I do this lab with my advanced biology class called “The Mystery Milk” lab. I give them 4 unknown liquids and they have to use a series of indicators to figure out what these milk-like products are. The students always ask me if they can taste the products (Answer: No, it’s a science lab). But imagine if there WAS a device that acted like a tongue for mystery liquids…
Beer, Wine, Milk
Scientists have used a combination of techniques to create electronic tongues (and noses). The tongue was able to identify wine from beer from milk. In fact, it was so sensitive that it even picked up souring milk changes! This was something the scientists were *not* expecting.
Using surface plasmon resonance imaging, cross-reactive sensor arrays, and machine learning, researchers trained these tongues to identify distinctive characteristics of test liquids through pattern recognition. Before I go into why we even care about electronic tongues, let’s take a look at the technology behind electronic tongues.
Vibrating electron frequencies
Surface plasmon resonance imaging uses the energy from electrons to figure out what molecules are in a substance. Since different molecules resonate at different frequencies, the resulting images are also different.
Cross-reactive Sensor Arrays
Cross-reactive sensor arrays are like a really really really generalized ELISA test. Instead of having one receptor binding to one molecule, it recognizes general biochemical similarities. This means one test can sort through thousands of molecules.
When you combine the data from both tests, machine learning algorithms can be trained to pick out patterns. Remember that post about neural networks and deep learning? You can use this type of data to train a machine!
How can we use electronic tongues in real life? Imagine a polluted river. You could collect samples and analyze them to figure out what the pollutants were. Or you could stick your electronic tongue in the river and have it collect data immediately.