Researchers use machine learning to understand how brain cells work

For something so small, neurons can be quite complex – not only because there are billions of them in a brain, but because their function can be influenced by many factors, such as their shape and genetic make-up. .

A research team led by Daifeng Wang, professor of biostatistics, medical informatics, and computer science at the Waisman Center at the University of Wisconsin-Madison, is adapting machine learning and artificial intelligence techniques to better understand how a variety of traits together affect how neurons function and behave.

Called multiple learning, this approach can help researchers better understand and even predict brain disorders by examining specific neural properties. The Wang lab recently published its findings in two studies.

Daifeng Wang

In the first study, published in November 2021 in the journal Communications Biology, the researchers showed that they could apply multiple learning to predict the characteristics of neurons. By applying existing machine learning techniques, which use computer algorithms to analyze large amounts of data and automatically make predictions, they found they could classify cells based on their genes and electrophysiological behavior. This behavior encompasses the electrical activity of neurons, which is crucial for communication between neurons and, ultimately, brain function.

Using information from around 3,000 neurons in the mouse brain, the scientists applied multiple learning to align gene expression and electrophysiological data. Their goal was to establish whether there was a measurable relationship between the two.

They found that these two characteristics of neuronal cells showed similar patterns – high values ​​in the same group of cells, but low values ​​in the rest of the cells – and were aligned in “multidimensional space” or demonstrated a relationship. with each other. . This defines their so-called multiple form, a complex mathematical description of the properties of neurons.

“Based on this multiple shape, we found that cells can be grouped into different groups,” says Wang, also a professor of biostatistics and medical informatics at the UW School of Medicine and Public Health.

Clustering cells using a single feature, either gene expression alone or electrophysiology alone, did not result in clusters as clearly separated as when the two features were used in tandem.

The scientists then asked how the genes might work together to influence cellular electrophysiology. Using clusters of cells, they found links between electrophysiological characteristics and specific genes that control the expression of other genes. Some of these genes are also involved in the control of the immune system, suggesting an interaction between neuronal communication and inflammation.

With this data, Wang and his students then explored whether they could make predictions about the electrophysiological characteristics of a neuron based on gene expression. Wang likened this to trying to predict the relationship between traffic patterns in a particular part of a city and the number of takeout orders from restaurants in the area at any given time of day.

Multiple learning can help researchers better understand and even predict brain disorders by examining specific neural properties.

“If you compare traffic with the number of takeout orders from restaurants in a particular area, they are two different things, but I think they share similar patterns – as they can both have the same peak times” , he says. “Here, we would use multiple alignment to align patterns between traffic (electrophysiology) and takeout amount (gene expression), and then find the shared pattern between the two.”

With this information, Wang says, you can start predicting when takeout orders will peak based on traffic data alone, or you can start predicting the gene expression of neurons based on their electrophysiological characteristics.

Once the concept was developed, Wang’s team then used the collected data to inform their second study, published in January in Nature Computational Science. It describes a new and improved type of multiple learning that addresses the limitations of previous models and could help researchers better understand neuronal function in the context of health and disease.

Called deepManReg, the new model improves the prediction of neuronal characteristics based on gene expression and electrophysiology. It is also more generalizable to other types of cellular data, can integrate more than two types of neural features, and can reveal how multiple features connect or influence each other.

Using machine learning for these applications could help reduce the time and money needed to study certain brain features. Although the researchers’ most recent studies are based on healthy cells, Wang intends to use the techniques to learn more about brain disorders and diseases.

“Basically, (we can study) how these genes are regulated to affect the electrophysiology or behaviors of diseased cells,” says Wang.

Both studies were supported by the National Institutes of Health (grants R01AG067025, R21CA237955, R03NS123969, U01MH116492, and P50HD105353.) The studies also received financial support from the Office of the Vice Chancellor for Research and Higher Education.

Sherry J. Basler