Dynamic memory management and plasticity of the cortex

Part 2.10. Dynamic memory management and plasticity of the cortex

The human being is born with a nervous system that is subject to strong changes even after birth. The climbing fibre substitution described in the previous chapter is highly suitable to explain the process of plasticity of the brain and especially of the cortex cortex. At the same time, the prejudice is dispelled that the brain has a static, fixed "wiring". Instead, new neurons are constantly being added, which are integrated into the existing system. It has already been proven that neurons also die off.

Of course, the following is only an unconfirmed theory of the author.

We imagine a primary cortex cluster of n signal neurons. In the associated primary cerebellum its already k complex signals have been learned. The k-th complex signal is the last learned. We interpret the learning of complex signals as a multistage process.

So if there are already learned complex signals, there are k Purkinj groups in the primary Cerebellum. Each of them is assigned exactly one positive and one negative nuclear neuron. We use the circuit variant according to sketch 1.25 of page 81. We now think of a new and imprintable complex signal of the number k + 1, which is still unimprinted, but now acts as strongly and as long as it takes to imprint it. Since none of the first k Purkinj groups recognizes this new signal as their own signal, the kth distribution neuron will be active in the nucleus olivaris. This is because no negative nuclear neuron has inhibited the signal propagation along the sequential distribution chain. Therefore, the k-th distribution neuron in the olivaric nucleus is active as long as this new complex signal is active. According to the author, this k-th distribution neuron in the nucleus olivaris activates a previously inactive proneuron, which converts into a glutamatergic neuron. This new, (k+1) -th distribution neuron is excited by the predecessor neuron, if the latter is active.

During its "maturation", this new distribution neuron forms an axon that grows into the associated cerebellum cluster, possibly using the axon of the kth distribution neuron to the cerebellum cluster as a guiding structure. An additional marker-controlled target search is also possible. Upon arrival in the cerebellum, this axon activates a number of different proneurons with its neuronal activity, which ultimately mature into Purkinje cells, basket cells, star cells and golgi cells. Likewise, this axon activates two proneurons in the cerebellum nucleus, one of which becomes a positive (glutamatergic), the other a negative (GABAergic) nuclear neuron.

Only after all these cells have matured can an active, imprintable and previously unimprinted complex signal imprint the newly formed Purkinj group, so that the cerebellum cluster now contains k+1 imprinted complex signals.

Here the dynamic process of memory formation from proneurons continues. The (k+1) -th positive nuclear neuron forms an axon, which grows towards the secondary thalamus and awakens a sleeping proneuron there, which as a glutamatergic thalamus neuron sends an axon towards the secondary cortex. Therefore, a new cortex neuron is formed in the secondary cortex from a suitable proneuron, which is active exactly when the new complex signal is active.

The matrix system forms a new climbing fiber signal to this cortex neuron, which is also generated successively by activating the required proneurons of the required type and converting them into neurons. The result is a new climbing fibre axon which belongs to the (k+1)-th complex signal and takes its marker-controlled path to the (k+1)-th Purkinje group in order to substitute - i.e. replace - the magnocellular climbing fibre.

Thus (k+1) complex signals are stored in the cerebellum, each of which is (gradually) addressed via the parvocellular climbing fiber axon. And the parvocellular climbing fiber signals are known to be used for replication. Therefore probably only these signals are part of the long-term memory.

With each new complex signal that is transferred into long-term memory, exactly one new neuron is added to the secondary cortex, as are new neurons in the associated primary cerebellum cluster. The more complex signals are learned, the more cortex neurons represent the area in the cortex cortex that is assigned to the original receptors. However, this is precisely the plasticity where the size of a cortex area corresponds to the "importance" and "differentiation" of the associated receptors.

For example, exercise can increase the size of the areas of the cortex cortex that are assigned to the fingers. This can be observed in violin lessons.

In terms of memory requirements, the computer and the brain are therefore different. In a computer, memory exists before it is needed. In the brain it is created dynamically during the learning process.

It is still unclear how existing memory is cleared, i.e. deleted, in order to be occupied by new, different signals. This must be possible because the mass of the brain remains relatively constant after a certain development, i.e. no constant (mass) new formation of neurons from proneurons can take place.

The algorithm of forgetting allows several neural variants. Perhaps part 3 or part 4 of the monograph will present more findings on this.

In the next section we will deal with a circuit in which a Cerebellum is operated quasi backwards. In many electronic circuits you can swap input and output. Such "backwards" operated circuits will be called inverse circuits. They sometimes produce astonishing results.

Theory of the neuronal circuity of the brain and analytical thinking

Part 2.10. Dynamic memory management and plasticity of the cortex