Chapter 3: The limbic system

 

Dedicated to James Papez, the neurologist whose opinion that the structures of the limbic system are linked to each other by a conduction chain of high-capacity circuits inspired this part of the author’s monograph. The Papez Circuit, which has reaped much praise but has also much criticism, is the theoretical basis used to explain signal processing and information management in the limbic system

The Oranienburg prophesy:

It is common sense, and not computer simulation, that will enable us to broaden our knowledge of the world!

(Andreas Heinrich Malczan - June 2013)

 

Section 3.1. An excursion into the primeval ocean of epochs long past

 

Let us risk an excursion into the murky depths of times long past, when the first animal species were working out ways of actively obtaining their food. Let us select one of these primitive animal organisms and take a closer look at it. To facilitate identification, we can call him Fred.

Like all living things Fred tended to get hungry. But food availability was not a problem; he lived in the primeval ocean where food was available in abundance. The ocean was enormous and deep and its depths were pitch dark. Fred had never seen what we know as daylight. But this was not just because it was dark. The shallower parts of the ocean near the coasts of the land masses were actually quite bright. The problem was that the primitive animals of that epoch had not yet developed eyes. At the very most, they may have had a few photosensitive skin cells enabling them to distinguish between light and dark. But they could not actually see anything. Creatures like our Fred had never seen daylight because they were blind. It was with his sense of smell that Fred located his food. He was not herbivorous, but he depended on herbivores for his survival; they were his prey. But there was one problem. When Fred got a whiff of something that smelled good, it nearly always swam away. Because the animals on whom Fred preyed were mostly good swimmers. And that was not all; almost as soon as Fred had smelled his prey, he forgot about it.

Fred’s olfactory receptors were excited when the prey swam past. The resulting action potentials were transmitted to a neuron layer that would later be described as the olfactory cortex. As Fred’s muscles were now activated via motoneurons by the transmitter substance acetylcholine, the axons of the cortex had to be switched over from glutamate to acetylcholine. This was done by a transactional core which would later be named the septum. The signals switched over to acetylcholine in the septum were transmitted to the motoneurons allotted to them in the muscle groups, and these generated Fred’s mobility and controlled his steering function.

It obviously made sense to activate Fred’s whole system, not just the muscle groups responsible for the actions. That is why modern-day vertebrates still have cholinergic projections from the core of the septum to the hypothalamus, the part of the diencephalon controlling not only the autonomic nervous system but also the nucleus suprachiasmaticus which, in turn, controls the body’s circadian rhythm. Similarly, the septum projects via the habenula to the monoaminergic cores in the tegmentum, i.e. to the dopaminergic VTA, the adrenergic locus coeruleus and to the serotonergic nuclei raphes magnus. The transmitters in the two last-named organs activate the prosencephalon in vertebrates and mammals. And it is still the habenula in present-day animals which integrates the olfactory, septal and amygdaloidal  signals and transmits them via the nucleus dorsalis tegmenti to the neuron nuclei responsible for activating an animal’s masticatory and deglutitory muscles after it has killed its prey. All these systems are, moreover, present in duplicate. Work sharing between the left and right halves of the body was a relatively early evolutionary development.

Consequently, activation of the muscles and the neuronal systems by olfactory perception when either prey or enemies made their appearance was a step in the right direction. It was actually quite a good strategy, as long as the prey or the predators were slow to react and sluggish in their movements. But evolution did the rest. Many prey animals optimized their systems and escaped being eaten by reacting more quickly. The result was that they multiplied rapidly. And displaced the slower species which fell victim to the predators.

This gain in speed created a new problem. There was no parallel increase in the rate of diffusion of the scent emitted by the prey and this tended to decline sharply as it dispersed. This meant that, although perceptible, the scent of a rapidly moving prey would be located at a point where the prey no longer was and, in any case, it quickly evaporated. So the information received by the predators was extremely short-lived and the sensory receptors quickly closed down again. Fred had lost the scent of his prey.

It seems highly probable that Nature’s initial solution to this problem will have been enhancement of the sensitivity of the olfactory receptors and evolution of the sense of smell to a remarkably high level of efficiency. But this was nullified by further increases in the prey’s reaction speed, which frequently enabled it to escape the predator’s fangs. And its scent soon disappeared.

This meant that predators not wishing to die of starvation had to develop a memory – some sort of nerve cell circuit to remind them that a tasty morsel had just swum by at high speed. This memory would enable the predator to pursue that prey or search for other prey that had found itself a hiding place. The reminders emitted by this memory would generate action potentials that would ultimately control the muscle groups used for pursuit of prey.

The predator’s sense of smell provided the input for these signals. Specifically designed receptors – olfactory cells – generated action potentials  when the right kind of prey swam by. These receptors could detect the metabolites produced by the prey and diffused through its skin – even under water. Of course, the same applied to the hungry predators, whose greedy, wide-open jaws were disgorging similar metabolites and betraying their presence. This is the reason why an animal’s sense of smell was and had always been so important and so incredibly sensitive. It was equally important for procreation of the species. Even in Fred’s day, sexual attractants, now called pheromones, were already the agents responsible for bringing male and female of the same species together for the purposes of sex and procreation.

 

There was no lack of prey in those days. The only problem was that Fred forgot about it as soon as it had swum past him; the scent molecules were quickly dispersed in the ocean’s enormous mass of water. The signals stopped as soon as excitation of the olfactory cells ceased. The prey was forgotten. What needed to happen?

Some brain mechanism that repeated the signals had to evolve. Something akin to the forgetful housewife, who keeps murmuring “We’re out of coffee, we’re out of coffee”, until she finally notes it down on her shopping list. The technical solution was repetition!

It was a new neuronal subsystem functioning as an echo generator that produced this constant repetition of the original signals. This generated a stream of echoes from each input that was persistent enough to achieve the same result as the forgetful housewife’s addition to her shopping list. The original input came from the olfactory cortex. The ensuing output was a series of echoes repeating the input signal as often as was necessary to achieve the desired result.

Each of the output neurons in the olfactory cortex had its own output axon, through which Fred’s physical functions were activated via the septum when his olfactory cortex was excited by the action potentials signaling the presence of prey. The autonomic nervous system was responsible for activation and deactivation of these physical functions. But how did Nature solve the forgetfulness problem? What sort of repetition could provide the technical solution? What sort of circuit would provide a suitable echo generator?

The action potentials of each output neuron in the olfactory cortex were also conducted to a new neuron with a considerably longer axon. This could best be described as a transactional neuron. During the course of evolution this mass of tiny transactional neurons formed its own layer, in which they were embedded as granules. That is why they are now called granule  cells. The axons of granule cells were not myelinated and this meant that each action potential was transmitted relatively slowly, for example, at a rate of only 0.2 millimeters per millisecond. That is equivalent to 0.2 meters per second – roughly the same speed as present-day action potentials transmitted along unmyelinated fibers. This speed will certainly have been slower back in Fred’s day, simply because there was not yet any need for anything quicker.

If an axon of this type was twenty millimeters long, an action potential took 100 milliseconds to cross it. This caused a time lag. In physical terms, the axon was a conductor delaying the action potential travelling along it. For example, if this conductor was tapped 40 times at different points and at roughly equivalent intervals, forty echo signals would be received at intervals of 2.5 milliseconds. But there was as yet no neuron group to do the tapping, so one had to be evolved. These forty echoes were conducted to an integration neuron which put together a new series of signals containing the forty echoes. This integration neuron was also a new type of neuron playing a key role in the evolving echo generator.

From now on we will call these neurons doing the tapping ‘echo  neurons’. In our example they tap the delaying conductor to produce forty echoes spread over a period of 100 milliseconds. We will call these primary echoes. After their collection in the integration neuron a secondary echo of considerably longer duration is produced.

This creates a memory lasting one hundred milliseconds overall – the time taken for the action potential to be delivered in the form of output from the final echo neuron. An integration neuron has collected the forty separate echoes and used these to produce a time-lagged output signal, which was a secondary echo of, for example, 100 milliseconds’ duration at a firing rate of approx. 400 Hz.

If, however, the transmission speed of the action potentials had been only 0.1 m/s instead of the 0.2 m/s assumed in our example, the secondary echo would have lasted 200 milliseconds. The drawback was that, in order to reach a firing rate of 400 Hz, it would have been necessary to tap the delaying conductor 80 times. But even an echo lasting 100 milliseconds was better than nothing. Fred would at least be able to remember for that length of time that a potential prey had just swum by, and receive forty reminders of this from the 40x duplication of the original action potential. This was a memory of sorts, even if it was extremely short.

Animals using this kind of echo generator were able to remember the prey longer, but it must be assumed that this type of system was not particularly efficient in its early days. At the start there would perhaps be only one single echo neuron, but even this would double the number of action potentials generated from one olfactory excitation, as compared to a system producing no echoes at all. And the number of echoes went one higher with the inclusion of each additional echo neuron to tap the conductor. As each action potential was an electrical impulse, the number of these impulses multiplied accordingly. Electrical engineers nowadays call this frequency multiplication. It is a principle that was perfected in neurological systems many millions of years ago.

 

All the technical elements of the neuronal echo generator now have their own names. The axons functioning as delaying conductors are called mossy fibers. The neurons belonging to them are granule cells. The axons responsible for rapid transmission of the echoes from the action potentials to the integration neuron are called Schaffer collaterals, named after the famous neurologist Károly Schaffer (1864-1939) who did most of his work in Budapest.  The neurological structure of echo generators in a modern mammal’s brain is called the hippocampus. The Latin name is Cornu Ammonis (CA for short), in English Ammon’s horn.  This is subdivided into four serially numbered sections. Section 3, the CA3 region of the hippocampus, contains the echo-generating neurons, which are called CA3 pyramidal cells because of their shape. The CA1 section contains the integration neurons which unite the many separate echoes received from the mossy fibers into a secondary echo. These neurons are called CA1 pyramidal cells.

The degree of freedom of interpretation possible in any description of the circuitry of neuronal echo generators is quite astonishing. Mathematicians and electrical engineers will tend to visualize a delaying conductor of this kind as a simple conductor forming a tiny branch at each echo neuron, with the echo neuron tapping the conductor at this branch. Each echo neuron has its own output conductor to the integration neuron responsible for collecting the individual time-lagged echoes.

 

Fig. 1: Simplified diagram of an echo generator

 

 

Neurologists will take a totally different view of the same set of facts. They know that the CA3 pyramidal cells in the hippocampus are quite large neurons which (according to the author’s theory) should function as echo neurons. A single synapse is not capable of generating an action potential in the mossy fiber acting as delaying conductor. Consequently, the relevant mossy fiber, like nearly all axons, creates a dense network of axodendrites that either surrounds it or completely replaces it. These axon structures in the hippocampus are called mossy fibers because of their moss-like network structure; they have to be elongated because it is their length that enables them to function as delaying conductors. And an echo neuron – those in the hippocampus are called CA3 pyramidal cells  – now taps this ramified axon structure containing hundreds (or even thousands) of synapses, which together transmit an excitation that is sufficiently strong to generate an action potential.

But the axon conductors of the CA3 pyramidal cells functioning as echo neurons, which are called Schaffer collaterals, also form their own telodendrons on the C1 pyramidal cells functioning as integration neurons; these also have large ramified telodendrons making countless synaptic contacts with these equally large axodendrites. This type of Schaffer collateral is extremely numerous and they tend to run parallel to each other to form a network of nerves called a plexus. The result is that echo generators have possess some marked characteristics that make them readily identifiable.

 

Fig. 2: More realistic diagram of an echo generator

 

 

One advantage of a low diffusion speed of the action potentials along the delaying conductor was that this stretched the duration of the echo signals. This prolonged an animal’s memory of its prey and enabled it to continue the pursuit. Animals developing this faculty won the evolutionary race. But once this type of echo generator had evolved in all surviving animals of this subphylum (i.e. vertebrates), this evolutionary advantage was effectively neutralized. Predator and prey once again had equal fighting chances. But some still had a slight edge over the others by using low-speed action potentials along the delaying conductors to maximize echo generation. Further stretching of the prey’s echo would have been advantageous, because this would enable the predator to seek out a prey that had managed to hide itself for as long as the echo signals remained active, and because the action potentials generated by these signals controlled the muscles via the septum. But stretching the echo presented technical problems. For example, the delaying conductors would have to be twenty centimeters long to yield an echo lasting one second. The animals of that period were not that long overall. Could this even be the explanation for the giant size of some extremely primitive forms of life present in the ocean during the early stages of evolution? But the equation: Large Body = Longer Neuronal Delaying Conductors has a fatal flaw, because volume increases by the power of three in proportion to body length and higher water resistance causes an exponential increase in energy requirements that would sooner or later become insatiable.

So, was there some other way to enhance memory time and prevent the predator’s memory of the prey disappearing in a fraction of a second?

The solution was a technical trick – the output of the integration neuron was recycled as input. The olfactory receptors were excited when prey approached. The action potentials were transmitted from the cortex into the delaying conductor. This was tapped as often as possible by the echo neurons. The resulting echoes were transmitted to a common integration neuron along axon conductors that were better myelinated and therefore “faster”. Another new neuron – the recycling neuron – received these signals and routed them quickly along its myelinated axon to the starting point, i.e. to the granule cell of the olfactory cortex, thereby closing the circuit. Modern-day physicists call this recycling of input to point of origin ‘feedback’. This is the mode of action of modern electronic oscillating cycles.

 

Fig. 3: Use of feedback to produce continuous oscillation

 


 

An oscillation is generated by excitatory, i.e. positive, feedback of output to input. This enables a single impulse from an input signal to generate a continuous oscillation. This type of oscillation signal is generated in the limbic system by positive feedback into the echo generator. The action potentials circulate continuously on axons forming a closed circuit.

 

Fig. 4: Continuous oscillation in limbic system:

 

 

The benefits obtainable from gigantism degenerated into impediments with the evolution of echo generator feedback. The olfactory memory became independent of body size, especially in insects, many of which gradually morphed from their once gigantic forms into smaller and even very tiny ones. Systems with feedback-capable echo generators no longer needed elongated delaying conductors. Dwarfish animal forms were equally capable of developing good olfactory memories. Millions of years later this sort of dwarf growth would prove remarkably beneficial in proto-mammalian life forms in an era when planet Earth was under violent bombardment from comets, its atmosphere darkened by dust, its vegetation withering away and food becoming scarcer and scarcer. Giant animal life forms like the dinosaurs were unable to cope with this catastrophe and became extinct. The smaller forms with feedback-capable echo generators survived.

Thus, the first evolutionary step was the transformation of one action potential triggered by an animal’s olfactory sensors into the series of forty action potentials used in our example by using echo neurons to tap the delaying conductor forty times. The next step was to plug these echo conductors together on a single integration neuron, on which the forty time-lagged  action potentials were brought together in a time-lagged sequence. This yielded a highly desirable duplication of frequency that enhanced the system’s reactive sensitivity. The duplicated signals originally produced by the granule cells were routed back to them as input by a recycling neuron. To enable this, the delaying conductors, the echo conductors for the individual echoes, the output conductor for the secondary echo emitted by the integration neuron and the feedback conductor of the recycling neuron had to form some sort of circular structure, in which the end was in contact with the start. Many million years later, neurologists would name a structure of this kind in the limbic system the Papez Circuit. And they named the neurological structure of granule cells, delaying conductors, echo neurons and integration neurons the hippocampus.

The recycling neuron was now able to transmit its action potentials to two different points, either to the hippocampus or back to the olfactory cortex. Nature opted in favor of security and used both solutions. The result was that the recycling neuron routed the output signal back to both the olfactory cortex and the subiculum of the hippocampus, thereby closing the oscillation circuit at two points.

The recycling neuron linked to each delaying conductor was also able to transmit the output signal to other cerebral subsystems, for example, to the muscles, to the systems responsible for controlling blood pressure and to all the other functions enabling pursuit and capture of prey. The result was that the recycling neurons came together during the course of the evolutionary process to form a neuronal core emitting the output responsible for control of the key physical functions facilitating pursuit of prey, attack and defense and, of course, flight (in the sense of disengagement) – because the sense of smell not only helps to locate prey; it also warns of the presence of other dangerous predators capable of either competing for the prey or transforming a weaker predator into prey. We now call this neuronal core the amygdala. This lateral amygdala later acquired some important subsidiary cores that are now integrated into the amygdala and perform other functions. The section of the amygdala solely responsible for onward transmission of olfactory signals is now called the cortical core of the amygdala. As a host of other signals not originating in the olfactory system found their way to the hippocampus, the recycling  neurons handling these signals formed the lateral amygdala. The evolution of the lateral amygdala is proof that the limbic system receives signals not originating in the in the olfactory system. Several researchers, including Pitkänen, Aggleton and Saunders, distinguish between the following substructures in the amygdala: anterior amygdaloid area, accessory basal nucleus, central nucleus, cortical nucleus (anterior and posterior parts), basal nucleus (magnocellular and parvicellular parts), lateral nucleus, periamygdaloid cortex, paralaminar part of basal nucleus. Part 4 of this monograph will show how the evolutionary process has pushed the amygdala’s original function of routing signals received from the hippocampus back to the hippocampus and the olfactory cortex as feedback into the background, in order to make way for a much more urgent development – digitalization of the analog output of the cortex layers. The lateral amygdala would have been quite adequate for feedback of the signals alone.

The lateral amygdala, which had originally been the core formed by the recycling neurons feeding signals originally emitted by the hippocampus back to the hippocampus, subsequently established contact not only with the sympathetic and parasympathetic nervous systems controlling physical activity, but also with the hypothalamus, that centrally-located cavity of gray matter responsible for suppressing pain during battle or flight, as well as the locus coeruleus and the tegmentum controlling alertness and reticular formation. Nature is a great believer in system theory.

Interlinked neurons form a chain transmitting the input signal. We could call this chain ‘signal-linked neurons’. Its axons and dendrites form a sort of data link, along which a given signal is transmitted. For example, a signal transmitted from a neuron in the olfactory cortex goes to the relevant granule cell in the hippocampus, then along the axon of the mossy fiber to the echo neurons, from there on to the integration neuron and then to the recycling neuron in the amygdala and so on. If one assumes that these linked neurons can not only exchange neuronal signals – which are known to be in the form of chemical substances, so-called transmitter molecules –, it is also reasonable to assume and has already been demonstrated that exchange of other substances, in particular marker substances, along the same chain is a feasible proposition. And if these marker substances are position markers for the intended target structures, all of the neurons in the chain will possess these target markers, which guide their axon collaterals to the locations specified by the position markers. The neurons of a chain of this kind are to a certain extent marker-related. This is understandable when they are all transmitting their output to the same target structures.

This would explain why, for example, not only the septal, but also the amygdaloid and the hippocampal neurons, and even the original neurons in the olfactory cortex have contact with, for example, the hypothalamus. All these signal-linked neurons form   ‘tuples’ of marker-related neurons. And these then also have the same target regions. This strategy helps to circumvent problems arising from malfunction of one of the neurons in the chain. The target regions then receive the necessary input via the other marker-related neurons. This is the reason why both the olfactory cortex and also the septum, the amygdala and the hippocampus project in some cases into identical target structures.

The habenula already referred to above receives input from the septum and the amygdala, but also from the area olfactorius and the regio praeoptica, all of which interchange signals. This yields a greater degree of protection against malfunction and explains the remarkable robustness exhibited by neuronal systems.

On the other hand, it gives neurologists the problem of having to explain why each neuronal region is interlinked with most of the others for reasons that are mostly buried in Nature’s approach to system theory.

 

It should also be remembered that every single output neuron in the olfactory cortex has its own mossy fiber in the hippocampus, with its own echo neuron, Schaffer collaterals, integration neuron and recycling neuron in the amygdala. Also that the output neurons of the olfactory cortex represent different types of smell, each of which had its own oscillation cycle in the hippocampus where the signals are in constant circulation in order to prevent them from being forgotten.

To complete the picture, it should also be remembered that the olfactory cortex back in those days already had the capability to block neighboring receptors. Evolution had found it useful to hook up this blockade by installing interneuron links between every output neuron in the olfactory cortex and all the output neurons in their immediate vicinity. This enabled the stronger olfactory signals to suppress the weaker ones, thereby amplifying the stronger signals and attenuating the weaker ones. Insects already had this capability to block neighboring receptors in their olfactory systems. It was retained in many other neuronal cores of the cerebral system and amplified contrast in the output signals.

In principle, the primitive hippocampus described above generated for every single input signal (or more precisely for every cortical output neuron excited by an input signal) a continuous signal of higher frequency that kept on reminding the animal of the prey swimming past it and activating the physical motions enabling its pursuit. Similarly, the animal was prevented from forgetting the presence of an enemy by an oscillating signal triggered by hostile input. This mobilized and maintained for an adequate period the necessary action potentials for reactions to the enemy – either attack or flight. This was more or less equivalent to a memory. Once received, the oscillation triggered by the input signal circulated in the limbic loop and enabled the animal to remember the prey or the presence of an enemy. This memory function enabled pursuit of prey or, alternatively, flight or attack. The action potentials required for these reactions were supplied by the continuous circulating signals.

However, if, for example, each and every input action potential from the olfactory receptors generated forty new action potentials that were fed back to the granule cell at the start of the mossy fiber and each one of these was once again multiplied by forty and the process continued ad infinitum, wouldn’t this geometric progression produce a super-high-frequency oscillation of more or less infinite duration?

This is where the so-called ‘dead time’ of the neurons comes in. This was already an essential characteristic even at this early period of evolution. Every nerve cell took a very short break of roughly one millisecond after each action potential. This break is called the dead time. During this dead time the neuron is totally incapable of creating an action potential in response to even the most powerful excitation. This effectively put a ceiling on the firing rate. No neuron was capable of firing at a faster rate than that dictated by the reciprocal of the dead time. This is how Nature prevented a self-destructive increase in resonance frequency. It was Nature’s solution for providing positive signal feedback generating stable continuous oscillation without totally overloading the system.

The ceiling imposed on the firing rate by the dead time of the neurons lay somewhere within the range of 0 - 1000 Hz, i.e. approximately three powers of ten, and this created another problem. Signal strengths in Nature had in many cases much larger intervals than 0 - 1000. This made linear reproduction of signal strengths in firing rates by the receptors impossible. There was only one possible technical solution to this problem – logarithmic reproduction of natural signal strength in the firing rate. This is the reason why most receptors have a logarithmic curve. This keeps firing rates within the range dictated by the dead time. Only certain special neurons are capable of attaining firing rates of around 5000 Hz; most of them have firing rates significantly lower than 1000 Hz.

This maximum firing rate was now reached in the echo generator as well, as it fed back the output of the integration neurons as input into the granule cells. Unfortunately, this neurological solution had two serious design flaws. Firstly, it put a heavy load on the neurons involved because, once activated, they were compelled to keep on generating action potentials. There was a tetanic, i.e. higher-frequency, continuous oscillation in the circuit, in which the neurons continued firing without interruption. This was because each output from the hippocampus was recycled back to the hippocampus as input. These neurons fired at the maximum rate. This eventually led to overloading and premature death of the affected neurons, which consequently had to be replaced. It is now possible to demonstrate that the limbic system is still capable of producing new neurons and also new receptors.

The second flaw was much more serious. The rotation of the signals in the elementary Papez Circuit persisted not only for minutes, but also for hours and even days after the prey had managed to escape and was now miles away. The same applied when it had been caught, eaten and digested. Nature had to call a halt to this ongoing signal rotation, because it created the illusion that the prey was still around even after it had long since either fled or been eaten.

While it made sense to continue searching for a prey that had managed to find a temporary hiding place, if a further check of the immediate environment proved negative, it was reasonable to assume that it had escaped and was no longer in the vicinity. But the continuing rotation of the action potentials generated by the original olfactory fingerprint created the illusion that it was still somewhere close.

Although an illusion or mental image was basically a positive phenomenon, and could possibly be described as the birth of intelligence and consciousness, it was, in the context of search for prey, a dangerous illusion because it indicated the presence of prey that no longer existed.

Evolution had to supply a technical solution to both these problems, which will be designated here as hippocampus problems 1 and 2 respectively:

 

Hippocampus problem No. 1:

 

A way had to be found to interrupt the continuous rotation of the input signals with either short or longer breaks, during which the neurons could recuperate and, consequently, prolong their active life significantly.

 

Hippocampus problem No. 2:

 

A way had to be found to stop the continuous rotation of the input signals automatically after a certain time, when the intensity of the prey’s scent started to wane, either because it had escaped, already been eaten or the scent of new prey had made its appearance. There was also the possibility that other predators emitting a characteristic hostile scent could appear on the scene. In this case, the memory of the  original prey had to be deleted and replaced by neuronal reactions enabling either flight, defense or attack.

 

The first solution was to make the continuously firing neurons gradually less sensitive to further excitation by the repeated input. The initial excitation of inactive neurons often produced a higher firing rate than subsequent excitations, i.e. the neurons’ sensitivity to repeat excitation declined with time. This kind of excitation curve can still be found in many of the receptors in the nervous system. But this solution was totally inadequate.

 

The solution to hippocampus problem No. 1 was provided by a neurological structure now called the septum. Septal cores likely existed prior to evolution of the hippocampus. These cores produce the transmitter substance acetylcholine, which has been responsible for activating not only muscles but also existing neurological structures ever since evolution of primitive animal life. The cholinergic activation system still performs an important function in the brain, even though it no longer directly controls activation of the motoneurons in the muscles.  It therefore made good sense to use output from the hippocampus that had been triggered by scent of prey to activate the cholinergic system. Consequently, the excitation from the hippocampus projected into the septum, thereby enabling the animal to activate the necessary systems quickly when prey appeared or danger threatened.

The fact that cholinergic axons in the septum found the way back into the hippocampus may have been purely accidental. But excitation of the output neurons in the hippocampus was, in any case, of only limited benefit, because they were already excited. So evolution opted for another approach. Cholinergic axons from the septum excited the numerous GABAergic interneurons present in the hippocampus. Nearly every neuron core contains inhibitory neurons, whose principle function is to enhance contrast between output signals. This enabled the septum’s excitatory cholinergic output to reach the GABAergic interneurons in the hippocampus, because neurons automatically seek each other out.

Similarly, it made sense to give some of the GABAergic interneurons in the septum, those that were then and still are responsible for blockade of neighboring receptors and consequently contrast enhancement in the septum, the capability to form longer axons, because these could then form glutamatergic projection neurons in the septum, which would, in turn, block the glutamatergic projection neurons in the hippocampus.

We have decided to use the definition ‘direct blockade’ in this monograph to describe  blockade of a core’s or a region’s output neurons by means of inhibitory projection onto those output neurons. Direct blockade acts directly at the signal’s point of origin, i.e. at the neuron whose output has to be suppressed.

Coining of the term ‘direct blockade’ automatically implies the existence of indirect blockade. We define indirect blockade as the excitation of neighboring inhibitory interneurons located in the vicinity of the relevant output neuron. As already also postulated, nearly all neuronal cores contain interneurons, most of which are GABAergic, whose main responsibility is to block weaker signals in order to enable stronger signals get through, i.e. they block neighboring receptors.

When projection neurons from other cores excite these inhibitory interneurons, the ultimate effect is blockade of  a whole group of neighboring output neurons. We define this as indirect blockade. It enables the use of excitatory projection signals, whose inhibitory action stems from indirect excitation of inhibitory interneurons that, in turn, triggers blockade of the signal that would otherwise have been transmitted. It seems quite possible that these interneurons form a specific category responsible for switching from an excitatory to an inhibitory transmitter and no longer used for blockade of neighboring receptors.

When attempting to describe cytoarchitectural structures and linkages, it is always essential to analyze whether input into any region or core blocks the output-producing projection neurons located there or excites the inhibitory interneurons. The first case is direct blockade, the second indirect blockade. Use of these terms greatly simplifies description of inhibitory activity. But, as there is a third possible variant in the form of excitation of the projection neurons and also a fourth through blockade of the inhibitory interneurons, there are, in fact, at least four possible versions of neuronal contacts. This is the reason why identification of the neurotransmitter and the type of neurons contacted so important for analysis of neuronal circuits.

Thus, the hippocampus caused excitation of the septum and the septum had an inhibitory effect on the hippocampus, thereby using both direct blockade by GABAergic projection neurons and indirect blockade through excitation of inhibitory interneurons.

This demonstrates that the hippocampus and the septal cores already formed a closely cooperating unit back in the early days of evolution history. But why were these cores in the septum able to impose breaks that interrupted the higher-frequency continuous oscillation in the hippocampus? How was this possible?

According to one widely-held theory, the hippocampus and the septum were located at diametrically opposite points in the primeval brain, which still closely resembled the so-called neural tube. This meant that they were (by accident) at the maximum possible distance from each other. Consequently, the excitatory, glutamatergic signals emitted by the hippocampus had quite a long distance to cover before reaching the septum. And the return journey was just as long after they had been switched to GABA, an inhibitory transmitter, in the septum and the system’s topology routed them back to the hippocampus.

Neuronal signals were unable to follow a straight-line pathway through the neural tube, because this contained a fluid-filled cavity. Modern-day brains still have this cavity in a substantially modified form; it is filled with cerebrospinal fluid and is called the ventricular system. The body of the hippocampus still forms the border of the ventricular system in vertebrates and validates the neural tube theory. This border, which also contained nerve cells, had a curved surface that was the only available route for transmission of signals.

Assuming that the neural tube had a radius of around 10 millimeters and that the signals traveled in a semicircular curve from the hippocampus to the septum and took roughly the same route back, the total distance traveled corresponded to the circumference of the full circle. A circle with a 10 millimeter radius has a circumference of approximately 63 millimeters. An action potential traveling at a speed of 1m/s would cover this distance in 63 milliseconds. And this is the time that an action potential would take to travel from the hippocampus to the septum, be switched there to GABA and returned to the hippocampus. Thus, an active oscillation of 63 milliseconds’ duration was followed by an inhibitory GABAergic oscillation from the septum. And this totally suppressed the original signals from the hippocampus – for a period of exactly 63 milliseconds.

The original continuous oscillation was transformed by the strongly inhibitory feedback into a modulated, higher-frequency continuous oscillation with a time interval of 63 milliseconds. A tetanic excitation of 63 milliseconds’ duration is followed by signal-free interval also of  63 milliseconds’ duration. This would give the oscillation’s envelope curve an oscillation period of roughly 125 milliseconds, which would yield an envelope curve frequency of approximately 8 Hertz. All the foregoing numbers should be regarded as figurative.

This modulated oscillation of roughly 8 Hz in the hippocampus is nowadays called the hippocampal theta. It enabled primeval animals to remember an olfactory signal and to pursue prey that was trying to flee. Unlike an uninterrupted continuous oscillation, it does not overtax the neurons because they can take regular breaks of approximately 1/16th of a second, during which they can, for example, produce the necessary transmitter or freshen up their metabolism.

 

Fig. 5: The hippocampal theta of an active signal – a modulated titanic oscillation

 

 

It should be noted that the magnocellular striosomal system operates (coincidentally) at roughly the same basic frequency. In this case too, it would be useful to examine the need to  integrate suitably short intervals into continuous titanic oscillations (see Monograph Parts 1 and 2).

The interplay between the excitatory projection from the hippocampus to the septum and the inhibitory projection back from the septum to the hippocampus together with the time lag caused by the longer axon supplied the solution to hippocampus problem No. 1.

It is a well-known fact that Nature likes to provide fail-safe of backup systems. In the event of temporary or permanent failure of the direct GABAergic projection from the septum to the hippocampus this function could be taken over by an excitatory projection from the septum to the hippocampus. Cholinergic projection neurons in the septum could then excite the numerous GABAergic interneurons in the hippocampus and these would, in turn, inhibit the activity of the glutamatergic pyramidal cells in the hippocampus. This indirect variant is actually used in the mammalian brain to inhibit excitation in the hippocampus in addition to the one just described.

 

Fig. 6: Additional negative feedback from septum to hippocampus to induce intervals

 


 

Die Absicherung der sinnvollen Arbeitsweise einer neuronalen Schaltung durch mehrere Varianten bewirkt eine Verkomplizierung der Erkennung wesentlicher Zusammenhänge. Zur Sicherstellung der periodischen Schwingungsunterbrechung gibt es die direkte und die indirekte Hemmung des Hippocampus durch das Septum. Andererseits wurden auch analog GABAerge Projektionen des Septums in die Amygdala beobachtet. Diese können - falls die Hemmung des Hippocampus zeitweilig gestört sein sollte - die gleiche Wirkung haben: die Unterbrechung der ständigen Signaloszillation durch kurze Pausen. Ebenso könnte die erregende Projektion vom Hippocampus zum Septum zur Sicherheit durch eine gleichartige, glutamaterge Projektion von der Amygdala zum Septum ergänzt werden. Doch genau diese wird in der Realität nicht beobachtet. Dies liegt daran, dass die Amygdala und das Septum zu eng benachbart sind. Signale der Amygdala würden das Septum sofort erreichen. Einer Erregung der Amygdala durch das Septum würde unmittelbar fast ohne Zeitverzögerung eine Hemmung der Amygdala folgen. was die Entstehung der gewünschten tetanischen Erregung sehr stören würde. Daher gibt es nur die hemmende Einwirkung des Septums auf die Amygdala.

Das gleiche Ziel, positive und negative Rückkopplung im limbischen Schwingkreis kann also zur Sicherheit an mehreren Stellen realisiert werden.

In Schwingkreisen können unterschiedlichste Phänomene beobachtet werden. Ein interessantes Resultat von neuronalen Schwingkreisen mit starker Frequenzvervielfachung und positiver und negativer Rückkopplung besteht in der Normierung der Signale. Durch die starke Frequenzvervielfachung auf Grund der hohen Anzahl der Echoneuronen zu einer Moosfaser entsteht eine tetanische Schwingung, deren fast konstante Feuerrate nur noch durch die Refraktärzeit vorgegeben wird. Da die Integrationsneuronen dieser Subsysteme alle etwa die gleichen Parameter haben, haben sie auch alle etwa die gleiche Refraktärzeit. Dadurch entsteht im Falle der Aktivität der zugehörigen Signale in jeder Signalschleife etwa die gleiche Frequenz der tetanischen Schwingung. Andererseits wird durch die negative Rückkopplung diese tetanische Schwingung durch Pausen unterbrochen. Und da die Zeitverzögerungen der Aktionspotentiale vom Hippocampus zum Septum und zurück für alle Einzelschwingkreise etwa die gleiche ist, ist auch die Pausendauer durch die septale Hemmung in etwa gleich groß. Also sind sowohl die Frequenz der tetanischen Schwingung als auch die Pausendauer normiert. Zusätzlich sind die verschiedenen Schwingungen nicht nur normiert, sondern auch lose gekoppelt, also teilweise synchronisiert. Dafür sorgt die rezeptive Nachbarhemmung über die zahlreichen hemmenden Interneuronen.

Diese Normierung wirft viele klassische Vorstellungen über den Haufen. Normalerweise stellen wir uns vor, die Feuerrate zu einem Signal würde mit wachsender Signalintensität zunehmen. Im den limbischen Schwingkreisen trifft genau dies nicht mehr zu. Ein schwaches olfaktorisches Signal mit einer geringen Feuerrate erleidet während des Durchlaufens der Moosfaser eine Frequenzanhebung bis zur Maximalfrequenz, die nur von der Refraktärzeit abhängt. Zwischen schwachen, mittelstarken und starken olfaktorischen Signalen gibt es bei gut ausgebildetem Hippocampus frequenzmäßig kaum Unterschiede! Alle werden auf Maximalfrequenz angehoben und lediglich auf Grund der negativen Rückkopplung von Pausen unterbrochen, deren Pausenlänge ebenfalls normiert ist.

Daher musste die Natur Wege und Mittel suchen, die unterschiedliche Stärke von olfaktorischen und anderen limbischen Signalen anders zu codieren. Genau deshalb erfand das limbische System die Digitalisierung der Signale. Dann konnten verschieden starke Signale des gleichen Geruchs auf unterschiedlichen Binärleitungen aktiv werden, weil jede Signalstärke eine andere Signalleitung benutzte. Und alle aktiven Signale konnten die gleiche Feuerrate benutzen und waren daher innerhalb des Systems normiert. Diese Normierung der Signale finden wir bei Computern wieder. Nicht die untereinander identischen Bitsignale, sondern nur die Leitungsadresse bestimmt, welche Bedeutung ein normiertes Bit besitzt. Eine Signalstärke gibt es bei Digitalcomputern ebenfalls nicht, sondern nur aktive und inaktive Leitungen. Im Teil 4 dieser Monografie – als eigenständiges Werk verlegt – wird gezeigt werden, dass mit der Herausbildung der hemmenden Interneuronen in Neuronenkernen der wesentliche Grundstein für die Digitalisierung gelegt worden ist.

Die GABAergen Interneuronen des Hippocampus dienen auch der Lösung des Hippocampusproblems Nr. 2. War die Beute bereits verspeist worden, so begann deren Verdauung im Magen. Die erhöhte Aktivität im Magen war jedoch nicht nur chemischer Art. Die Rezeptoren im Verdauungssystem wirkten als Messfühler und konnten die höhere chemische Aktivität in eine neuronale Aktivität umwandeln.

Wenn nun geeignete Rezeptoren des Verdauungssystems während der Verdauung stärkere erregende Signale zu den GABAergen Interneuronen des Hippocampus sendeten, wurde das hippocampale Theta der olfaktorischen Signale gehemmt. Die zugehörigen Schwingungen kamen zum Erliegen. Die Aktivität im Verdauungssystem wirkte hemmend auf den Hippocampus und löschte dort die in den Schleifen rotierenden Signalaktivitäten, die von den olfaktorischen Rezeptoren angestoßen worden war. Das satte Tier vergaß die Beute und wurde träge. Dies trifft sogar noch heute zu!

Später konnten auch andere starke Signale des „Gehirnsystems“ den Weg zu den hemmenden Interneuronen des Hippocampus finden, so dass ein ununterbrochenes hippocampales Theta nicht mehr „ewig“ anhielt, sondern nach Minuten oder maximal nach Viertelstunden oder Stunden wegen der GABAergen Dämpfung zum Stillstand kam.

Ein weiterer Aspekt der GABAergen Hemmung durch die Interneuronen besteht darin, dass jedes hemmende Interneuron eine größere Anzahl von hippocampalen Körnerzellen erreicht. Dadurch werden auch viele benachbarte Moosfasern synchron gehemmt, so dass eine Synchronität des Schwingens der Hippocampusneuronen bewirkt wird. Erst dadurch ist das hippocampale Theta überhaupt messtechnisch auffällig geworden. Je mehr Neuronen im Hippocampus synchron feuern, umso besser lassen sich diese Theta-Wellen z. B. im EEG an der Kopfhaut nachweisen. Eine Analogie gibt es im magnocellularen Striosomensystem, wo die vielen Systemschwingungen der Elementarschwingkreise durch dendrodendritische Kopplung der dopaminergen Neuronen in der Substantia nigra pars compacta (lose) synchronisiert werden (Siehe Teil 1 und 2 dieser Monografie).

Die Unterbrechung der Signaloszillation im Hippocampus nach erfolgreichem Beutefang, bei erfolgreicher Flucht der Beute oder durch neue Signale bei Feindkontakt war durch eine externe Erregung der GABAergen Interneuronen im Hippocampus möglich geworden.

Aber das System aus Hippocampus, Septum und Amygdala hatte inzwischen drei Eintrittspunkte für einen möglichen Input. Die Unterbrechung der Oszillation einer ganz bestimmten Signalgruppe konnte auf zwei weitere Arten erfolgen. Anstelle der Erregung der GABAergen Interneuronen im Hippocampus konnte auch das zugehörige glutamaterge Rücksendeneuron in der lateralen Amygdala gehemmt werden. Diese Hemmung musste jedoch zeitlich so lange andauern, wie die Umlaufzeit der Signale betrug, damit alle Einzelechos komplett gelöscht wurden. Die Neuronen, die diese Hemmung bewirkten, bildeten im Verlaufe der Evolution einen eigenen Neuronenkern, der heute auch zur Amygdala gezählt wird. Es ist der zentrale Kern der Amygdala, der hemmend in die laterale Amygdala projiziert, um die Signaloszillation der zugehörigen Signale dauerhaft zu unterbrechen. Erkennungsmerkmal der zentralen Amygdala sind die GABAergen Neuronen, die hemmend zu den anderen Nachbarkernen der Amygdala projizieren. Im Teil 4 dieser Monografie wird sich zeigen, dass diese Neuronen später eine viel wichtigere Aufgabe übernehmen konnten, die den Wirbeltieren einen ungeheuren Intelligenzschub bescherte: die digitale Verarbeitung von Signalen und die Entstehung von Lernfähigkeit. Doch dies soll nicht hier behandelt werden, da es hier nur um die Signalverarbeitung innerhalb des limbischen Systems geht. Die zentrale Amygdala konnte jedenfalls erregende Signale, die sie erhielt, in den hemmende Transmitter überführen und so Stoppsignale zur lateralen Amygdala senden.

Ebenso konnte bezüglich eines bestimmten Signals dasjenige zugehörige GABAerge Projektionsneuron im Septum stark und länger andauernd erregt werden, welches die zusammengefassten Echosignale vom Integrationsneuron erhielt, um sie als GABAerges Hemmungssignal zu den Pyramidenzellen zu senden und deren Aktivität zu unterbinden. Die langdauernde Erregung des GABAergen Septumneurons führte zu einer landdauernden GABAergen Hemmung der zugehörigen Hippocampusneuronen und zur Unterbrechung der Signalrotation dieses Signals.

Und drittens gab es den alternativen septalen Weg, der schon angedeutet wurde. Im Hippocampus gibt es jede Menge hemmender Interneurone, die den Transmitter GABA verwenden. Ihre Aufgabe ist die rezeptive Nachbarhemmung, die zu einer Kontrastverstärkung des Outputs führt und schwächere Signale eliminiert. Cholinerge Projektionsneuronen des Septums könnten daher (wie bereits beschrieben) im Hippocampus die GABAergen Interneuronen länger andauernd erregen. Dies würde ebenfalls zur Hemmung der Signaloszillation führten. Wenn also die cholinergen Septumneuronen längere Zeit von außen dauererregt würden, so bräche die zugehörige Signaloszillation im Hippocampus zusammen.

Entweder GABAerge Projektionen in den Hippacampus oder die laterale Amygdala konnte die Signalrotation durch direkte Hemmung zum Erliegen bringen, oder eine Erregung der GABAergen Interneuronen des Hippocampus oder der lateralen Amygdala war in der Lage, die Signalrotation entsprechender Signale zu beenden. Die Erregung entsprechender Neuronen der zentralen Amygdala bewirkte den gleichen Effekt.

Falls aber die Beute entfliehen konnte, war ihre Verdauung unmöglich geworden. In diesem Falle fehlten die Signale der Verdauungsorgane, die z. B. die Interneuronen des Hippocampus oder der Amygdala erregten, so dass die Oszillation der Beutesignale unterbrochen werden konnte. Hier erwies es sich als zwingend notwendig, über innere Taktgeber zu verfügen, die nach einer gewissen Zeit – der maximalen Beuteverfolgungsdauer – für eine Unterbrechung der Signaloszillation sorgten. Es war anfangs sicherlich kein circadianer 24-Stunden-Rhytmus aus dem Nucleus suprachiasmaticus, aber es gab mit Sicherheit einen Neuronenkern, dessen neuronale Aktivität beispielsweise im Zwei-Stunden-Rhythmus schwankte. Wenn in diesem Beispiel die Aktivität der Neuronen dieses Kerns alle zwei Stunden ein Maximum erreichte, konnte deren Aktionspotentiale die GABAergen Interneuronen des Hippocampus erreichen und dort alle rotierenden Signale löschen. Alternativ konnten die GABAergen Projektionsneuronen des Septums länger andauernd erregt werden, die ihrerseits hemmend zu den Pyramidenzellen des Hippocampus projizierten. Der Taktgeber konnte aber auch über die direkte oder indirekte Hemmung der Amygdala für die Unterbrechung der Signalrotation sorgen. Wer nicht über einen solchen Taktgeber verfügte, vergeudete Zeit und Energie. Er verfolgte tagelang, wochenlang oder gar sein ganzes Leben lang eine Beute, deren Beuteduft ihn vor Tagen, Wochen oder Jahren erstmalig erregt hatte. Die in seinem Hippocampus rotierenden Beutesignale waren also eine fortwährende Illusion. Während er durch die völlig sinnlose Verfolgung dieser Illusion zunehmend Energie vergeudete, konnte er schon bald dem Hungertode anheimfallen, wenn nicht zufällig eine andere Beute mit stärkerem Beuteduft auftauchte. Wer aber periodisch, z. B. alle zwei Stunden, seine limbischen Rotationssignale einfach löschte, konnte still und energiesparend vor sich hinwarten, bis eine neue, fette Beute mit ihrem Duft die Signaloszillation neu startete, die Körperaktivierung in Gang setzte und Verfolgung und Kampf einleitete.

Die periodische Löschung der rotierenden limbischen Signale im Papez-Kreis erfolgt auch heute noch durch Zeitrhythmen, deren Periode von wenigen Stunden bis zu einem 24-Stunden-Rhytmus reichen kann. Die Verhinderung der periodischen Signallöschung z. B. durch Schlafentzug führt bei Lebewesen zu gravierenden Problemen.

Überhaupt erforderte das Leben einen zyklischen Ablauf. Beutefang, Beuteverzehr, Beuteverdauung, Ausruhen wechselten in stetiger Abfolge. Und die reine Existenz des sehr empfindlichen olfaktorischen Sinnes führte zur Notwendigkeit, spezielle Organe zu entwickeln, die solch ein zyklisches Wechselspiel unterstützten. Wer beispielsweise die Reststoffe der Verstoffwechselung der Nahrung kontinuierlich ins Wasser entließ, hinterließ eine markante Duftspur, die genau zu ihm führte. Wer aber geeignete Hohlorgane wie eine Blase oder einen Mastdarm entwickelte, in denen diese Stoffwechselreste zunächst zwischengelagert wurden, konnte dem Geruchssinn seiner Fressfeinde längere Zeit verborgen bleiben, um sein „Geschäft“ unbeobachtet in ungefährlicher Umgebung zu erledigen.

Das Arbeitsprinzip des rückgekoppelten Hippocampus förderte gewissermaßen einen zyklischen Lebenswandel. Signale wurden olfaktorisch, aber auch zyklisch durch Taktgeber gestartet, bewirkten länger anhaltende Reaktionen und benötigten anschließend Stoppsignale, um inaktiviert zu werden. Notfalls übernahmen Taktgeber eine periodische Inaktivierung.

Die periodische Löschung der limbischen Signale war auch deshalb nötig, weil jedes dieser Signale während seiner Rotation neue Aktionspotentiale erzeugte und dadurch Reaktionen aktivierte. Traf ein Aktionspotential vom olfaktorischen Cortex in einer Moosfaser ein, so generierten die CA3-Pyramidenzellen durch Anzapfen dieser Moosleitung eine ganze Folge von Aktionspotentialen. Diese erreichten die Amygdala, das Septum und das Corpus mamillare und erregten die zugehörigen Neuronen, die wiederum mit Aktionspotentialen antworteten. Ein bereits oszillierendes Signal konnte also nur durch ein benachbartes, neueres Signal rezeptiv gehemmt werden, weil eine Umlaufermüdung bei älteren Signalen zur relativen Abnahme der Feuerrate führte. Ansonsten gab es nur noch die Möglichkeit der periodischen Löschung durch geeignete Taktgeber. Diese konnten sich durchaus auf ausgewählte Signalgruppen spezialisieren. So wären Taktgeber denkbar, die beispielsweise ausschließlich Signale des Verdauungssystems löschten, damit das Lebewesen wieder neuen Hunger verspürte. Blieb das limbische Sattheitsgefühl ungelöscht, verhungerte man eventuell, obwohl man ein wohliges Sattheitsgefühl verspürte. Die Selbstreizung von Ratten, die eine elektronische Reizung ihrer limbischen Strukturen mittels Tastendruck dem Fressen vorzogen, zeigt die Gefährlichkeit der Nichtlöschung elementarer limbischer Reize.

Andererseits bewirkt ein „Reset“ (Computertechnik: „Zurücksetzen“) der limbischen Signale z. B. während des Schlafes eine höhere Bereitschaft für die neuen Signale des folgenden Tages. Man erwacht morgens mit „geleertem“ Kopf. Wer dies nicht kann, vergeudet Kraft und Energie in die Auswertung imaginärer, weil längst ungültiger Signale.

Sicherlich war es günstig, die in den limbischen Schleifen rotierenden Signale, speziell die bisher unbekannten, also die neuen, vor der Löschung zu erlernen. Daher entwickelten sich Algorithmen, die während spezieller Schlafphasen (REM-Phasen) zur dauerhaften Abspeicherung dieser temporären Signale im Cerebellum führten. Eine wesentliche Voraussetzung war der signaltechnische Zugang limbischer Signale ins Cerebellum. Die andere Voraussetzung war bereits erfüllt: LTP und LTD im Cerebellum waren möglich geworden, weil diese limbischen Signale bereits in Form tetanischer Erregungen vorlagen, die auch durch die nötigen Kontrollpausen unterbrochen wurden. In diesen Pausen prüften die speichernden Purkinjezellen den Erfolg der bisherigen synaptischen Veränderung der Kopplungsstärke bei den Eigen- und Fremdsignaldetektoren (Siehe Teil 1 und 2 dieser Monografie). So wurden die Signale aus dem limbischen Aktivspeicher in den cerebellaren Passivspeicher überführt.

Zusammenfasseng kann gesagt werden, dass die Neuronen, die eine periodische Unterbrechung der Schwingung im sogenannten Theta-Rhythmus bewirkten, ebenso geeignet sind, die Signaloszillation auch endgültig zu beenden. Dazu müssen sie nur so lange erregt werden, bis jedes der rotierenden Aktionspotentiale durch Hemmung gelöscht worden war.

Aber nicht nur die Signale der entflohenen Beute galt es zeitnah zu löschen. War diese Beute nicht entflohen, sondern konnte gefressen werden, so erzeugten die gustatorischen Rezeptoren während des Verspeisens der Beute entsprechende Geschmackssignale. Diese fanden ebenfalls den Weg zum Hippocampus, aber von dort ebenso zum Magen, wo sie z. B. die Erzeugung der nötigen Magensäure für die bevorstehende Verdauung bewirkten. Diese starke Aktivierung des Verdauungssystems, deren zugehörige Aktivierungssignale ja ebenfalls in der limbischen Schleife rotierten, war am Verdauungsende völlig überflüssig und sogar schädlich. Daher mussten geeignete Rezeptoren das mögliche Verdauungsende ermitteln und mit ihren Signalen die Signalrotation der Beutesignale im limbischen System beenden.

Es gab also durchaus viele Gründe dafür, eine einmal angestoßene Signalrotation zu beenden.

Aber auch die Umkehrung war möglich. Es erwies sich als sehr zweckmäßig, auch den externen Start einer Signalrotation zu ermöglichen. Ein inaktives, aber auch ein völlig neues, anderes, möglicherweise auch ein nichtolfaktorisches Signal (z. B. Hungergefühl) aus anderen Bereichen konnte in den Hippocampus eingespeist werden, indem es beispielsweise geeigneten Neuronen der lateralen Amygdala erregend zugeführt wurde. Diese Neuronenklasse war ja bereits existent. Man konnte den erregenden Input an ein speziell dafür erzeugtes Rücksendeneuron der lateralen Amygdala senden, damit dessen Erregung die Signaloszillation dieses speziellen Signals im limbischen Schwingkreis startete. Ebenso war es möglich, ein Integrationsneuron (CA1-Pyramidenzelle) oder ein Echoneuron (CA3-Neuron) im Hippocampus direkt zu erregen. Drittens konnte man im Septum ein cholinerges Neuron erregen, welches - völlig abweichend vom bisherigen Konzept - die Körnerzellen oder die Pyramidenzellen des Hippocampus erregte. Ob dieser Weg in der Realität genutzt wird, mögen Neurologen abklären. Rein theoretisch steht er offen, bedingt aber die Ausbildung einer neuen Neuronenklasse, die im ursprünglichen Konzept fehlte. Er würde eine Mehrdeutigkeit in das System bringen, die eigentlich als Störung wirken würde.

Es gab jedenfalls mehrere Möglichkeiten, eine Signaloszillation im limbischen System zu starten oder auch zu beenden.

 

Skizze 7: External initiation of an inactive signal and external termination of an active signal

 

 

One of the main functions of the oscillating signal was to keep the memory of the initial input signal alive. The memory of the prey or the danger was preserved for a longer period. During its existence the action potentials from this signal activated the relevant muscles and other physical systems enabling correct reaction to the situation. The signal was present in a type of oscillatory form and was consequently stored, but not statically stored. It was not stored in cerebellar storage cell, as described in Part 2 of this monograph. Storage in the limbic system was dynamic in the form of an oscillation, during which the signal was present and active, as though the stimulus that had triggered it, e.g. characteristic scent of the prey, was still present. In the meantime the actual original scent had frequently faded or even disappeared altogether, and the oscillating signal was merely a memory creating the illusion that the scent was still present.
Even today, a memory is still only a mental concept and therefore, in the final analysis, an illusion. But without this type of illusion the past would no longer exist; there would only be the extremely short present of the immediate moment. In this respect, there is a close relationship between memory and illusion.
Signal oscillation in the limbic system serves the purpose of temporary or preliminary storage. The nearest analogy to this in the computer world is the use of interim stack storage, from which data is retrievable when needed. The limbic system is a neuronal stack memory of the same kind, in which data are stored temporarily in the form of rotations around the various circuits until they are needed again. The oscillations generated by these rotations make them perceptible. An active signal is present in the limbic system in a form defined in this monograph as an oscillation. Whereas present-day computers simply have memories, neuronal systems have two distinct types of memory – the active memory and the passive memory. Storage of a neuronal signal in an active memory generates ongoing computer output. The storage neurons generate action potentials continuously. This constant output only ceases when signal rotation is completely deleted. A pendulum as an example of this. The signal remains active as long as the pendulum is swinging. When the pendulum stops or is stopped, its signal becomes inactive.
In contrast, passive memories emit no output at all when a signal is stored in them. Although the strength of the synapses containing the data is modified at time of input, their neurons deliver no output whatsoever. This type of memory is present in the cerebellum. A passive memory delivers output in one case only – when it receives input that is either identical or very similar to the signal stored in it. Then, and only then, does a passive neuronal memory deliver output.
A passive neuronal memory compares the incoming input with the relevant stored data. If these are identical or very similar, it will deliver an output signal for as long as an adequate degree of similarity between input and stored signal persists. This is roughly how passive electronic computer memories function.
This would perhaps be a convenient moment to provide some explanations of the way in which computer memories work, as experience shows that misconceptions exist in this area.
It would be misleading to compare a computer memory with a cabinet full of drawers containing a selection of articles, e.g. screws. Comparison with a full football stadium or a busy DIY store would be nearer the mark. There is a loudspeaker announcement: “The parents of Johnny Doe from Philadelphia are requested to come to the entrance to pick up their 5-yearold son.” Daddy Doe stands up and runs to the entrance where his son is waiting.
This is roughly how a computer memory works. An address signal (like the announcement stating the name) is input and travels through the address conductors to the memory cells. And the memory cell bearing the same address as the one transmitted via the conductors delivers its content onto the data conductor. The user needs to know which memory cell contains the content he wants to retrieve and specify this in the input inquiry. The computer understands the inquiry as meaning: “What does the memory cell with the address x contain?”
Present-day (year 2013) computers only have passive memories. Active memories comparable to the limbic system have not yet made their appearance in the computer world.
The fact that active and passive neuronal memories exist has consequences for our consciousness. The multimodal image of the world contained in our brain is produced by the active memory. Our perceptions, sensations, imaginations, ideas and feelings are represented by action potentials spinning around and oscillating in limbic circuits and creating signals that interact with each other.
In contrast, our life story is stored in the passive memory of the cerebellum. If this were to be permanently active, we would constantly be watching a film of our past life. This would inevitably have a seriously disruptive effect on our normal everyday life. This is the reason why memories not at present needed are transferred to passive storage. As already explained in Part 2 of this monograph, specific portions of these inactive data are reactivated when the inverse cerebellum converts these data into oscillations and outputs these to the active memory.  It is only when the data have been converted into oscillations that we once again become capable of generating ‘tangible’  memories. A good example of this is the anecdote of the bricklayer’s accident already discussed in Part 2 of this monograph. A bricklayer falls off a high building. As he plummets to almost certain death, his past life flashes past him in countless vivid mind images. A lucky chance saves his life and he is able to tell people about this experience. What had happened was that his inverse cerebellum, activated by the stress situation, delivered these images to the thalamus, which is incapable of distinguishing between external and self-generated input. These signals gained access to the limbic system  in oscillation form and became tangible memories.
As demonstrated in Part 2 of this monograph, active signals in the non-limbic system can also be present in oscillation form, if this is enabled by blockade of neighboring receptors (Malczan’s Oscillation Theorem).
In addition to the aspects already mentioned, another effect was caused by blockade of receptors of the pyramidal cells in the hippocampus. If every hippocampal output neuron were to occupy its own GABAergic interneuron and excite this exclusively with its own output, this interneuron would, when excited, blockade all output neurons within its reach in a fairly large catchment area. In this monograph we use the definition “blockade of neighboring receptors” to describe the process by which output neurons inactivate each other reciprocally inhibition of their receptors. This phenomenon ensures that, from the mass of signals input into the system, only the stronger ones prevail. The weaker signals circulating in the vicinity are either partially or completely suppressed by this blockade of neighboring receptors.
It seems possible that malfunction was involved in the present case. If GABA blockade of neighboring receptors does not occur, the hippocampal system is no longer able to continue inserting short breaks and suppress the weaker signals. This triggers an avalanche of activity resulting in the condition known as epilepsy. Thus, it seems possible that epileptic fits are caused by a disruption of circuits following failure of the interneurons in the hippocampus to perform their GABA blockade function. Failure of the GABAergic septal neurons had the same effect. The periodic interruption of oscillations in the hippocampus ceased and the system started to oscillate at highest frequency without any breaks, i.e.  another epileptic fit. The reason why severe epileptic fits are so serious is that a hyperactive hippocampus causes extreme activation of the septum and the amygdala, and this in turn leads to total activation of all body systems.
This explains why there is the alternative of an inhibitory GABAergic projection from the septum to the amygdala. This enables the amygdala itself to initiate periodic interruption of an oscillation when temporary malfunction of the hippocampus is threatening to make it continuous. This ensured that a fail-safe function was in place, but made it considerably more difficult to understand how it functions.
One special feature of this blockade of neighboring receptors has already been mentioned in Part 2 of this monograph; it enabled the evolution of intelligence and meditation. Blockade of neighboring receptors has a limited radius of action. A GABAergic interneuron can only inhibit an output neuron whose dendritic process is within reach. The area of receptor blockade if finite and usually small in diameter. This is of relevance for the signals in the higher association areas, most of which are spatially scattered. The Potentiation Theorem (see Part 2 of monograph) postulates that the number of signals in a complex increases, and the spatial distance of the neurons in a higher complex tends to increase statistically with the degree of abstraction. Consequently, if signals from a complex in the higher cerebral levels find their way into the hippocampus, they will tend to rotate within the hippocampal system for a significantly longer time – perhaps hours or even days. The more abstract the signal, the longer will it continue rotating in the hippocampal system, because the probability that other similar and stronger signals will arise is statistically low and, consequently, blockade of neighboring receptors declines to a low level.
The question whether signals of this kind from higher complexes reach the hippocampus will be answered later.
Rhythmic tetanic oscillation is an extremely important factor in the hippocampus-septum-amygdala system and we will refer to it in future as the limbic system rhythm. The term “hippocampal theta” is also used.

The existence of periodic high-frequency oscillation paved the way for evolution of another innovation. Neurologists like to describe high-frequency oscillations as tetanic excitation. But tetanic excitation is an essential prerequisite for development of LTP and LTD.
Without long-term potentiation and long-term depression it is impossible for a long-term memory to develop. It has already been demonstrated in Parts 1 and 2 of this monograph how LTP and LTD in the cerebellum lead to learning processes. This monograph still has to show that LTP and LTD are used in the limbic system.
It has already been explained how signals – both olfactory and other types – can be stored preliminarily, i.e. temporarily, in the limbic system in the form of signal rotation. It still needs to be explained how these signals can be transferred to a permanent form of storage. This will be done in Part 4 of this monograph, where it will be explained how the cerebellum commits new limbic signals to memory.
At the end of this excursion into epochs long past we intend to make some suggestions as to how the Papez Circuit evolved. The signals emitted by the hippocampal oscillation system prolonging the memory of prey or warning of the presence of other predators were needed for further processing in the nervous system. On the one hand, the septal system was activated; on the other, output from the lateral amygdala triggered a strong system reaction when attack or flight became necessary, while the central amygdala was responsible for weakening excitation by blockade of the rotating signals.
This limbic system was the first temporary memory in the cerebral system. Its signal oscillation process was so successful that all neuronal signals gradually found their way into the limbic system.

The output from these subsystems, especially from the hippocampus, was collected in a newly-evolving  switching core, the corpus mamillare and proceeded from there to the nucleus anterior thalami, a newly-evolved core in the thalamus specifically designed to handle limbic signals. This projected into a new cortical structure, the gyrus cinguli. This structure is a cortical area ,over which the new cortex of modern vertebrates was later superimposed to form the inner gyrus in the brains of modern-day vertebrates. This closes the circuit, because it is the gyrus cinguli which transmits most of its output to the hippocampus. This small inner limbic oscillation circuit, comprising hippocampus, septum and lateral amygdala, in which signals are temporarily stored by rotation and signal oscillation, acquired as a result of this redesign an additional outer oscillation circuit. Neurologists now call this outer limbic oscillation circuit, consisting of a cyclic linkup of hippocampus, anterior thalamus and gyrus cinguli the Papez Circuit. In the opinion of the author of this monograph, its principal theoretical role as a system responsible for handling information is temporary storage of limbic signals in the form of oscillations in closed circuits, in order to prevent these signals from disappearing into oblivion.
While the inner limbic oscillation circuit mostly handles olfactory signals and signals from the autonomic nervous system, the outer circuit handles signals of a non-olfactory nature that have found their way into the limbic system. Many of the latter are so-called unmarked signals. These are the signals that the cerebellum has not yet been able to commit to memory completely. After being marked in the cerebellum, signals of this kind oscillate between the direct and the indirect cerebellum (see Malczan’s Oscillation Theorem – Part 2 of this monograph) to reach the central amygdala and from there the  lateral amygdala, where they delete the original signals still spinning around the limbic memory in temporary storage. Thus, the only signals rotating in the limbic system are unmarked, active ones not yet committed to memory in the cerebellum. More precisely, they are in all cases the new signals received by the creature. It is for this reason that we always associate the hippocampus and the amygdala with new signals, because it is always in the non-limbic system that signals already committed to memory have their activity phases. Loss of limbic system function equates with loss of learning capability. This is because new signals first have to stay spinning around the limbic loop until they have been committed to memory by recruitment of the neurons required for storage of memorized data in the cerebellar system. These are obtained from the available stock of proneurons.  And this process of transformation of proneurons and formation of axons to link up the resulting neurons (i.e. learning) can last hours or even days. If it proves impossible, the creature loses its learning capability.
Signal oscillation – which is also observed in the non-limbic system (see Malczan’s Oscillation Theorem – Part 2 of this monograph) is a phenomenon that should be examined in an attitude totally devoid of emotion. Although we always associate limbic signals with emotions and feelings, this is because of the significance of the signals that have to be assessed – like, for example, those from the amygdala for a decision to either fight or run away and fight another day. The context-related and emotional significance of the limbic system’s signals must not be allowed to distract attention from the mission actually entrusted to the limbic system by evolution – the preservation of important signals for a sufficiently long period. The limbic system created the first-ever memory based on neuronal circuitry in the history of evolution.
The concept of temporary storage of a signal form by oscillation in a closed loop was so successful that it was also used in the non-limbic system during the subsequent course of evolution, where it was responsible for the genesis of intelligence and meditation, as already demonstrated  in Part 2 of this monograph.
It is now therefore time for a meticulous test of this theory. A test of the theory by the author himself revealed some flaws that will have to be corrected (in Parts 4 and 5 of this monograph). These relate, in particular, to the double and quadruple negation in the basal ganglia system. We can anticipate some new, corrected insights in this area that will help to clear the way for better understanding of the brain’s digital functioning.

In summary, it can be concluded that use of the limbic system and the signal rotation procedure made it essential for all active signals to be interruptible. Hunger created an urge to hunt for prey. A full stomach canceled that urge. Pairs of signals were also formed, in which the two parts were diametrically opposed or, perhaps, complementary to each other – hunger/satiety, desire/aversion, tasty/disgusting, warm/cold, attack/flight, courage/cowardice, love/hate, every limbic signal had an inverse signal representing the other extreme. And all these paired signals could only be stopped by activating the other extreme. Thus, in all pairs of this kind, one of the extremes was always active, and the other inactive. The system could be described as dual or bivalent. And because there were so many of these dual of complementary signals, animals lived in a multidimensional dual world. We therefore interpret the limbic system itself as a multimodal assessment system. And because many limbic signals are extremely complex and contain a mass of sub-elements, we frequently interpret them as emotions. The elementary signals concealed within them are too complex for us to understand. But the fact remains that every emotion in the limbic system has its  opposite emotion and only one of these can be active at any time.
It got really complicated when it became possible to gauge the strength of these emotions, because at some point in evolution the limbic system acquired the ability to digitalize signals. Suddenly, nothing was simply black or white anymore, there was a whole selection of different shades of gray in between. This may sound illogical. But can you always answer questions like “Are you warm enough?”, “Are you hungry?”, “Do you feel well?” with a clear yes or no. There can always be intermediate nuances like: “I feel very warm”, “I’m quite warm”, “I feel a bit cold.” We all know the statement used by the weathermen: “There is a 45% probability of rain”. Percentages are classic examples of intermediate assessments. The arrival of interim values in the limbic system’s formerly dual raster made the world multidimensional and its apparent continuity questionable, because a finite number of neurons can only produce a finite number of interpolations in the form of intermediate values. The apparent continuity of the world and its parameters is a neuronal illusion of the mind perceiving it. Although the world is mostly continuous (unless you are looking at it through the eyes of a quantum physicist), the fact that the brain contains only a finite number of neurons means that it can only be depicted discontinuously in digitalized form. The still elusive secret of how continuous signals from the environment are digitalized in the brain will be reviewed for the first time in Part 4 of this monograph.
Before we start to summarize the insights gained so far, there is one simple question about the limbic system that needs answering: Approximately how many different limbic signals are present in a human being?
If we assume that most of these signals, when active, are spinning around the outer limbic oscillation circuit – the Papez Circuit –, the answer to that question has been lying around for quite a time. Each one of these signals has a projection neuron in the corpus mamillare. As far back as 1957, Powell, Guillery and Cowan calculated that the nucleus medialis of the human corpus mamillare contains some 400000 nerve cells. That figure is stated on page 147 and in Table 163 on page 333 of Blinkow &Glezer’s up-to-date work The Central Nervous System in Figures and Tables. This means that a human being can learn and distinguish between roughly 400000 different limbic signals. Whereby the limbic system is present in duplicate.

 


ISBN 978-3-00-045141-6

Monograph  von Dr. rer. nat. Andreas Heinrich Malczan