Author: Asher Agarwal, Class of 2027

Figure 1. A brown mouse eating food.

The ability to discriminate similar, partially overlapping sensory stimuli is critical for an animal to survive in its environment. For example, in taste, learning discrimination becomes the difference between consuming a nutritious food item and a toxic one. Previous research done on other senses supports 2 models: one that learning may enhance sensory representations in the brain, and the other that learning enhances decision-making activity. Despite the importance of taste discrimination, taste discrimination learning and its neural basis have been largely understudied. Researchers from the Graduate Program of Neuroscience at Stony Brook University set out to understand the dominant mechanism of this learning in the gustatory cortex (GC), the region of the brain that is responsible for the sense and discrimination of taste.

To analyze GC plasticity that occurs during discrimination learning, they chose six mice to undergo a two-alternative choice task with taste mixtures of sucrose and sodium chloride (NaCl). The mice’s training involved progressively from pure sucrose and NaCl to more overlapping taste mixtures, increasing the difficulty for discrimination. Two-photon calcium imaging was used to record neural activity in the mice before and after learning. Specifically, neural responses were analyzed for sensory signals which occur during the sampling period, when the mice taste the sample, and decision-related signals which occur after a delay period, when the mice make a decision about the sample. 

Before learning, most of the GC neurons encoded linearly, which represents sensory signals via stimulus intensity. Decision-related signals were weak and dispersed during the delay period. For highly similar taste mixtures, performance was close to levels that would occur simply by chance, suggesting the mice were not yet able to discriminate the tastes. After the mice’s learning, behavioral performance improved significantly across all mixture pairs presented to them, particularly for the more difficult ones. Sensory representations (linear encoding) did not significantly improve. However, decision-related signals became more intense and consistent, with stronger calcium peaks closer to the decision. The signals also became more specialized–neurons were tuned to make more categorical choices rather than just reflecting the percentages of NaCl and sucrose.

The finding that sensory representation did not change after learning but decision-related activity did, indicated that taste discrimination learning does not alter how stimuli are perceived, but rather how decisions about them were made. This study highlights how the plasticity of decision-related neural circuits in the GC is the mechanism for taste discrimination learning. Future work should involve high-density electrophysiology for a richer data set and should also look across the various layers of the GC to reveal further insights into the neural mechanism underlying taste discrimination learning.   

Works Cited:

[1] J. Kogan and A. Fontanini, “Learning enhances representations of taste-guided decisions in the mouse gustatory insular cortex” Current Biology 34, 1880-1892 (2024). doi: https://doi.org/10.1016/j.cub.2024.03.034.

[2] Image taken from https://www.pexels.com/photo/brown-rat-eating-food-2189599/.