[ad_1]
A neuron (also called a nerve cell) is an electrically sensitive cell that captures, forms and transmits data through electrical signals and concoction. It is one of the essential components of the nervous system.
The dendritic tree of neurons badumes an important role in the treatment of information in the brain. Although it is thought that dendrites require independent subunits to perform the most substantial part of their computations, we still do not see how they compartmentalize into functional subunits.
In a new study by scientists from the EPFL Blue Brain project, a Swiss brain research initiative, it is shown how these subunits can be deduced from the properties of dendrites. They have developed a new framework for determining how a neuron works in the brain.
The survey was conducted using cells from the virtual rodent cortex of Blue Brain. Scientists predict that different kinds of neurons – non-cortical or human – will function in the same way.
Their results demonstrate that when a neuron gets an input, the specific tree-like receptor parts, called dendrites, that come out of the neuron, cooperate in virtually a way that is adjusted according to the complexity of the neuron. # 39; entry.
The strength of a synapse decides with what emphasis a neuron feels an electrical sign coming from different neurons, and the demonstration of learning alters that force.
By examining the connectivity matrix that determines how these synapses connect, the algorithm defines when and where synapses are grouped into autonomous learning units from the structural and electrical properties of the synapses. dendrites.
The new algorithm decides how dendrites of neurons functionally separate into separate recording units and discover that they interact dynamically, depending on the remaining task, to process the data.
Scientists are comparing their results to the work of understanding the innovation actually realized today. This recently observed dendritic utility acts as parallel computing units, implying that a neuron can treat various parts of the parallel contribution, such as supercomputers.
Each of the parallel computing units can autonomously determine how to change its output, much like the deep learning system nodes used in artificial intelligence models today. Compared to cloud computing, a neuron separates powerfully from the number of separate computing units required by the workload of the input.
Marc-Oliver Gewaltig, Head of Neuroscience Simulation at Blue Brain, said, "In the Blue Brain project, this mathematical approach allows for the determination of functionally relevant neural input clusters that feed the same unit. parallel processing. This then allows us to determine the level of complexity at which to model cortical networks during digital brain reconstruction and simulation. "
The study also reveals how these parallel processing units influence learning, that is, the change in connection strength between different neurons. The way a neuron learns depends on the number and location of parallel processors, which therefore rely on the sign that lands from different neurons. For example, specific synapses that do not adapt autonomously when the level of neuron information is low, begin to adapt freely when information levels are higher.
Willem Wybo, Senior Scientist and Lead Author, said, "The method shows that in many brain states, neurons have far fewer parallel processors than predicted by dendritic branch patterns. Thus, many synapses appear to be in "gray areas" where they do not belong to any treatment unit. However, in the brain, neurons receive different levels of background. Our results show that the number of parallel processors varies with the level of background, indicating that the same neuron could have different computer roles in different brain states. "
"We are particularly excited about this observation as it sheds new light on the role of up / down states in the brain and also gives a reason why cortical inhibition is so site-specific. With this new information, we can begin to look for algorithms that exploit the rapid changes in the coupling between treatment units, allowing us to better understand the fundamental question of calculating the brain. "
The study is published in the journal Cell Reports.
[ad_2]
Source link
