The evolution of the human brain has been the defining factor in creating the superstructure that enables consciousness. Generally as one moves to the anterior of the brain, the higher the cognitive abilities.

Of course this is not always the case as some areas are efficient at what they do and are left in situ and improved. The cerebellum was pushed to the posterior by the emerging expansion of the newer brain mass. It was left intact as it controlled the life support and fast reactive necessary actions. It still contains a very large number of neurons but not the connectivity of later structures. As a crude partition device, you could divide the brain into three areas, the ‘thinking’ part, the data consolidation part and the sensor accumulation part. Of course the brain is highly connected and many actions are duplicated, shared and enriched by other areas. There are many factors that created the human level of higher cognition which is thought to include better nutrition (cooked meat for one), a diverse social structure and its lack of protective body parts. Other animals use tools as food utensils but none have evolved the sheer killing power that more than compensates for the lack of sharp parts.

It could be said that consciousness is at the top of a scale that includes sentience, cognition and self awareness. There is a definite correlation between ‘brain power’ and density of neurons and networks. Some animals may have the potential to have higher brain potential but have no pressing need to fuel the high demands that it would entail. A classic case is the pilot whale that has a large brain but is efficient at finding food and lives in a generally benevolent environment. The usual benchmark on cognition is what evolutionary level a species is at present and historically. There is a widely held belief amongst philosophers that we live in the  21st century with a stone age brain. This shows that cognition can expand past its logical and knowledge boundaries but not its emotional ones. We make better and better technological marvels but our interaction with each other and the environment is still stuck in the past. Usually articles start at the bottom and work their way up. As consciousness is the most interesting of a whole fascinating subject, we will start there. Both human intelligence and ‘AI’ are contentious and fast moving subjects that can shift continuously.

When I first thought about trying to define thought in a purely organic way, I read many discourses on the subject but was sorely disappointed. Of course one could say that we have no definitive answer but this did not seem to stop people writing reams about the subject and basically say nothing. It also true to say that others have failed to mention the subject but were not afraid to voice their opinion about subjects that were a lot easier.  I believe that we should look at this phenomena with the clues that are present and  the fact that we are the most advanced cognitive animal with the most advance brain must highlight the fact that these things are related. Resources are limited in the real world and organisms do nothing unless there is an advantage in keeping a characteristic. Mutations may cause unnecessary parts forming but they are wheedled out in the long run by predation or other exterior factors. My own thoughts (I know), are highlighted in the Brodmann Area 10 page.

There seems to be a great drive in AI to further certain ‘weak’ algorithms as they have produced some fairly good results. One of these is artificial neural networks, in the form of GANs etc. This is like trying to make a better and better diesel engine. The enormous resources thrown at these and the dependency on more and more data, even if its made up, seems almost maniac. It is estimated that Google spent $35 million dollars on winning the Go challenge, which was mainly to prove a selling point, as winning at Go is not going to change the world. The fervour and addiction of the ‘shiny toy’ syndrome has stopped designers stepping back and thinking, “what are we trying to do here?”

AI and the brain have basically the same goals, to recognise patterns and model or react to a certain input. Some feel that the two are so different that there is no longer any advantage in studying them and learning from each. This is a very short term approach and ignores the information that each one holds. Natural neuron structure was the blueprint for the original insight into artificial neuron architecture (ANN). To avoid millions of years of organic innovation would be folly. AI offers the advantage of trying out different techniques without killing something. Technology still cannot show us how small groups of neurons work, we can look at one (usually a squid) or many (voxels) but not say a 100 linking to each other. AI has the problem that ANN produce a black box that at present only allows us to see the inputs or the outputs. One thing they have in common is the linking of neurons that change the strength of connections (weighting). Of course there is a whole range of ANN free algorithms such as Bayes, Decision Trees, SVM etc. that are more visible and faster.

There is a lot of work being done on relating actions to parts of the brain. This has been very beneficial in Parkinson’s disease and others but does not explain what is going on under the hood. Advances in genetics and epigenetics also help us understand how neurotransmitters and hormones affect the human body. Neurons are really small and even now we can either work on a single one that is found in the squid or many (voxels) but not in the tens or hundreds. Also C. elegans is a roundworm with approximately 300 neurons and 7000 connections and we still cannot explain how it moves. The many successes sometimes hide the fact that we have a long way to go in our hopeful discovery of what makes things tick. Other techniques such as microRNA, CRISPR, protein and virus substitution and transportation still show us that nature is best.

See me here

Copyright © 2020 | Maldwyn Palmer