IISc researchers use GPUs to explore human brain activity in recent study

A new Graphics Processing Unit (GPU)-based machine learning algorithm developed by researchers at the Indian Institute of Science (IISc) can help scientists better understand and predict the connection between different regions of the brain.

The algorithm, called Linear Evaluation, Acceleration, Linear Distributed, or ReAl-LiFE, can quickly analyze the huge amount of data generated from Diffusion Magnetic Resonance Imaging scans ( dMRI) of humans Brain.

Using ReAL-LiFE, the team was able to evaluate dMRI data 150 times faster than existing modern algorithms, according to IISc Press Release released on Monday.

Devarajan Sridharan, Associate Professor at the Center for Neuroscience (CNS), IISc, and corresponding author of research published in a magazine Natural computational science.

Millions neuron fires in the brain every second, generating electrical impulses that travel through neural networks from one point to another in the brain via connecting cables or “axons”. These connections are essential for the calculations the brain does.

“Understanding brain connectivity is crucial for exploring relationships between brain and behavior on a large scale,” said Varsha Sreenivasan, PhD student at CNS and first author of the study. However, conventional approaches to studying brain connectivity often use animal models and are invasive. On the other hand, dMRI scanning provides a non-invasive method to study brain connectivity in humans.

The cables (axons) that connect different regions of the brain are its information highways. Because the axon bundles have a tube-like shape, the water molecules move along their length in a directional direction. dMRI allows scientists to track this movement, in order to create a comprehensive map of the network of fibers in the brain, known as the interconnection network.

Unfortunately, it’s not easy to pinpoint these connections. The data obtained from the scans provide only the net flow of water molecules at each point in the brain, the release notes.

“Imagine that water molecules are cars. The information obtained is the direction and speed of the vehicles at each point in space and time without road information. Our task is similar to inferring road networks by observing these traffic patterns,” explains Sridharan.

To accurately identify these networks, conventional algorithms closely match the predicted dMRI signal from the inferred connection with the observed dMRI signal.

Scientists have previously developed an algorithm called LiFE (Linear Evaluation) to perform this optimization, but one of its challenges is that it works on Central processing unit (CPU)This makes the calculation time consuming.

In the new study, Sridharan’s team tweaked their algorithm to cut down the computational effort involved in a number of ways, including eliminating redundant connections, thus significantly improving the performance of LiFEs.

To speed up the algorithm even more, the team has also redesigned it to work on electronic chips – high-class goods gaming computer – Is called Graphics processing unit (GPU)helping them analyze data at a speed 100-150 times faster than previous approaches.

This innovative algorithm, ReAl-LiFE, can also predict how a human test subject will behave or perform a particular task.

In other words, using the connection strengths estimated by the algorithm for each individual, the team was able to account for variations in behavioral and cognitive test scores within a group of 200 participants.

Such analysis could also have medical applications. “Large-scale data processing is becoming increasingly essential for big data neuroscience applications, especially for understanding healthy brain function and brain pathology,” said Sreenivasan.

Using the resulting connections, for example, the team hopes to be able to identify early signs of aging or decline in brain function before they manifest in behavior in Alzheimer’s patients.

“In another study, we found that an earlier version of ReAL-LiFE could outperform other competing algorithms to distinguish Alzheimer’s patients from healthy controls,” said Sridharan. .

He added that their GPU-based implementation is very generic and can be used to solve optimization problems in many other areas.

Source link


News7h: Update the world's latest breaking news online of the day, breaking news, politics, society today, international mainstream news .Updated news 24/7: Entertainment, the World everyday world. Hot news, images, video clips that are updated quickly and reliably

Related Articles

Back to top button