Flexible Steering and Conflict Resolution: Pro-Goal/Anti-Goal Gating in Drosophila Lateral Accessory Lobes
Liao, C.-C.; Chang, N.; Liu, Y.-S.; Lo, C.-C.
Show abstract
Navigation comprises multiple processes: sensory integration, position estimation, decision, and finally, motor command generation, which is the control problem. Previous studies discovered that PFL3 neurons compute a heading-goal error and directly project to the descending neurons that generate the steering commands. Yet, this direct pathway is necessary but not sufficient for the flexible control required to pursue goal while avoiding threats and resolving conflicts. Using the Drosophila connectome and circuit modeling, we uncover a layered control architecture in the lateral accessory lobes (LAL) that arbitrates between goal pursuit and stimulus-driving overrides. First, a dorsal pathway that forms a dominant indirect "pro-goal" push-pull circuit that amplifies left-right asymmetries. Second, a faster ventral pathway forms an "anti-goal" circuit that recruits and inverts the pro-goal circuit to redirect movement, and only engages when a goal is present. The architecture also resolves two symmetry challenges: Rear Stalemate (goal 180 behind) and Front Stalemate (threat dead ahead), by rapidly amplifying tiny perturbations into decisive turns. Together, these motifs instantiate comparator, override, and gating principles in a compact neural controller, yielding testable predictions for insect motor control and design rules for bio-inspired, embodied intelligence.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Behavioral state and stimulus strength regulate the role of somatostatin interneurons in stabilizing network activity 97%
- Computational functions of precisely balanced neuronal assemblies in an olfactory memory network 97%
- The spinal premotor network driving scratching flexor and extensor alternation 97%
"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.