A new AI system makes it possible to find precisely where and when mouse brain cells are activated during movement, learning and memory.

The data gathered from these experiments, conducted at Johns Hopkins University School of Medicine, could someday allow scientists to understand how the human brain functions and is affected by diseases. An article published in Nature Communications explains how the researchers were able to develop this new algorithm.

“When a mouse’s head is restrained for imaging, its brain activity may not truly represent its neurological function,” said Dr Xingde Li, leader of the research group. “To map brain circuits that control daily functions in mammals, we need to see precisely what is happening among individual brain cells and their connections, while the animal is freely moving around, eating and socialising.”

Li’s team did not initially set out to use AI in its research. It started by developing ultra-small microscopes that the mice wear on their heads. However, the researchers soon ran into a significant challenge: due to the small size of the microscopes – measuring only a couple of millimetres in diameter – the imaging technology they can carry is very limited. Disturbances such as the mouse’s breathing or heart rate would affect the accuracy of the data these microscopes can capture.

To eliminate the error caused by these disturbances, the researchers estimated that the miniature microscope would need to exceed 20 frames per second.

“There are two ways to increase frame rate,” explained Li. “You can increase the scanning speed, and you can decrease the number of points scanned.”

Out of the two options, the latter proved to be successful. However, similar to reducing the number of pixels in an image, this strategy would cause the microscope to capture lower-resolution data. Enter AI. 

Li hypothesised that an AI program could be trained to recognise and restore the missing points, enhancing the images to a higher resolution. 

One significant challenge in the proposed AI approach was the lack of similar images of mouse brains to train the AI against. To overcome this gap, the team developed a two-stage training strategy. The researchers began training the AI to identify the building blocks of the brain from images of fixed samples of mouse brain tissue, until it was capable of recognising these building blocks in a head-restrained living mouse under their ultra-small microscope.

“The hope was that whenever we collect data from a moving mouse, it will still be similar enough for the AI network to recognise,” said Li.

The researchers then tested the AI program to see if it could accurately enhance mouse brain images by incrementally increasing the frame rate. Using a reference image, the scientists reduced the microscope scanning points by factors of 2, 4, 8, 16 and 32 and found that the AI could adequately restore the image quality up to 26 frames per second. 

By combining the AI and tiny microscope attached to the head of a moving mouse, Li’s team was able to precisely see activity spikes of individual brain cells activated by the mouse walking, rotating, and generally exploring its environment.

“We could never have seen this information at such high resolution and frame rate before,” said Li. “This development could make it possible to gather more information on how the brain is dynamically connected to action on a cellular level.”

AI tools have been very successful in driving scientific research, having been recently used to assist early diagnosis of Alzheimer’s disease, identify symptoms of Covid-19, and remediate some everyday mental health challenges. 

The researchers said that, with additional training, the algorithm may eventually be able to accurately interpret images up to 52 or even 104 frames per second.

A new AI system makes it possible to find precisely where and when mouse brain cells are activated during movement, learning and memory.

The data gathered from these experiments, conducted at Johns Hopkins University School of Medicine, could someday allow scientists to understand how the human brain functions and is affected by diseases. An article published in Nature Communications explains how the researchers were able to develop this new algorithm.

“When a mouse’s head is restrained for imaging, its brain activity may not truly represent its neurological function,” said Dr Xingde Li, leader of the research group. “To map brain circuits that control daily functions in mammals, we need to see precisely what is happening among individual brain cells and their connections, while the animal is freely moving around, eating and socialising.”

Li’s team did not initially set out to use AI in its research. It started by developing ultra-small microscopes that the mice wear on their heads. However, the researchers soon ran into a significant challenge: due to the small size of the microscopes – measuring only a couple of millimetres in diameter – the imaging technology they can carry is very limited. Disturbances such as the mouse’s breathing or heart rate would affect the accuracy of the data these microscopes can capture.

To eliminate the error caused by these disturbances, the researchers estimated that the miniature microscope would need to exceed 20 frames per second.

“There are two ways to increase frame rate,” explained Li. “You can increase the scanning speed, and you can decrease the number of points scanned.”

Out of the two options, the latter proved to be successful. However, similar to reducing the number of pixels in an image, this strategy would cause the microscope to capture lower-resolution data. Enter AI. 

Li hypothesised that an AI program could be trained to recognise and restore the missing points, enhancing the images to a higher resolution. 

One significant challenge in the proposed AI approach was the lack of similar images of mouse brains to train the AI against. To overcome this gap, the team developed a two-stage training strategy. The researchers began training the AI to identify the building blocks of the brain from images of fixed samples of mouse brain tissue, until it was capable of recognising these building blocks in a head-restrained living mouse under their ultra-small microscope.

“The hope was that whenever we collect data from a moving mouse, it will still be similar enough for the AI network to recognise,” said Li.

The researchers then tested the AI program to see if it could accurately enhance mouse brain images by incrementally increasing the frame rate. Using a reference image, the scientists reduced the microscope scanning points by factors of 2, 4, 8, 16 and 32 and found that the AI could adequately restore the image quality up to 26 frames per second. 

By combining the AI and tiny microscope attached to the head of a moving mouse, Li’s team was able to precisely see activity spikes of individual brain cells activated by the mouse walking, rotating, and generally exploring its environment.

“We could never have seen this information at such high resolution and frame rate before,” said Li. “This development could make it possible to gather more information on how the brain is dynamically connected to action on a cellular level.”

AI tools have been very successful in driving scientific research, having been recently used to assist early diagnosis of Alzheimer’s disease, identify symptoms of Covid-19, and remediate some everyday mental health challenges. 

The researchers said that, with additional training, the algorithm may eventually be able to accurately interpret images up to 52 or even 104 frames per second.