summary: Researchers used artificial intelligence (AI) to track and visualize changes in synaptic strength in live animals. Synapses are communication points in the brain and are essential for the processes of learning, memory and aging.
Scientists have employed machine learning to improve image quality and enable long-term detection and tracking of individual synapses. This advance may provide important insights into how the brain is affected by aging, disease and injury.
Important facts:
- The technology uses artificial intelligence to track changes in synaptic strength, the points at which nerve cells in living animals communicate.
- In this study, we implemented machine learning to improve the quality of images composed of thousands of synapses, enabling long-term individual detection and tracking.
- This revolutionary technology helps us understand how synaptic connections in the human brain change with learning, aging, injury and disease.
sauce: johns hopkins medicine
Scientists at Johns Hopkins University have developed a way to use artificial intelligence to visualize and track changes in the strength of synapses (connections between nerve cells in the brain) in living animals.
techniques described in nature methodScientists say it should lead to a better understanding of how such connections in the human brain change with learning, aging, injury and disease.
“If you want to know more about how an orchestra performs, you have to observe individual players over time, and this new method does that for synapses in the brains of living animals.” said Dr. Dwight Bergles of Diana Silvestre University. Professor Charles Homsy of the Solomon H. Snyder Department of Neuroscience, Johns Hopkins University (JHU) School of Medicine.
Dr. Bergles co-authored the study with fellow biomedical engineering assistant professor Dr. Adam Charles, Dr. Adam Charles of Maine, Dr. Jeremias Slam, and JHU Bloomberg Distinguished Professor Richard Hugania. Director of Neuroscience, Solomon H. Snyder. All four researchers are members of the Kavli Neuroscience Discovery Institute at Johns Hopkins University.
Nerve cells transmit information from one cell to another by exchanging chemical messages at synapses (“junctions”). The authors explain that in the brain, various life experiences, such as exposure to new environments or learning a skill, are thought to trigger synaptic changes that strengthen or weaken these connections that enable learning and memory. increase.
Understanding how these subtle changes occur across the trillions of synapses in our brains is a daunting challenge, but it also explains how our brains function in health and disease. important in understanding how it changes with
To determine which synapses change during specific life events, scientists have long visualized the changing chemistry of synaptic messaging, necessitated by the high density of synapses in the brain and their small size. I’ve been looking for a better way to convert. It has a new state-of-the-art microscope.
“From the difficult, blurry, noisy image data, we had to extract the portion of the signal that we wanted to see,” says Charles.
To that end, Bergles, Sulam, Charles, Huganir and their colleagues turned to machine learning, a computational framework that enables flexible development of automated data processing tools.
Machine learning has been successfully applied to many areas of biomedical imaging. In this case, scientists have leveraged this approach to improve the quality of images made up of thousands of synapses.
This can be a powerful tool for automatic detection that greatly exceeds human speed, but the system must first be “trained” to make the algorithm learn what a high-quality image of synapses should look like.
In these experiments, the researchers worked with genetically modified mice whose glutamate receptors (synaptic chemosensors) glow green (fluoresce) when exposed to light.
Since each receptor emits the same amount of light, the amount of fluorescence produced by synapses in these mice indicates the number of synapses, and therefore the strength of synapses.
As expected, imaging of the intact brain produced low-quality images, making it difficult to clearly see individual clusters of glutamate receptors at synapses, much less to detect and track them individually. was difficult.
To convert these into higher quality images, the scientists trained machine learning algorithms using images of brain slices (ex vivo) taken from the same strain of genetically modified mice.
Since these images were not from live animals, it is impossible to use a different microscopy technique to produce much higher quality images, nor to produce images of lower quality similar to those taken in live animals of the same view. I was also able to generate an image.
This cross-modality data collection framework enabled the team to develop enhanced algorithms capable of generating high-resolution images from low-quality images, similar to those collected from live mice.
In this way, data collected from intact brains are greatly enhanced, allowing individual synapses (thousands) to be detected and tracked during experiments over several days.
To follow changes in receptors in living mice over time, the researchers then used a microscope to repeatedly image the same synapses in mice over several weeks. After taking baseline images, the team placed the animals in a room with novel visual, odor, and tactile stimuli for five minutes.
They then imaged the same areas of the brain every other day to see if and how the new stimuli affected the number of synaptic glutamate receptors.
Although the focus of their research was to develop a suite of methods for analyzing changes at the synaptic level in a variety of contexts, the researchers found that this simple environmental change caused changes in the fluorescence spectrum across synapses in the cerebral cortex. We found that it induced changes, indicating synaptic-synaptic relationships. Animals exposed to new environments tended to have stronger muscles, with some gaining strength and others losing strength.
The research has been made possible by close collaboration between scientists with different expertise, from molecular biology to artificial intelligence, who usually don’t work closely together. But such collaborations are encouraged at the multidisciplinary Kavli Neuroscience Discovery Institute, Bergles says.
Researchers are now using this machine learning approach to study synaptic alterations in animal models of Alzheimer’s disease, and the method may shed new light on synaptic alterations that occur in other disease and injury contexts. i think i can make it.
“We’re really looking forward to seeing where and how others in the scientific community take this,” Slam said.
Funding: The experiments in this study were performed by Yu Kang Xu (PhD student and JHU Kavli Neuroscience Discovery Institute Fellow) and Dr. Austin Graves. (JHU Biomedical Engineering Assistant Professor) and Gabrielle Coste (JHU Neuroscience PhD Student). This study was funded by the National Institutes of Health (RO1 RF1MH121539).
About this AI and neuroscience research news
author: Vanessa Wasta
sauce: johns hopkins medicine
contact: Vanessa Wasta – Johns Hopkins University School of Medicine
image: Image credited to Neuroscience News
Original research: open access.
“Cross-Modality Supervised Image Restoration Enables Nanoscale Tracking of Synaptic Plasticity in Living Mice” by Dwight Bergles et al. nature method
overview
Cross-modality supervised image restoration enables nanoscale tracking of synaptic plasticity in living mice
Learning is thought to involve alterations in synaptic glutamate receptors, submicron structures that mediate communication between neurons in the central nervous system.
The small size and high density of synapses make them difficult to degrade in vivo, limiting our ability to directly link receptor dynamics to animal behavior. Here, we have developed a method that combines computational and biological methods to overcome these challenges.
First, we trained a deep learning image restoration algorithm that combines the advantages of in vitro super-resolution and in vivo imaging modalities to overcome the limitations inherent in each optical system.
Applied to in vivo images of transgenic mice expressing fluorescently labeled glutamate receptors, this repair algorithm super-resolves synapses and allows us to track behavior-related synaptic plasticity with high spatial resolution.
This method demonstrates the ability of image enhancements to learn from ex vivo data and imaging techniques to improve in vivo imaging resolution.
