Federal state autonomous educational institution of higher professional education “Southern Federal University”. Taganrog, Russia. firstname.lastname@example.org
The clone hypothesis has been proposed after algorithmisation of radio assay data of the cerebral cortex genesis and interpolation of these laws to those early proliferation stages where the laying of the hierarchical structures takes place starting with receptor and motoneuron levels.
According to our hypothesis one cell division step increases the number of the cells four times. If we satisfy these rules we shall obtain the structure of nearest neighboring cells originated from the same predecessor.
The neuroelement in the suggested network is a typical threshold formal neuron which is from the point of view of Ramon-y-Cajal the model for the neuron as a structural and functional unit of the nervous system and the model for the synapse as an apparatus of the interneuron connection presented by Sherrington. The formal neuron in the clone network is complemented by the rhythmical activity generator (RAG) and system-organizing inputs.
It is well-known that biorhythm is a very important functional manifestation of living forms. By the present time biorhythm logy is one of the parts of biology.
In the model the RAG operates in some individual cycle where the neuroelement can be in the electric activity state or in the memory ability state. The system-organizing inputs simulate control actions typical of the whole organism. The motivation (M) and reinforcement (R) signals analogous with those in learning and self-learning in animals come on them. The motivation signal in the given model changes weight coefficients of the synapses at L current clock periods for those neuroelements which at those moments of time were and are in the memory state and the sums of weight coefficients of which has exceeded the threshold value to generate the impuls. The reinforcement keeps modified only those synapses that belong to the preceding L clock periods.
The clone reflectory network performs the transformation of the spatial-temporal pattern of the input image signals into the spatial-temporal pattern of the control signals of the working organs by the way of the connection passage .
1. Стадников Е.Н. Клоновая нейросеть на основе нейропроцессора NM6403. Нейрокомпьютеры, №5, Москва, 2005.