Biological life has a great deal of interdependence between the layers of the biological system in an organism. The architectural framework section suggests that only parts of these interdependencies have been characterized. 5. How many types of neurons exist? Research is still actively exploring this challenging question to quantify the result. Studies indicate there are many neuron classes, each optimized slightly differently for different purposes (Matani 2018; Yuste 2015). Within the brain, there could be thousands of neuron types. Exploring neuron categories is a logical next step to bound the number of neuron types that need to be explored and learn how to consider neuron function better. This paper generally characterizes neurons, and it shows how they feed into an interneuron network from a top-level perspective. In the brain are sensory neurons, interneurons (processing neurons), interneurons interfacing with the central nervous system (connecting neurons with some processing capability), and motor neurons that control movement. 6. Methylation adaptation and signaling responses DNA methylation is a form of adaption that also results in signaling response changes. Sensitivity or on-off switching of gene function is accomplished. Epigenetic memory keeps this setting until some other influence changes it. Methylation occurs by attaching a phosphate group between a cytosine and guanine pair, forming a CpG site. Demethylation occurs through the removal of the phosphate group from this site. Other types of methylation help cellular processes adapt to change. Research has shown that epigenetic modifications result in stable patterns in the brain that can be transcribed into new cells. Methylation can occur in specific brain parts in response to various environmental and other inputs (Rutishauser 2021). Methylation modification is not the type of memory that makes sense to consider as part of a neural network Von Neumann computing architecture. Instead, methylation modifications would equate to a more global type of change rather than a frequently changing memory cell in a neuron. To give an analogy, methylation would be like updating a computer’s Basic Input-Output System (BIOS) or reprogramming a more straightforward Internet of Things (IoT) device program by reimaging the Programmable Read-Only Memory (PROM) memory. Thus, methylation changes (like BIOS or PROM changes) will persist and impact the control logic of the biological system operations. 7. Neuron synapse signaling Several steps occur when information is sent from one neuron to another. The axon prepares information for proper transmission to the synapses. Then the neurotransmitter takes this processed information and makes it ready to leave the synapse and transition the physical bond from the transmitting neuron to the receiving neuron. Next, the receiving neuron takes in the information, and the transmission is complete. Fig. 19 shows this neuron synapse signaling process. Since sensory input is a comprehensive function done with a collection of neurons, it is important to characterize the relationships between components that are part of this activity. Fig. 20 shows this sensory neuron activity process flow. 8. At what levels can neuron hardware implementation be done? Various abstraction levels can be utilized to create a biomimicry implementation of brain neural networks. Typically, these do not consider the physical layer since biological circuitry fundamentally differs from semiconductor hardware. With chip electronic circuit size approaching the molecular level, this could be revisited. Molecular biology with its various signaling paths and paradigms should be considered for biomimicry inspiration. A spiking neural network (SNN) is the biological neural network signaling approach. The network activates a subset of connected neurons, optimized to process a scenario based on previous learning when necessary to complete a computational task. Neurons are linked through learning mechanisms. Once learning activities are completed, the spiking neural network remembers the proper computational function and response, so it can recall it again when the same scenario occurs. Various learned connections in the brain’s spiking neural network computing fabric can be recalled when needed. 9. Human brain neuron models Research has been done to comprehend how the human brain operates. The brain contains nearly 100 billion neurons, each with the same flexible configuration. Not only is there a vast number of neurons, but there are 100 trillion synaptic interconnections. Neurons have input connections from sensors, a cell body that receives ACT Synapse Model Chemical Synapse Neuron Cell 1 Neuron Cell 2 Presynaptic Axon Synaptic Vesicles Neurotransmitter Postsynaptic Receptors Postsynaptic Cell Synaptic Cleft Figure 19. Synaptic communication activity model. JOHANSEN Human brain function and the creation model 2023 ICC 304
RkJQdWJsaXNoZXIy MTM4ODY=