Analysis of inter-leg coordination mechanisms in cricket locomotion: insights from thoracic Ganglion network...

Adaptive insect locomotion depends on interactions between the nervous system and body dynamics, yet how these components contribute to inter-leg coordination remains unclear. We investigated the role of direct neural coupling by unilaterally transecting the posterior intermediate connective linking the mesothoracic and metathoracic ganglia in the cricketGryllus bimaculatus. We analyzed kinematic changes and compared them with predictions from a purely neural phase oscillator network model featuring hierarchical asymmetric coupling. The experiments showed a clear dissociation: transection disrupted anti-phase coordination between the contralateral hind legs (from 188to 104, Cohen's) while frequency synchronization across all six legs persisted despite a substantial overall frequency reduction (from 4.8 Hz to 2.0 Hz). Amplitude and mean angle also changed in legs ipsilateral to the transection. The neural model reproduced the phase shift quantitatively (98.25, 5.5% error) but selectively departed from the biological data in three respects-the overall frequency reduction, the maintenance of synchronization among decoupled legs, and local amplitude and mean angle changes. Each departure points to a distinct role of embodied dynamics beyond neural connectivity alone. These results demonstrate a functional two-layer architecture: neural coupling establishes the baseline phase relationships (timing coordination), whereas descending excitatory drive, embodied dynamics, and sensory feedback cooperatively regulate movement frequency and amplitude (tempo and magnitude control). By experimentally isolating neural connectivity and comparing biological responses with a neural-only model, this study disentangles the distinct contributions of central circuits and physical dynamics. The findings provide biological evidence for a hybrid control strategy in which fixed neural patterning establishes coordination templates continuously adapted by mechanical interactions, offering bio-inspired design principles for resilient legged robots.