Unsupervised learning reveals rapid gait adaptation after leg loss and regrowth in spiders.

Abstract

Many invertebrates voluntarily lose (autotomize) limbs during antagonistic encounters, and some regenerate functional replacements. Because limb loss can have severe consequences on individual fitness, it is likely subject to significant selective pressures, making this an excellent phenomenon with which to investigate biomechanical robustness. Spiders frequently autotomize one or more legs. We investigated the time course of locomotor recovery after leg loss and regeneration in juvenile tarantulas (Arachnida: Araneae) naive to autotomy. We recorded high-speed video of spiders running with all legs intact, then immediately after, and again 1 day after they had autotomized two legs. The legs were allowed to regenerate, and the same sequence of experiments was repeated. Video tracking analysis revealed that the spiders resumed their pre-autotomy speed and stride frequency after leg regeneration and in ≤1 day after both autotomies; path tortuosity was unaffected by these treatments. Autotomized spiders widened the spread of their remaining legs for stability and to compensate for missing functional space. To analyze how their gaits changed in response to leg loss, we applied unsupervised machine learning for the first time to measured kinematic data in combination with gait space metrics. Spiders were found to robustly adopt new gait patterns immediately after losing legs, with no evidence of learning. This novel clustering approach both demonstrated concordance with hypothesized gaits and revealed transitions between and variations within these patterns. More generally, clustering in gait space enables the identification of patterns of leg motions in large datasets that correspond to either known gaits or undiscovered behaviors.