The Crunchometer: A Low-Cost, Open-Source Acoustic Analysis of Feeding Microstructure
Elucidating the neuronal circuits that govern appetite requires a detailed analysis of the microstructure of solid food consumption. A significant barrier in this field is that existing techniques for monitoring feeding are either prohibitively expensive, limiting their use, or lack the high temporal resolution necessary to align feeding events with neuronal activity. To overcome these limitations and democratize this technology, we developed the Crunchometer, an open-source sound-based system that utilizes a cost-effective condenser microphone and video recordings, along with two mathematical methods (supervised thresholding and an unsupervised Support Vector Machine, SVM), to precisely capture bite sounds with unprecedented temporal resolution and built a feeding ethogram for both Chow and High-Fat Diet (HFD) pellets. While human observers tend to estimate longer, more frequent feeding bouts, our automated methods more precisely and reliably detect the start and end times of most feeding bouts. In consequence, both automated methods fragmented feeding bouts more frequently than human observers. We demonstrate the Crunchometer's utility by monitoring feeding patterns across different energy states, hunger (fasted), and satiety, and by characterizing the pharmacological effects of the anti-obesity drug semaglutide. This revealed that semaglutide not only suppresses solid food intake but also decreases the preference for HFD pellets. Additionally, we differentiate the genetically-evoked gnawing from consummatory feeding behaviors involving GABAergic neurons in the Lateral Hypothalamus (LH). Importantly, the Crunchometer integrates seamlessly with in vivo electrophysiology and calcium imaging in freely behaving mice. The Crunchometer, paired with recordings of LH activity, revealed the presence of meal-related neurons that monitor the entire meal, rather than just individual feeding bouts. Microendoscopes and calcium imaging on Vgat-IRES-cre and Vglut2-IRES-cre mice further suggest that GABA neurons in LH (but not glutamatergic neurons) are also strongly modulated by eating a solid food pellet. Notably, the acts of licking liquid sucrose and eating solid food activate distinct sets of neurons. Its compatibility with electrophysiology and calcium imaging makes the Crunchometer an ideal tool for elucidating the neuronal correlates of feeding behavior for solid foods.