Researchers mount the brains of worms to predict odors with machine learning

scientists use worms for better understanding the human brain: In particular, how it processes the sense of smell. So far, they have accomplished a rather impressive task: they are able to recognize what the worm smelled a few seconds ago just by looking at its brain.

“We discovered some unexpected things when we started looking at the effect of these sensory stimuli on individual cells and connections in the worm’s brain,” said Associate Professor Salk. Srikant Chalasani, a member of the Laboratory of Molecular Neurobiology and a senior author of the new work.

Chalasani wanted to study how the brain processes information from the outside world at the cellular level. However, this was very difficult for the human brain to do, as it has 86 billion cells. IN microscopic worm,, Caenorhabditis elegans, there are only 302 neurons, making it easy to observe.

Chalasani and his team designed C. elegans to give each of their 302 neurons a fluorescent sensor that will light up when the neuron is activated and will monitor the worms under a microscope while they have been exposed to five chemicals: benzaldehyde, diacetyl, isoamyl alcohol, 2-nonanone and sodium chloride.

These five chemicals smell like almonds, popcorn with butter, banana, cheese and salt to humans.

But researchers have had difficulty distinguishing the effects of different odors. Then they turned to the mathematical approach. The approach, called graph theory, analyzes the collective interactions between pairs of cells.

Finally, they combined their new approach with machine learning to be able to distinguish even more discrete interactions. They found that this new algorithm was able to clearly distinguish the nervous response to salt and benzaldehyde, but often confused the other three chemicals.

Researchers now hope to use the lessons of this study to gain a deeper understanding of how people encode information in the brain and what goes wrong in sensory processing disorders such as attention deficit hyperactivity disorder (ADHD) and autism. The study was published in the journal PLOS computational biology.

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button