Monday, April 2, 2018

Reasons for Doubting a Brain Could Do the Super-Complex Encoding Needed to Neurally Store Episodic Memories or Concepts

Scientists have published innumerable papers on memory that used "encoding" in their title. Encoding refers to a supposed translation effect in which information in our minds is translated so that it can be stored in our brains. But the existence of all these papers does not actually establish that any such thing as memory encoding actually occurs. In the past scientists published innumerable papers and essays on caloric, phlogiston and the ether, none of which is now believed to exist. And scientists in recent decades have published many thousands of papers on supersymmetry,  superstrings, and dark matter, none of which has been proven to exist.  The vast majority of papers that use "encoding" in their title are simply papers about human memory experiences and experiments, papers that do nothing at all to show that perceptual experiences or learning are translated into neural states.  The term "encoding" is often used in such papers to give a hard-science aura to research that is purely psychological.  No scientist has ever provided convincing evidence that experiences or conceptual learning are actually translated or encoded in some way that allows them to be stored as neural states. 

The topic of encoding actually gives another argument for disbelieving that human memories are stored in brains. I can summarize the argument as follows: when we analyze in detail what a brain would need to do on a low-level in order to store memories, we find that the brain would need to use a whole set of encoding protocols so sophisticated that they could not have naturally arisen.

To explain this argument I must first explain what is meant by an encoding protocol. To understand the idea of encoding memories, you have to realize that there is pretty much no such thing as “just storing” information. Almost every type of information that we know of involves some type of encoding. Encoding is when information is stored using some particular system of representation. 

Let me give some examples to clarify the idea. Consider the storage of written words. A system of symbols (the alphabet) may be written to store the words. Each time someone types the letters “cat” to mean a cat, that person is using an encoding protocol known as the alphabet.

When words are stored on a computer or smartphone, multiple levels of encoding are involved. First, there is the alphabetic encoding. Then there is what is called an ASCII encoding. The individual letters are converted to numbers, using an ASCII table that signifies how particular letters will be stored as particular numbers (an example of an ASCII table is below). Then there is a binary encoding by which numbers such as 34 written in the base 10 system are converted to a series of 1's and 0's such as 100101101.

ASCII is an encoding protocol

 We see in the diagram below three different types of encoding going on when the word “cat” is stored on a computer.

When an image is stored on a computer, there are also multiple types of encoding going on. The image is broken down into a grid of pixels; each pixel is translated into a number indicating a particular shade of color; then those numbers are translated into binary form that is ideal for storing on a computer hard drive.

image to binary
Encoding of a visual image

Now let us consider how the brain might store memories. Just as there is pretty much no such thing as “just storing” information on a computer, without using some type of encoding, there could be no such thing as “just storing” a memory in the brain. If we are to imagine that the brain stores memories, we must imagine that the brain somehow does encoding to get the memory stored to the brain. 

It is hypothesized that the brain might store memories using things such as molecules or nerve cell electrical states. But just as information must always be encoded before it can be stored on a hard drive, there would need to be encoding before any memory information could be stored in something like molecules or electrical states. For example, we cannot imagine that a molecule would ever directly store some letters that we could see using an electron microscope that might show a tiny little “Y” and “G” on a molecule. Nor can we imagine that if we scan your brain with an electron microscopic, we would find tiny images representing scenes from your life, images that could directly be seen in the electron microscope's closeup photos. Instead, if the brain stores memories, there would need to be some low-level encoding causing the memories to be stored in a way that can efficiently be represented by molecules or nerve cell electrical states or neuron connections.

But here a very great difficulty arises. The problem is that human memories include an incredibly diverse collection of things. Consider only a few of the types of things that can be stored in a human memory:
  • Memories of daily experiences, such as what you were doing on some day
  • Facts you learned in school, such as the fact that Lincoln was shot at Ford's Theater
  • Sequences of numbers such as your social security number
  • Sequences of words, such as the dialog an actor has to recite in a play
  • Sequences of musical notes, such as the notes an opera singer has to sing
  • Abstract concepts that you have learned
  • Memories of particular non-visual sensations such as sounds, food tastes, smells, pain, and physical pleasure
  • Memories of how to do physical things, such as how to ride a bicycle
  • Memories of how you felt at emotional moments of your life
  • Rules and principles, such as “look both ways before crossing the street”
  • Memories of visual information, such as what a particular person's face looks like

Now given all these types of things that can be remembered, there would seem to be two possibilities if the brain is storing all your memories:

Possibility 1: The brain has a single “all-purpose” encoding system that somehow works for all of these extremely diverse types of memories.
Possibility 2: The brain uses many different types of encoding protocols; and when it is time to store a memory, the brain somehow figures out the appropriate encoding protocol to use.

The problem is that neither of these options seem credible.

Given the incredible diversity of the various types of memories listed above (which is not even a complete list of all the types of memories that can be stored), it seems unreasonable to imagine that there could ever be any universal “all purpose” encoding system that could work with things as dissimilar as memories of smells, miscellaneous learned facts, principles of living, and long sequences of musical notes or letters. It would seem that many of the types of memories would require their own distinctive type of encoding system.

But it also seems very hard to imagine that the brain could encode memories by using a variety of encoding schemes, and selecting the appropriate encoding scheme to use at the moment that a memory is stored. The problem with that idea is that it requires for us to believe that the brain subconsciously is able to instantly figure out what type of information a memory is, and to then select the appropriate encoding scheme. That would seem to involve some incredibly sophisticated instantaneous analysis, and how could such a thing all be going on subconsciously?

It seems very hard to believe, for example, that if you meet some young woman at a party, and ask for her phone number, that behind the scenes your brain subconsciously is acting like this: well, well, I have a face to store, so let me use the “visual data encoding system” I have stored in such and such a place; and for the phone number let me use the “numerical sequence encoding system” I have stored in some other place; and to store her name let me use the “alphabetic data encoding protocol” I have stored in some other place; and to store the scent of her perfume, let me use the “smell encoding system” I have stored in some other place. All this classification and selection seems to be far too much work to be done subconsciously. Nor can we believe that a selection of encoding systems occurs consciously, for we never (or virtually never) have any conscious thoughts about what type of encoding protocol to use when we remember things.

In short, explaining how encoding could work on a low-level to store memories is a practical nightmare. Our psychologists say that some encoding is going on when memories are stored, but none of them has succeeded in presenting a plausible and detailed description of exactly how such encoding could work. When discussing the issue, psychologists and neurologists will typically combine speculation and tangential findings in a way that will skillfully hide their lack of any mechanism for how such encoding could occur.

When we consider the storage of multifaceted or abstract ideas, it is hard to think of how encoding could possibly work. How, for example, could your brain possibly do some encoding that would represent your mother or the concept of treating people equitably or the idea of the United States as some series of electrical states in the brain, some molecular states, or some combination of the two? We cannot imagine any such encoding.

How can we escape such difficulties? We can abandon the materialist idea that your memories are all stored in your brain. We can believe that our minds are some mysterious spiritual or soul reality, and that memories are part of such a reality. If we no longer have to believe that all memories are being stored in particular parts of the brain, all of the difficulties in explaining memory encoding conveniently disappear. We could still continue to believe that our brains store some things that are rather loosely called memories, such as what are called motor memories, remembrance of how to do particular movements of parts of the bodies. But by believing that conceptual memories and experiential memories are part of some non-physical reality that the human mind is part of, we free ourselves from the seemingly impossible burden of having to explain how the brain could possibly have naturally developed encoding protocols that could represent such things.

We know of one encoding protocol that is definitely used by living things: the genetic code. This is the protocol by which particular amino acids are represented by particular trios of nucleotide bases. The genetic code is an encoding protocol roughly as complicated as the ASCII protocol. Below is how it is represented by Genomics for Energy and Environmental Science, US Department of Energy Office of Science:

But how did this genetic code ever appear naturally? This is a major explanatory nightmare for naturalists. The natural origin of the genetic code has never been plausibly explained.

But consider how great are the difficulties of a materialist who believes all of your memories are stored in your brain. It would seem that for memories to be stored in the brain, it would have to be that there naturally arose not just one sophisticated encoding protocol (the genetic code), but many such sophisticated encoding protocols –all the different encoding protocols needed to store our many different types of memories in our brains. The theorist who claims that your memories are all stored in your brain is like someone burdened by ten albatrosses hanging from his neck, each albatross being the burden of explaining how one of these encoding protocols could have naturally arisen. I haven't even mentioned the issue of decoding, which further doubles the explanatory burden of someone believing that the brain stores all your memories, by requiring such a person to believe that the brain also has a whole set of decoding protocols that are mirror images of the encoding protocols, decoding protocols used in retrieving stored memories that had been encoded using the encoding protocols.

It can be powerfully argued that if we have to believe that all of these sophisticated encoding protocols exist, this should make it ten times more difficult to explain how humans naturally came to exist. But there is an alternative. By believing that the mind has a spiritual reality or soul reality, we free ourselves from the burden of having to explain all these sophisticated encoding protocols that we would need to store all our memories in the brain – for we no longer have to believe that our memories are all stored in the brain.

I may note that there is zero direct evidence that brains or neurons actually do anything like an encoding of information as part of storing memories to brain. An article on a memory experiment states, "Press a scientist to tell you how memories are encoded and decoded in the brain, and you’ll soon find that the scientific community doesn’t have an answer." A psychologist notes, “Misleading headlines notwithstanding, no one really has the slightest idea how the brain changes after we have learned to sing a song or recite a poem.” A 2010 book by two neuroscientists says in its preface, "One can search through the literature on the neurobiology of memory in vain for a discussion of the coding question: How do the changes wrought by experience in the physical structure of the memory mechanism encode information about the experience?" 

When information is encoded, you may not be able to understand the encoding, but you can at least almost always tell that some type of encoding protocol was used (for reasons such as symbol repetitions). For centuries Europeans were unable to decipher the hieroglyphics of the ancient Egyptians, but at least they could tell that hieroglyphics were encoded information. There is nothing at all in the brain that we can point to and say, "This microscopic thing has some of the hallmarks of stored encoded information." Do a Google search for "signs of encoding in the brain" and you will find nothing referring to someone who found evidence that some matter in the brain showed a sign of encoding. 

We know for sure that there is a simple type of encoding that goes on in human cells: the encoding needed to implement the genetic code, so that nucleotide base pairs in DNA can be successfully translated into the corresponding amino acids that combinations of the base pairs represent.  To accomplish this very simple encoding,  the human genome has 620 genes for transfer RNA.  But imagine if human brains were to actually encode human experiential and conceptual memories, so that such things were stored in brains. This would be a miracle of encoding many, many times more complicated than the simple encoding that the genetic code involves. Such an encoding would require thousands of dedicated genes in the human genome. But the human genome has been thoroughly mapped, and no such genes have been found.  This is an extremely powerful reason for rejecting the dogma that brains store human experiential and conceptual memories. 

If you do a Google search for "genes for memory encoding," you will see basically no sign that any such things have been discovered, other than one of those hyped-up press stories with an inaccurate headline of "100 Genes Linked to Memory Encoding." The story is referring to a scientific study described by this paper, entitled "Human Genomic Signatures of Brain Oscillations During Memory Encoding." The very dubious methodology of the authors was to get data on gene expression, and try to see how much it correlated with oscillations in brain waves.  Out of more than 10,000 genes studied, the authors have found about 100 which they claim were correlated with these brain wave oscillations. They state, "We were successful in identifying over 100 correlated genes and the genes identified here are among the first genes to be linked to memory encoding in humans." But the correlations reported are weak, with most of these 100 genes correlating no more strongly than .1 or .2 (by comparison, a perfect correlation is 1.0, and a fairly strong correlation is .5). The results reported do not seem any stronger than you would expect to get by chance. We would expect that even if there was no causal relation between gene expression and brain waves, that if you compared gene expression in more than 10,000 genes to brain waves, you might find purely by chance a tiny fraction such as 1% of the genes that would look weakly correlated to brain waves (or any other random data, such as stock market ups and downs).   In one of the spreadsheet tables you can download from the study,  there is a function listed for each of these roughly 100 genes, and in each case it's a function other than memory encoding. So such genes cannot be any of the thousands of dedicated genes that would have to exist purely for the sake of translating complex conceptual, verbal, and episodic memories into neural states, and vice versa, if a brain stored memories. No such genes have been identified in the genome. 

The paper confesses, "All gene expression data are derived from different individuals than the ones that participated in the iEEG study."  This means the paper is an absurdity.  It is looking for correlations between one set of gene expression data measured in one set of individuals and another set of brain wave data measured in an entirely different set of individuals at a different time. That makes no more sense than trying to look for a correlation between some meal consumption in a 2016 woman's softball team and tooth decay rates in a 2017 male football team. 

No comments:

Post a Comment