In examining the possible causes and treatments for degenerative conditions such as Alzheimer’s disease or Parkinson’s, neuroscientists often struggle to identify cells needed for activity. understand the brain that causes changes in behavior such as memory decline or imbalance and vibration.
A multidisciplinary team of neuroscience researchers at the Georgia Institute of Technology, borrowing from existing tools such as graphical models, has found a better way to identify cells and understand the ways of diseases, which could lead to better understanding, diagnosis and treatment.
The results of their research on February 24 were reported in the journal eLife. The research was supported by the National Institutes of Health and the National Science Foundation.
The field of neuroscience studies how the nervous system works, and how genes and the environment influence behavior. By using new technologies to understand the natural and intolerant states of biological systems, neuroscientists hope to eventually bring cures to diseases. Before that can happen, neuroscientists must first understand which cells in the brain are driving but mapping brain-to-brain activity is not as simple as it appears.
No two brain cells are alike
Traditionally, scientists have established a coordinate system to map each cell location by comparing images to an atlas, but the idea in literature is not “that all brains look the same a real thing, “said Hang Lu, Love Family Professor of Chemical and Biomolecular Engineering at Georgia Tech School of Chemical and Biomolecular Engineering.
There are two main challenges in a coordinated way: first, the actual number of cells in which none look so different; second, cells vary from person to individual.
This is a big hurdle right now – you can record neuron activity all you want but if you don’t understand which cells are doing what, it is difficult to compare brain or situation and draw meaningful conclusions. “
Hang Lu, Professor, Georgia Tech School of Chemical and Biomolecular Engineering
According to graduate researcher Shivesh Chaudhary, there are also sounds in data that make it difficult to establish communication between two different regions of the brain. “There may be some deformations in data or some shapes may be missing,” he said.
Focusing on cellular relationships, not just geography
To overcome these challenges, the Georgia Tech researchers borrowed from two subjects – graphical models in machine learning and a metric geometric approach to match shape in mathematics – and devised a computational method to identify cells in their organism, the nematode C.elegans.
The team used frameworks from other areas such as natural language processing to build their own modeling software. In natural language processing, the computer can determine what sentences mean by capturing dependence between words in a statement.
The researchers adopted a similar model but instead of capturing dependence among the words, “We captured them among the neurons to identify cells,” Chaudhary said, noting that this approach limits error multiplication compared to other methods that study the geographic location of each cell. .
“The use of relationships among the cells was certainly more useful in explaining cell identity,” Lu said. “If you define one, you affect the identity of the other cells.”
The approach, say the research team, is far more accurate than the standard identification method. The algorithm, while not perfect, performs significantly better against imperfect data, and “gets less rattled” by sound or errors, Lu said.
The algorithm has a major impact on many developmental diseases, because once scientists are able to understand the mechanism of disease, they can detect interventions.
“You can use this to do drug and genetic scans to assess genetic risks. You can take someone’s genetic background and study how this background causes cells to behave. itself in a different way from the usual reference genetic background, “Lu said.
“One interesting thing about this approach is that it is data-driven, so it captures the changes among individual worms. This approach has a high potential to be relevant to a field. a wide range of studies on development and function under normal as well as disease.
Faster data analysis
The algorithm significantly accelerates the speed of overall brain data analysis. The researchers explained that, prior to this progress, their lab may take 20 minutes to record a data set, but it would take them weeks to identify cells and analyze data. With the algorithm, the analysis takes “at most overnight on a desktop,” Chaudhary said.
The device also supports crowdsourcing, collaborative online platforms that open up the algorithm to a larger community, which can test the algorithm and build atlases.
“Every researcher working on the same problem could make recordings and contribute to the construction of these atlases that can be widely used in all contexts,” Lu said.
The researchers believe the success of the project to be able to apply a variety of subjects across physics, biology, mathematics and chemistry. Chaudhary, who has an undergraduate degree in chemical engineering, took advantage of advances in computer science and mathematics to solve this unique neuro-science problem.
“In our labs, we have a physicist working on microscope construction, we have biologists, we have people like me who are more inclined towards computer science. We also collaborate with pure mathematics, “he explained. “The neuro-science field has it all. You can go whichever way you want.”
Source:
Georgia Institute of Technology
Magazine Reference:
Chaudhary, S., et al. (2021) Graphical modeling framework for automated sensing of cell identity in dense cell imaging. eLife. doi.org/10.7554/eLife.60321.