation

One way of measuring the stability of a signal is called variance, a calculation of the difference between each individual measure and the mean. In order to include both positive values (those above the mean) and negative values (below the mean) the values are squared in calculating variance. This makes them all positive.  Taking the square root of the variance provides a statistic called standard deviation.

Of course a larger signal, say 50u, is likely to have a larger variance than a signal like 23u. In calculating the degree of control in brain frequency levels, we use a measure called coefficient of variability, simply dividing the variance by the mean.  In effect this tells us how dispersed the signal is.  The higher this value, the less controlled that area or frequency.

Training Variability

One of the key rules of brain-training is that the more efficient and effective the brain is, the less energy it tends to use and the less variable it is likely to be. You will find as we proceed that the major focus of our approach will be to focus on reducing whatever frequencies are excessive in a given brain. Unlike some other systems which train up fast frequencies, we will tend to concentrate on training down.

Since the lowest signal (theoretically 0u) cannot be reduced, by reducing the maximum values will reduce the range between the highest and lowest. This automatically reduces variance.  By focusing on cutting off the highest 10-15% of amplitudes at a given frequency, we teach the brain to increase its level of control and reduce variability.

On the other had, a brain which is excessively controlled can be trained by increasing the maximum level a a particular frequency. We will often find ourselves training to increase frequencies from 6 Hz to 15 Hz in order to break loose these stuck places.