User:Graeme E. Smith/Allocortical areas, Implicit Memory with no Explicit Access

Recently there has been some work done mapping the areas of the brain that are Isocortical or the standard granular 6 laminae cortex structure, Allocortical or not the usual 6 laminae structure, and Agranular or having little or no fourth layer in an otherwise Isocortical structure. It has been noted that in the case of the Allocortical areas of the brain many not of the hippocampus have been found to have a three layer structure similar to the first three layers of the Cerebral Cortex. It has been suggested that this might indicate a precursor or archeological remnant of a previous cortex design that has been supplanted in most of the cortex areas with Isocortical architecture. In this article I propose the hypothesis that the Allocortical areas are simply areas that have only an implicit memory component and do not have an explicit address for place-code addressing.

        Because this article mixes terminology from more than one specialization, I will first attempt to show where the terminology comes from, so that people who are specialized are not put off by new concepts that do not make sense from their own specialist perspective. Essentially this article has to do with the Anatomical one might say Micro-Anatomical structure of cortex tissue. Classically the cortex has been divided into six layers or Laminae numbered I to VI from the outside in which is usually also thought of as the top - down toward deeper structures in the brain. This structure, it seems, is created by the ascension of the pyramidal neurons, which probably in fetal tissue starts off as a much smaller number of layers.
        It is important to note that the thickness of the cortex is primarily affected by the length of the Apical Dendrites of the pyramidal neurons, and thus areas of longer Apical Dendrite populations, have thicker cortex layers. The difference being filled in as it were with glial cells. If this is true, then the number of laminae is probably also truly just a function of the lengths of the apical Dendrites of the pyramidal cells.
        However, Cortex tissues have been classified, by their micro-anatomical structure resulting in a numbering system called Brodmann Areas that defines islands of different micro-anatomical structures that spread throughout the Cerebral Cortex at least. From this a formal classification nomenclature has arisen, that uses common terms for similar types of micro-anatomical structures. The classifications are: Isocortical or Granular areas, which have the standard 6 laminae cortex structure, Agranular areas which have a missing 4th layer or Laminae IV, also called the granular layer, and Allocortical areas which no longer maintain the 6 layer structure, many of which have three layers, and some of which notably the other Pyrimidal areas such as in the hippocampus, are coiled or wrapped around each other and no longer take on the aspect of tissues.
        These terminologies attain to the Micro-Anatomical structure of cortex tissues and are therefore the meat of one specific branch of Neuro-Anatomy. It should be clearly noted that when I talk about Allocortical Tissues, I am not necessarily talking about the Allocortex, Which is part of a different branch of Neuro-Anatomy that labels areas that have a particular type of micro-architecture. The difference is that there are many locations for Allocortical tissue in the brain, and only most of them lie in the Allocortex. Now we must add some ideas from Psychology.
        It has long been known by psychologists that there are two types of memory, The standard or conscious memories and a group of memories that is subconscious or unconscious until our attention is paid to them. Recent work has indicated that these subconscious memories are at least partially responsible for the richness of our thoughts, in that they flavor even the dry facts we attempt to recover from conscious memory. Gradually the terminology has shifted over time, so that we now call Subconsciuos memory implicit memory, and Conscious memory explicit memory but a case can be made that the new terms are necessary because the classification of conscious memory and subconscious memory, is too subjective in that it involves consciousness which is not considered germane to the nature of the memory, but is germane only in that that is the way the differences in the types of memory were detected. Care should be taken not to mix consciousness back into discussions about implicit memory and explicit memory, because while there is a connection, it is more in the nature of the fact that consciousness uses both types of memory in different ways, than any requirement for how the memory should work.
        So, why one might wonder is there two types of memory? We can label these two types of memory by information theory as content addressable and place-code addressable memory, but that is just a label for the way that the information stored in the memory is accessed. Content addressable memory is accessed by testing each memory location for specific content such as a particular stimulus, and place-code addressable memory is accessed by giving the place-code or address for a specific memory location, and demanding from that location it's contents. Why wouldn't there be just explicit (Place-code addressable) memory, instead of this strange add on that hides information from us? In fact that is a question most computer guys ask, when confronted with implicit memory, after all in computer technology, explicitly addressing memory is so basic to the design of the computer that it is almost taken for granted
        Some of us, working in the field that are familiar with both computers and network simulations of neurons, are beginning to get an inkling. In our opinion it is the question that is wrong, not the implicit memory. In matter of fact, building an implicit memory with neural networks seems to be a simpler problem than building an explicit memory.^[1] Perhaps we propose, the problem is one of the technology by which the memories are built, perhaps, just perhaps the implicit memory came first, and explicit addressing was an add on. Of course this heretical idea is all couched in philosophical terms like "phenomenal", and output of it is described as a Quale, and therefore seems remote to even neural network afficiandos, but I suggest that such a hypothesis, might be born out in the micro-anatomy of memory areas such as the cortex and hippocampus. And, if what I suspect about the nature of the architecture of implicit memory areas in the brain is true, then, by studying Allocortical areas of the cerebral cortex, we can perhaps shed some light on the subject.
        Hypothetically, speaking of course, The three layer allocortical areas represent implicit memory. The 4th or granular layer was an add on created to allow the brain to reroute most of the sensory systems via the thalamus, at or after the time the 5th and 6th layers ascended. If this is true, then all allocortical areas will be bereft of granular layers and only Isocortical areas will be classified as granular and agranular depending on the development of the 4th layer.
        If we look at the Allocortical areas that have three layers we see that they are layers 1, 2, and 3 also known as Laminae I, II and III respectively, suggesting that if allocortical areas are more basic, they may have evolved first. According to Eccles^[2] Laminae I is mostly a tangled web of connective processes with few neurons in the adult brain. According to Marr^[3] Laminae I consists mostly of Mossy Dendrite Neurons, I have taken this to mean that mossy dendrite neurons are more prevalent in the Immature brain and thus may play a role in learning.^[4]
        Layer 2 (Laminae II) and Layer 3 (Laminae III) are populated with pyramidal cells, and Martinotti cells. Martinotti Cells seem to have an inhibitive role that impacts on signals passing through layer 1, while Layer 2 pyramidal neurons have shallow, and Layer 3 pyramidal neurons Deep connections within the brain. A recent article has noted that pyramidal neurons in the cerebral cortex are more vertically oriented than pyramidal neurons in the hippocampus areas. This suggests a possible overlap of function between layer 2/3 cells, that might mean that the dual layer architecture merely indicates a need for multiple sources for the same essential signal. If so, we can treat layers 2 and 3 as a unit. which is where the layer 2/3 nomenclature comes from.
        An Alternate explanation, is that layer 2 memorizes the stimuli, and layer 3 deals with relative temporal relationships within the stimuli. This is certainly thought to be the case in the cerebellum which while supposedly Isocortical in construction still seems to have a similar bottom three layer architecture except for some fibers that connect directly from the Inferior Olive to the Purkinje cells. However in the cerebellums case a stronger case can be made for sequences of memories if only because of its purported role in Skill Memory which is presented as sequences of actions, triggered by context.
        If Marr's presentation about the first 4 layers of the cerebral cortex, is correct despite its tendency to assume mathematical functions for the Layer 4 neurons, Allocortical areas define a self-classifying content addressable memory. Such a function would trigger a rich response to stimuli, creating a parallel datafield containing all the subclassifications under which a particular stimuli can be detected. Such a datafield could be called a Quale, because the nature of the neural networks from which it was built, would not allow any form of addressing within the allocortical area except that of content. This is important because it means that output from allocortical areas, can't be subdivided it has to be taken as it is, a parallel data-field filled with redundant data, in a poorly organized form, and retrieval of the information is impossible except by experiencing the same stimuli again. This non-partitionality and non-retrievability is exactly the characteristics of implicit memory, though, once a stimulus has triggered a memory, the implicit memory often informs the retrieved memory by adding to the richness of the resulting data.

Conclusion

        Allocortical memory of the three layer type, and Isocortical memory, both share the top 3 layers (Laminae I, II, III) without any noticeable micro-anatomical changes. This suggests that in fact the primary difference between Allocortical memory and Isocortical memory is the ability for Isocortical memories to be retrieved in some manner other than a content addressable format, where only the same stimulus can trigger the memory. The hypothesis is therefore that Allocortical memory is implicit memory with no explicit access, and Isocortical memory is implicit memory with an added capability of explicit addressing.