Explore machine learning research on the study of brain adaptability to the environment
Synapses are connections that allow nerve cells deep in the brain to communicate with each other. Researchers at Johns Hopkins University artificial intelligence and ML Visualize and track changes in synaptic strength in real animals. The method, detailed in Nature Methods, could help researchers better understand how learning, aging, injury and disease affect these connections. I have. human brain
Dr. Dwight Bergles likens learning how an orchestra plays to observing individual performers over time. He holds the positions of Prof. Diana Silvestre and Prof. Charles Homsy in the Solomon H. Snyder Faculty. neuroscience At the Johns Hopkins University (JHU) School of Medicine. According to him, synapses in living animal brains are handled in this peculiar way.
Co-authors on the study include Dr. Richard Heughanier, Bloomberg Distinguished Professor and Solomon H. Snyder Chair of Neuroscience at Johns Hopkins University, and Dr. Adam Charles, Jeremias, Maine. Dr. Sulam, all assistant professors in the Department of Biomedical Engineering. All four researchers are affiliated with the Kavli Neuroscience Discovery Institute at Johns Hopkins University.
Synaptic junctions, or “junctions,” are where nerve cells exchange chemical messengers to transmit information. According to the authors, gaining new abilities and exposure to new environments can increase how different life events change synapses in the brain, increasing connections that enable learning and memory. These are just two examples of how it can weaken or weaken. As difficult as it is, understanding how these small changes occur across the trillions of synapses in the brain is critical to how the brain functions normally and how disease affects the brain. Essential to understanding what influences.
Synapses in the brain are dense and their small size, a property that makes them very difficult to visualize even with new state-of-the-art microscopy, has led scientists to understand the changing chemistry of synaptic messaging. We need to find better ways to visualize.
From difficult, blurry, and noisy image data, Charles says, you need to extract the portion of the signal that you want to see.
To achieve this, Bergles, Sulam, Charles, Huganir, and their colleagues machine learning, is a computing foundation that allows flexible development of autonomous data processing tools. Because machine learning is used effectively in many areas of biomedical imaging, scientists have used machine learning to improve the quality of images of thousands of synapses. The system first needs to be “trained” to tell the algorithm what a high-quality picture of the synapse should look like, which is a key tool for automatic detection that beats human speed.
The researchers used genetically modified mice whose synaptic chemosensors, glutamate receptors, to fluoresce green when exposed to light. In these mice, each receptor emits the same amount of light, so the amount of fluorescence a synapse produces is a measure of its intensity. This is because each receptor produces the same amount of light.
Because individual clusters of synaptic glutamate receptors have been difficult to clearly observe, much less to identify individually and track over time, imaging of the intact brain is of predictably low quality. image was obtained. Researchers used photographs of brain slices taken in vitro from genetically modified animals of the same species to train a machine learning system to convert them into higher quality photographs. Since these pictures were not taken from live animals, it is possible to produce images of similar low quality to those taken from live animals, or significantly higher quality images using different microscopy approaches. was.
