Semblance Hypothesis

  (towards a theory since 2017)

How constraints guide towards the solution?

 

Aim of this page is to explain how we can reach the solution using large number of constraints provided by findings from different levels of nervous system functions. This approach is motivated by the methods used in physics to understand particles and fields that we cannot sense directly using our sensory systems. The deep underlying principle of this is based on methods used in linear algebra for solving a system of large set of linear equations that has a unique solution. Here, the relationships between the variables in the equations guide us towards the solution. In mathematics, it is possible to find quick methods to arrive at the solution. In fact, we invent those quick methods. The natural question at this point is that mathematics can develop equations. Neuroscience is different. Yes, in mathematics all the derivations can be carried out even without any equations. Equations were invented by us so that others can compute the results of similar problems very quickly. Always the first person who invent such short cuts need to spend lots of time to design it (In fact, students who study only the equations do not understand the concept behind the process and they will not like mathematics. Once one understands the process behind an equation, one will enjoy it and most likely go for graduate studies in mathematics!). So the point here is that if we are ready to spend time and energy, we can slowly arrive at the solution for the nervous system using the deep principle behind solving a system of linear equations. 

Since in neuroscience, we cannot have such equations or short-cuts, we have to arrive at the solution using the hard way. Since we cannot create an easy equation in neuroscience, everyone who tries to understand the derived solution has to take the
same hard way to appreciate the solution. Since studies of the brain has specialized and super-specialized into large number of levels, those who are interested in understanding how the solution was derived will have to spend time to understand different fields of these specializations. This is a reality. 

Here we will use subsets of disparate findings from the list in Table 2 on the first page of this website. We need to use trial and error methods to reach at the solution. By repeating this approach using different subsets of findings, we are expected to arrive at the same solution, which is expected to be the correct solution. Why do we have this much optimism? The optimism is due to the fact that there can only be one unique solution for the system and since we are using very large number of findings from different levels of operation of the system, it must be correct. At this point one may ask the following questions. “What is the problem with already published work in neuroscience?” Research work in neuroscience have been carried out by examining finding only from few levels to reach a solution. This has been the practice since one person can only specialize in a few levels of studies and journals have space limitations for articles. “What is the problem with already published work in neuroscience that explains synaptic plasticity?” Here, we made an assumption that synaptic connections make changes and it will be responsible for learning-induced changes from which memories are retrieved. This was initially set up not based on any derivations. Now that we have better knowledge of approaching a system that exhibits disparate features at different levels, we are able to derive an operational mechanism. Due to this reason plastic changes anticipated at the synapses become a weak candidate capable of explaining disparate findings from different levels. Reaching the correct solution implies that it can explain findings from all the levels of the systems to such an extent that we will be confident in replicating the mechanism in engineered systems, which should be the gold standard criterion in understanding the system. 

We have to use all the findings from different levels of operation of the nervous system and work hard to find the solution that can remain invariable under all the conditions. It is hard; but this is what we have to do to get to the solution. In this approach, we should be ready for the following. 1) Whatever is the solution that can explain all the findings, we should be ready to tentatively accept it and try to verify it further. 2) Always consider the solution as a hypothesis until we use large number of triangulations to confirm its accuracy. Once we get exhausted and fail to reject the hypothesis, we should be accepting it as further testable hypothesis. 3) Once we agree that there can be no other way that this system can function and is in agreement with all the expected features of an evolved system, then we should be ready to accept it. So let us begin.

In order to become successful in solving a system, we have to include all the variables within the system in different non-redundant linear equations. This is a basic principle for success. Ignoring any single variable will not allow us to solve the system. The main function of the nervous system is generation of first-person inner sensation, within it (which we call as “mind”). Therefore, we have to include a variable for first-person inner sensations within the equations (findings) from appropriate levels for solving the system. Findings from the following levels are to be examined. 1) Systems, 2) Behavior, 3) First-person inner sensation, 4) Electrophysiological, 5) Cellular, and 6) Biochemical. By listing major findings from each level and the major constraint that they bring (Table 1), we will be able to derive a solution for the system.

From the above list, we can see a new level – first-person inner sensations - is introduced. This is of paramount importance in the case of the nervous system. Without this, we will not be able to find a mechanism that generates first-person inner sensations. This is the unique property of nervous system that makes it different from all other systems in the body. So the real challenge is to understand at what locations and by what mechanism does the system generate first-person inner sensations. It will also help us to understand how this function is related to other features of the system – for example behavior. We currently use behavior to study several higher brain functions such as perception and memory and interpret the results. We have been consistently failing to solve the system since we haven’t taken the variable of first-person inner sensation into consideration while solving the system. So here we are using this level and trying to generate equations that contain this variable. What we meant by this in this approach is to define the relationship between findings from different levels with that of the generation of first-person inner sensations. For example, if a drug blocks memory retrieval as evidenced by lack of behavioral motor action indicative of memory retrieval, now we have to consider that the drug is blocking either generation of first-person inner sensation or its connected pathway towards the behavioral motor action. By continuing this approach, we hope to clarify the pathway of generation of inner sensations and its relationship with behavior. We also hope to understand how this pathway is generated during learning so that it can be reactivated at the time of memory retrieval. We should also make sure that the unique solution for the system should be compatible with all the previous experimental observations. All these functions are expected to operate at physiological time-scales of milliseconds. Therefore, we will avoid any delayed operations observed within the system for the purpose of solving the system. Explaining all these will ultimately clarify the mechanism of nervous system functions.

Finding

Constraint offered by the finding

Necessary feature of the solution that can accommodate the constraint
Memories are virtual first-person inner sensations - which can be viewed as hallucinations (a sensory experience of something in its absence). The system should have an operational mechanism to generate hallucinations. (This was also the view of Marvin Minsky, a pioneer in Artificial Intelligence research (Minsky, 1980). Operational mechanism should have a specific feature for generating internal sensations (Vadakkan 2007, 2018).
A finite system has to generate infinite number of internal sensations.  There should be sharing of unitary mechanisms of operation depending on specific shared features of internal sensations induced. This can be achieved by combinatorial action of unitary mechanisms of operation. This allows usage of common shared units of internal sensations for shared features of items and events whose memories are retrieved. This increases the efficiency of the system. In conditions that require infinite number of properties, nature adapts such a mechanisms. For example, variations in the light and heavy chain regions of immunoglobulins (Tonegawa, 1983). Operational mechanism should have a specific feature for generating unitary mechanism for internal sensations.  There should be a mechanism that integrates all the units of internal sensations to generate the first-person internal sensation of perception, memory and other higher brain functions (Vadakkan 2016).
Associative learning can take place within milliseconds. Memory is retrieved in milliseconds of time. Learning mechanism should be able to get completed within milliseconds of time. Memory retrieval mechanism should be able to use the learning-induced change to induce inner sensation of memory within milliseconds of time.
Learning should take place at physiological time-scales of milliseconds. A cue stimulus should be able to induce first-person inner sensations within milliseconds (Vadakkan 2018).
Higher brain functions can operate only at a narrow range of frequency of oscillating potentials recorded from extracellular matrix space. The operational mechanism is tightly associated with the vector components that determine the frequency of these oscillations.
Operational mechanism of both learning and memory retrieval should be associated with the vector components of oscillating extracellular potentials (Vadakkan 2013, 2016, Vaz et al. 2019).
Most learning-induced change will reverse back leading to forgetting. This memory is called working memory. Some of the learning-induced changes will persist for short period of time responsible for short-term memories. Some changes may persist of long periods of time responsible for long-term memories. The learning-induced change should be able to explain changes that are responsible for these different types of memories that are classified based on the duration after which they can be retrieved following learning. It should be possible to demonstrate that most of the learning-induced change is reversible quickly that can then explain generation of working memory during the short period of time before those changes reverse back. Some of the learning-induced changes should be able to demonstrate mechanisms by which they can continue to persist for both short and long periods of time that can then explain generation of short-term and long-term memories during the period of time when learning-induced changes persist (Vadakkan 2018).
    If we can derive a solution that can accommodate all the above constraints, then we are moving in the right direction even though our sensory systems cannot directly sense the first-person inner sensations. If this mechanism can then explain all the remaining features of the system, then it is expected to make predictions. Once the predictions can be verified, we can confirm the mechanism. 

Table 1. List of five unique findings, the constraints offered by them and the necessary feature of the solution. The solution is expected to have all the above four necessary features. A derived solution with all the necessary features can then be examined whether it can explain all the remaining features, such as a) why do the system need sleep? b) what is the explanation for the electrophysiological finding of long-term potentiation and its correlations with memory? When a satisfactory solution is found, it can be further tested to examine whether it can satisfy constraints offered by all the findings listed in Table 1 on the first page of this web site. Only a correct solution can provide explanations for all these findings.

If we further analyze the constraints, we can see that we have not yet searched for a mechanism that generates inner sensations. Therefore, we can reasonably say that we have to discover a mechanism that is not familiar to us. In this context, research community should maintain a low threshold for immediately verifying the validity of the arguments in different hypotheses and if found tenable, they should be subjected to further verification.

References

Minsky M (1980) K-lines: a theory of memory. Cognitive Science 4: 117–133 Article

Tonegawa S (1983) Somatic generation of antibody diversity. Nature 302: 575-581 Article

Vadakkan KI (2007) Semblance of activity at the shared post-synapses and extracellular matrices - A structure function hypothesis of memory. ISBN:978-0-5954-7002-0 Download

Vadakkan KI (2013) A supplementary circuit rule-set for the neuronal wiring. Frontiers in Human Neuroscience  7:170  Article

Vadakkan KI (2016) The functional role of all postsynaptic potentials examined from a first-person frame of reference. Reviews in the Neurosciences (Explained how a neuronal soma is flanked by a large number of internal sensory processing units and their relationship with neuronal firing). Article

Vadakkan KI (2018) A learning mechanism completed in milliseconds and capable of transitioning to stabilizable forms can generate working, short and long-term memories - A verifiable mechanism. Peerj Preprints  Article

Vaz AP, Inati SK, Brunel N, Zaghloul KA (2019) Coupled ripple oscillations between the medial temporal lobe and neocortex retrieve humanmemory. Science 363: 975-978  Article