Introduction to the Study of the Neural Basis of Action and Thought

The excitability of neurons is due to electrical and chemical phenomena. Motion of ions is described by currents and potential, moreover intricate chemical cycles are described by amplitudes and delays. Starting from what is usually known about synapses and Hodgkin-Huxley axon, we describe first all the parts of a standard neuron (from its input till the input of the following neuron). Then, looking at various physical constraints, we generalize to any kind of neuron. Thus, the synthesis of the physicochemical properties of the cell enables us to compute when spikes occur. Now, we show that the firing rate is the significant nervous message. So, from the input – output relations, we are able to compute the processing abilities of any neuron (some of them make linear additions; others exhibit an ON-OFF behavior, and so on). The next step is to evaluate limitations caused by noise and to study little sets of neurons (of simple animals or very well localized in man, for instance the center controlling breathing). Then, we begin to look the three main functions of huge neuronal sets: to code sensorial message, to choose between competitive signals, to modulate other neuronal sets.

From the simplest animal to the man, the structure of neurons does not strongly change. But the number of neurons increases continually: sea anemone owns 1 neuron, frog and honeybee 10⁷, man 10⁷ with a three layered brain (limbic system, thalamus, cortex). Although human behaviors are very sophisticated, frog behaviors can be described by simple software. Thus, we had to compare brain with computer. Neurons are 10⁷ times slower than electronic chips. Brain capacity for data storage is very weak when compared with the newest machines. Thus, books and magnetic memories appear as extensions of human memory. But the massively parallel circuitry in brain, the specialization of cortical areas and the modulating interactions between the three layers allow to animals and men a global efficiency not achieved by machines. Artificial intelligence (which mimics the human one) allowed for instance the building of almost autonomous automatic pilots. But a man has to intervene to choose the destination. Self learning machines have been conceived for instance to read postal codes. But men have to furnish them samples of each figure: robots can do anything except to choose the goal of their working.

External and endogenous phenomena are detected by sensors. Most of them are just modified neurons. Some others (in human retina, in ears) are hypersensitive cells. Sensorial data are processed in the brain which send orders to the muscles. From the signals of a great number of olfactory receptors, brain extracts the ordered list of the six more strongly excited. Man can learn to recognize 2 10³ smells (and dog 10⁶). Eyes of bees or frogs have poor keenness sensitivity. Human’s retina has very large keenness and sensitivity. In frog, visual signals are processed on the retina. In man, they are sent to the brain and processed by several parallel paths. Adaptable software governs inborn or learned behaviors (for instance walking). Only one out of several competitive behaviors (for instance the various possible directions of eyes) has to be chosen. Choosing centers (for instance superior colliculus) select the more strongly excited among several competitive instinctive drives, which are sophisticated combinations of several sensorial signals. Behaviors control the muscular activity. They are simple automatisms in frogs. In men, they consist in layers of control devices. The lower-level unit is based on two antagonist muscles and their embedded sensors. This unit governs muscular tonus. The cerebellum is in charge of the phasic excitations. Various signals combine to build a representation of the body position.

Learning occurs when excitations of a neuron change its excitability. At the cell level, some effects are observed just after birth. From then two phenomena occur: the first, which governs for instance sleep duration, is the adaptation of the neuron metabolism to its needs (learning delay: one month; instantaneous reading; forgetting delay: one month). The other, which explains pavlovian linking, is supported by special post-synaptic receptors (learning delay: 10 minutes; instantaneous reading; forgetting delay, several years). At the system level, recording needs the simultaneous excitation of a conditional, a conditioning and an emotional (limbic) modulating pathway. Circuitry of recognizing flowers device in honeybee is very simple. In Pavlov’s dog, learning has to link a great number of possible conditional pathways to a great number of possible conditioning ones (a problem solved by the building of the web to link a great number of phones): the switching circuitry uses many elementary Pavlov’s link. Disappointment (a modulating signal) causes differentiation. Sequential links allow memorizing a melody or a sequence of words. If supplemented with a backward transmission, such circuitry generates recollection, building up of abstract concepts and partial recollection of never transmitted signals, known as imagination. Links between real or virtual sensorial inputs and motor orders play a main role. The abstract “me personally” is an abstract concept resulting from links between motor orders and report of an environmental change.

Social instincts need an exchange of messages. In animals, real or imaginary heard signals act as instinctive drives to trigger a behavioral response (flight, attack). In men, the labels of real or imaginary heard sequences of phonemes are words and words act as instinctive drives to utter other words. Real or imaginary speeches allow long term strategies.

A man can learn and recognize about 40 phonemes and less than 10⁴ words (when modern dictionaries quote 10⁴ words). The output of a word device is used as a partial instinctive drive. The total drives result from a fluctuating mosaic of partial drives, a word calls for another word. Sentences are seldom stored in a long term memory, but compacted in a short-term memory. Primary words are related to motor behaviors. Others are abstract concepts related to primary words or abstractions built from abstractions. They allow the bottom-up building of long term strategies. Their results are evaluated by downward imaginative tests. The frontal cortex confronts the immediately most pleasant behavior (insulting my boss) and the long term most pleasant one, explaining how works self-control. Mechanisms of language explain fiction, creativity and rational thinking. In dog as in man, another complex motosensorial device governs complex motions. Simple motions play the same part that phonemes, simple actions replace words and complex actions replace sentences. Combining language and complex motions, man becomes able to conceive new actions, to invent and use tools. Human consciousness is possibly related to the imaginative working of these two moto-sensorial devices. Then, dog’s consciousness would be related to its imaginative motions.