Chapter 1: The multi-level, magnocellular mean value system of the brain

Dedicated in highest consideration to Prof. Dr. Ulrich Ramacher,

who inspired the author to suspect a multi-level mean value system in the brain *) and who considered an electronic circuit diagram of the brain to be possible - may readers also be inspired by his work "on the construction of artificial brains

*) Page 323 - Fig. 16.5 Resolution pyramid

Part 1.1 Cortex cluster, activity neuron and mean signal level

According to the classical view of neurologists, the cortex cortex is the place of conscious perception. A topological organization of the cortex cortex is known to consist of constantly repeating substructures, the main basis of which are the cell columns. The barrel columns in the rat cortex or the colour blobs and orientation columns in the visual cortex are described. It does not seem necessary to mention the source here, since these facts have become part of general knowledge.

In the visual cortex, color blobs and orientation columns could be combined to form hyper columns because their signal-processing affiliation to a certain pixel (image point) of the retina of an eye had been recognized.

While the color blobs and cell columns could be made visible by suitable methods, the hypercolumns are an arbitrary but meaningful summary based on recognized interconnection structures.

Such an arbitrary but meaningful summary of neurons of the cortex cortex seems to be urgently required with regard to the function of the brain and is therefore carried out by the author of this monograph.

Definition 1.1: Cortex cluster

We think of the cortex cortex divided into small spatial units, which we call clusters. We imagine a single cluster as a cylinder that comprises all six cortex layers. Since the cortex cortex is organised in cell columns, such a cluster might also comprise several cell columns.

The cortex clusters may overlap each other so that neighbouring clusters have a common subset of neurons. This overlapping is useful from a systems theory perspective, as will be shown later.

In each cluster we now imagine a large pyramid cell, which may lie in layer V. We assume that this neuron is a so-called mean value neuron. With its huge dendrite tree, this neuron taps all output deliverables in the cluster and forms a weighted average from these signals.

Definition 1.2: Mean neuron (activity neuron) of a cortex cluster

Each cortex cluster has exactly one magnocellular output neuron in the cortex layer V, whose input is supplied by all other output neurons of the cluster. The input-providing neurons are called signal neurons of the cluster. The signals they represent are called elementary signals of the cluster. The mean-value neurons of the cortex clusters are also called activity neurons. The signal strength (fire rate) of the activity neuron may be described as the signal level of the cluster.

The activity neurons postulated by the author use the well-known principle of convergence of neuronal excitation. Many input neurons project their output onto a magnocelluar neuron. While the projecting input neurons are usually grouped together in a receptive field, clusters are spoken of here in view of the spatial extension of the input field.

A cortex cluster is thus the spatial input catchment area of a large, magnocellular mean neuron in the cortex layer V.

We assume the possible existence of excitatory interneurons, whose task it is to supply the larger dendrite tree of a mean value neuron with excitation of the more distant neurons, which this mean value neuron cannot reach with its own dendrites.

We also assume the possible existence of inhibitory interneurons, which dock input-wise to the axon collaterals of the activity neurons and in turn inhibit the activity neurons of the neighboring clusters when excited, which would result in contrast enhancement within the signal levels of the cluster set. However, this inhibition may be relative and not total, i.e. the inhibition of the neighbouring cluster neurons increases with increasing rate of fire and decreases with increasing distance.

We attribute the limitation of the input suppliers of an activity neuron to a cortex cluster to the existence of a distance-dependent attenuation of the signals. The fact that predominantly unmyelinated axons transmit excitation from the signal neurons to the activity neuron is another reason for the limitation of the maximum signal range.

At this point it is not yet decided whether the signals are transmitted via action potentials or only subliminally via membrane voltages.

Sketch 1.1: Cortex clusters and their mean value signals

What benefit could nature derive from the existence of cortex clusters and associated mean value neurons?

One of the main abilities of the brain is its ability to learn. In learning, different signals are linked together in a meaningful way. To do this, however, the brain must first recognize whether signals exist at all, and if so, where they are located.

Signals reach our consciousness in the cortex cortex. It is therefore useful to divide the cortex cortex into observation areas and to determine the signal activity in each observation area.

This is precisely why, in the author's opinion, cortex clusters exist. Every cortex cluster is an observation area. Its mean signal activity is measured by means of the mean neuron located there.

If the cluster neurons are inactive, this mean value is the zero signal. The mean neuron is silent.

However, if a certain minimum number of cluster neurons are active, the mean neuron delivers a perceptible signal whose signal strength - i.e. fire rate - also increases with increasing cluster activity. In this respect, this mean value neuron is an attention detector and may also be called an activity neuron. The more active the signal position in the cluster, the more attention the nervous system should pay to this cluster.

Theorem 1.1: surveillance theorem

The firing rate of the mean neuron (activity neuron) of a cortex cluster is a measure of the neuronal activity in the cluster and thus indicates whether signals are present there that might be worth learning.

By convention, the activity neurons may all be located in the V layer. However, this does not exclude the possibility that there cannot be other, different types of neurons in layer V. However, the magnocellular, large pyramid cells in the cortex layer V, which are referred to here as activity neurons, have already been detected.

Therefore, the question now arises as to which subsystems of the brain evaluate the output signals of the activity neurons.

It seems important to keep the output of the cortex clusters strictly separate in future. The cluster neurons form one output group, the activity neurons the second output group. It will be shown that this fundamental division will be found in many brain structures.