Scientists have always been curious about the human brain. How it works, how it senses, all these are a great mystery. Sensing smells is an interesting ability of our brain. So, how does our brain manage to do it? To find this answer, Recently scientists are researching worm’s brains to predict smell with machine learning.
Researching a human’s brain is very difficult. It’s big and contains 100 billion-plus neurons. That’s why getting a specific result on such a sensitive ability is close to impossible. For this reason, the researchers needed an alternative to finding out from where the core smelling sensation begins.
Why use the worm’s brain to predict smell with machine learning?
A microscopic worm’s brain contains only 302 neurons. That makes this hard task kind of simpler for the researchers. A Salk Associate Professor named Sreekanth Chalasani who is a member of Molecular Neurobiology Laboratory revealed something interesting.
He said that they found something unexpected when they started looking at the effect of these sensory stimuli on individual cells and connections within the worm’s brain. He wanted to study the brain of the worm to know how it processes information from the outer world.
It was so tuff to execute. Chalasani’s team built C.elegans to implant fluorescent sensors in each of their 302 neurons. These sensors would light up as an activity indicator. Then under a microscope, the worms were exposed to 5 different chemicals. Each chemical had different smells.
We are including the chemicals name with referenced smells:
● Benzaldehyde smells like almond,
● Diacetyl smells like buttered popcorn,
● Isoamyl alcohol smells like banana,
● 2-Nonanone smells like cheese, and
● Sodium chloride is the smell of salt.
Was this research successful?
It was actually very hard to differentiate the effects of different smells. So they started analyzing with a mathematical approach which is called graph theory analysis. At last, they successfully paired the new approach with machine learning to differentiate more accurately.
Surprisingly the new algorithm could differentiate the response of neurons in the case of almond and salt smell. But the result of the three others was confusing. It was just an initial step. The researchers wish to study deeply in this area. It can make a revolutionary change in our current artificial intelligence.
If the study continues with this great aim, we will see some unimaginable technologies in the future. We have seen different gadgets that can identify chemicals, radiation, etc. But when a machine will learn to sense the smell, some of our most complicated tasks will be easier.
I think it is possible. We just have to wait patiently to see if we can really use a worm’s brains to predict smell with machine learning. This research will also help us a lot to know how a human’s brain reacts and recognizes different smells. This whole research and related information are published in PLOS Computational Biology.