Scientists Develop Innovative AI Method for Tracking Neurons
Scientists from EPFL and Harvard have developed an innovative artificial intelligence method for tracking neurons in moving organisms, which can significantly accelerate brain research. According to research published in the journal Nature Methods under the leadership of Sakhand Jamal Raha from the school of basic sciences at EPFL, this groundbreaking method has the potential to revolutionize the field of neuroscience.
Traditionally, analyzing neuron activity required manual marking of a large number of images, which complicated research, particularly when studying brains that change and deform, such as in worms. The new method employs convincing neural networks (CNN), an artificial intelligence system trained to recognize samples in images, to address this challenge.
The unique feature of this method is the use of “targeted augmentation” technology, which automatically generates reliable annotations based on a limited number of manual segments. This enables the CNN to effectively study internal brain deformations and create annotations for new poses, significantly reducing the need for manual markings.
In a successful test on the round Caenorhabditis Elegans, a popular model organism in neuroscience, scientists were able to measure the activity of the worm’s interneurons and identify complex changes in their behavior in response to various stimuli, such as smells.
Lead researcher Sakhand Jamal Raha highlights that this new method greatly reduces the amount of manual work required for analysis compared to completely manual markings. This breakthrough has the potential to accelerate research in brain visualization and deepen our understanding of neural patterns and behavior.