Back

Social feedback structures stay/switch decision-making in Drosophila

Aragon, M. J.; Murthy, M.

2026-10-08 neuroscience
10.64898/2026.10.01.756036 bioRxiv
Show abstract

Within natural environments, different sensory stimuli can drive mutually-exclusive behaviors such as fighting, feeding, or courting. When experiencing competing cues, how do animals decide to continue executing the current behavior or to switch to a new behavior? To address this question, we developed a closed-loop behavioral paradigm that presents actively courting male Drosophila with a fictive heat stimulus, forcing the male to choose between two competing behaviors: continue courting (stay) or escape the heat (switch). By modeling the sensorimotor transformations underlying this decision process, we discovered that male flies continually evaluate social cues, including female speed and her distance, in order to make a stay/switch decision. We investigated the neural correlates of stay/switch decision-making via brain-wide calcium imaging in head-fixed male flies presented with both fictive female visual cues and heat stimuli. We built models that predicted stay/switch decisions from neural data, which revealed that activity in a small number of brain regions, including the anterior optic tubercle, predicted decision outcomes up to 4 seconds prior to stimulus onset, whereas after stimulus onset, distributed activity across the brain was correlated with the behavioral decision. Together, these results demonstrate that Drosophila continually evaluate social feedback to inform decisions about how to respond to environmental cues, and that brain activity changes prior to presentation of the environmental cue can inform upcoming decisions. Overall, this study provides new insight into the processes underlying naturalistic decision-making in Drosophila.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.