The primary objective of the research team at RIKEN Center for Integrative Genomics Research () was to use single-cell transcriptome differences and statistics from 16 individuals with multiple sclerosis to evaluate the common neuronal networks that are affected by MS. They did this with a newly developed system for analysis that has seen sustained use in recent years. The system uses single cells RNA-sequencing and other approaches to work out how the two main starting-point cells work together to perform any given function. Furthermore the systems is automated in operation. In their paper published in the journal Science Translational Medicine the group discusses the latest advances in this field and how they used single-cell networks to define the what and how of neuronal networks.
Prior to the development of artificially intelligent systems that can assemble the data sets of billions of nerve cells simultaneously many systems are first exploited as experimental systems. In this way the differences between the input and output processes by the neurons can be used to explain and predict the way that they perform to an arbitrary system. For the current artificial intelligence systems this is possible because synthetic networks are associated with a number of the main procedures everything from performing typing and handwriting to drawing and sorting. However systems cannot be used to define what exactly are function and non-function and how they can be used to generate useful information for licensed health professionals the National Institute for Physiological Sciences who have to constantly take account of the interaction between different units of the nervous system as well as diagnosing diagnosing and treating diseases.
For the new functional classification of neuronal networks the system developed by the research group uses the classification architecture of the RIKEN biological cell maintenance plan an integrated human body system that is one of the most sophisticated in existence. Those with MS have a 30-fold higher incidence of spinal cord injuries than the general population and about 50 of those affected die from the disease. The researchers chose this classification for the new study because they were worried that differences in these network characteristics might be over-represented in the study groups. Additionally differences in different neural networks might influence whether an artificial intelligence system using the biological systems could predict the outcome of individual patients.
To their surprise the researchers found specific network differences between the known type 1 and type 2 MS population and also found that system differences between the starting-point cells and the immune cells were related to the outcome of neurosurgery patients with MS. Initial indications in the literature suggested that synapse is not entirely maintained in MS and it is largely destroyed by trauma and imaging. In addition such a disruption might cause MS but not cause new symptoms (that most severe forms of the disease are avoided by knowing the disease process). Their LINK project of the brain tissue cell mineralization which has been studied in numerous types of disease was open-sourced to the outside world suggesting a potential cause-and-effect.
The group was able to conclude that they were able to manually determine the course results of patients with MS using the form-and-function state-ten memory test and the age-adjusted form-and-function test based on blood biomarkers in their peripheral circulation and other extremities. This result was subsequently confirmed in a controlled experiment in the patients which we showed was as reproducible as the test performed in the real patients suggesting that their results are as accurate explains Satoshi Sato the senior author of the study and head of the Center for Integrative Genomics Research. The main authors are Yoshitake Murakami (RIKENColumbia University) RIKENFukuoka University and RIKENShimizuoka University.