Mouse Brain Cells Captured in Action As a result of AI

Mouse Brain Cells Captured in Action As a result of AI ...

In collaboration with specially designed ultra-small microscopes, Johns Hopkins biomedical engineers have developed an artificial intelligence training strategy to capture mouse brain cells in action. According to the researchers, the AI system, in combination with advanced ultrasound sensors, makes it possible to pinpoint precisely where and when cells are activated during movement, learning, and memory. This data could someday be used to understand how the brain functions and is affected by disease.

On March 22, researchers performed experiments in mice.

According to Zhang Zhang, a professor of biomedical engineering at the Johns Hopkins University School of Medicine, a mouse''s head may not truly be capable of sustaining its neurological function. Moreover, we must evaluate precisely what is happening among individual brain cells and their connections, while the animal is liberating, eating and socializing.

The size of these miniature microscopes, measured in a couple of millimeter in diameter, would inhibit the ability to process motion interference. Scientists estimate that a Lis miniature microscope would need to exceed 20 frames per second to prevent motion disturbances.

Li believes there are two ways to increase the frame rate. You may increase the scanning speed and decrease the number of points scanned.

Lis'' engineering team quickly discovered they reached six frames per second in the physical limits of the scanner, which maintained excellent image quality but was far below the required rate. So, the team decided to adopt a second strategy to decrease the number of points scanned. This strategy, however, would result in the microscope''s ability to recover lower resolution data.

When it is impossible or lengthy to assemble a computer program for an assignment, Li suggests that an AI program may be developed to recognize and restore missing points, thus improving the resolution of the images. Such AI training protocols are often used when it is difficult or tedious to perform such tasks as reliably recognizing a cluster of features as a human face. Instead, computer scientists utilize the technique of allowing computers to learn to program themselves through processing large sets of data.

The lack of similar mouse brain imagery to train the AI to overcome this obstacle, according to the researchers. This step began by training the AI to detect brain cells with natural structural variation and a minor bit of motion caused by the mouse''s breathing and heartbeat.

According to Li, whenever we collect data from a moving mouse, it would still be similar enough for the AI network to recognize.

The researchers compared the AI program to see if it might enhance mouse brain images with an increased frame rate. Using a reference image, the researchers measured how accurately the AI might improve the image and restore the image resolution.

The researchers said the AI might properly restore the image quality up to 26 frames per second.

The researchers analyzed how well the AI tool worked in combination with a mini microscope attached to the head of a moving mouse. Using the AI and microscope, they spotted exact activity spikes of individual brain cells activated by the mouse walking, rotating, and generally exploring its environment.

According to Li, this development might help gather more information on how the brain is dynamically connected to action on a cell level.

Researchers claim that with more training, the AI program may be able to accurately interpret images up to 52 or even 104 frames per second.

You may also like: