Researchers at the Indian Institute of Science (IISc.) Have developed a new graphical processing unit (GPU) -based machine learning algorithm called Regularized, Accelerated, Linear Fascicle Evaluation (ReAl – LiFE), which will help to obtain a better understanding and prediction. of connectivity between different regions of the human brain.
This algorithm can help analyze extensive data generated from diffusion magnetic resonance imaging (dMRI) scans that help scientists study brain connectivity at a speed 150 times faster than a normal desktop computer or existing state-of-the-art algorithms. . The study has been published in the journal Nature Computational Science.
With the study, the researchers tried to study the wiring of different parts of the brain that helps perform various calculations.
While these patterns can be studied in animals using invasive techniques, in humans, dMRI is used to infer white matter patterns. Through it, scientists can track the movement of molecules to create a complete map of the connectoma, which is a network of fibers through the brain.
“Although it is difficult to identify connectomes, we are trying to infer the information highway network from looking at traffic flow patterns (if molecules are like cars). We observe the movement of water molecules in the brain and “We try to infer where the wires are. Water molecules have to travel along the wires (axons), which have connected various parts of the brain. By measuring these lengths of water molecules, we can infer which areas are connected.” explain Devarajan Sridharan, associate professor at the Center for Neuroscience (CNS), IISc., and corresponding author of the study.
In addition, he added that this technique requires a large amount of computation that can be performed efficiently using GPUs.
“Tasks that used to take hours or days can be completed in seconds or minutes,” he added.
Accurate identification of information networks, conventional algorithms matched the predicted dMRI signals of the inferred connectoma with the observed dMRI signal.
Earlier a similar algorithm called LiFE (Linear Fascicle Evaluation) was developed to perform the optimization, but since it worked with traditional CPUs, the calculation took a long time.
“In the new study, Mr. Sridharan’s team adjusted its algorithm to reduce the computational effort involved in several ways, including eliminating redundant connections, thus significantly improving LiFE performance. To further accelerate the algorithm, the team also redesigned it to work with specialized electronic chips, of the type found in high-end gaming computers, called graphics processing units (GPUs), which helped them analyze data at speeds of 100 to 150 times faster than previous approaches, ”said an IISc press release.
This algorithm will have several applications in the field of health, including disease diagnosis and behavioral studies.
“Understanding brain connectivity is critical to discovering the relationships between the brain and behavior at scale,” said Varsha Sreenivasan, a doctoral student at CNS and lead author of the study.
While certain patterns of brain connectivity may explain the interindividual differences in care test scores that help determine behaviors, an earlier version of the same algorithm may also help distinguish between Alzheimer’s patients and healthy-age controls. just by measuring brain connectivity.