Fruity fly study uncovers neural circuits for sensing the pleasantness or unpleasantness of odors

Researchers led by Hokto Kazama at the RIKEN Center for Brain Science (CBS) in Japan have discovered how animals sense whether things smell pleasant or unpleasant, one of the abilities that allow us to appreciate the flavor of foods.

The study shows that these different sensations are computed by separate circuits of neurons in the brain but are not in fact opposites of each other. The findings were published in the journal Cell.

The sense of smell—olfaction—is one of the oldest senses, having its roots in the chemical receptors of ancient aquatic vertebrates. In mammals, chemical molecules in the air enter the nose, eventually encountering chemical receptors on olfactory neurons, which then send signals to the brain.

As there are countless airborne molecules, and because odors are often made from combinations of several molecules at once, a simple system with one receptor for each odor was never an option during evolution.

Instead, odors are encoded by thousands of overlapping neurons, distributed across the brain, which makes understanding how we judge smells very difficult.

In situations like this, scientists often look to the animal kingdom for less complex systems that still work very similarly. Kazama and his team focused their studies on the olfactory system of the fruit fly, where they can identify every single olfactory neuron and its connections. Even so, that’s thousands of neurons and hundreds of thousands of connections and it remained a serious challenge.

To overcome this difficulty, the researchers developed a method of recording the activity of all neurons in each brain region of the fruit fly by combining two-photon microscopy and optogenetic cell labeling. They also built a network model that reproduces the activity of these neurons based on the connectome—the connectivity between all the neurons in the brain—which was useful in understanding the circuitry underlying the computations.

They found that neurons in a brain region called the lateral horn represent the innate hedonic value of odors, pleasant or unpleasant. The model predicted that unpleasant odors were represented by feedforward excitation of neurons across the lateral horn, while pleasant odors were derived from additional local inhibition.

The most unexpected result was that the pleasantness and unpleasantness of odors are computed in circuits that are not only separate from each other, but also distinct in connectivity motifs. This means that in terms of the circuit, “good” is not simply the opposite of “bad”.

The researchers were able to use the optogenetic setup to verify the model’s predictions because optogenetics allows scientists to precisely excite or inhibit individual neurons of their choice. For example, when the model predicted that silencing a certain local circuit would make flies dislike an odor that they normally find pleasant, that is exactly what happened.

Because the olfactory circuit is very similar across animals, this work may contribute to a better understanding of the human brain.

According to Kazama, “by implementing the circuit mechanisms that we discovered in a compact fly brain, we may be able to develop more efficient algorithms as well as high-efficiency brain-inspired AI.”

“Indeed, our development of a connectome-based network model is a step towards creating a digital twin of a brain, which is useful in predicting the output of a system under various contexts.”