Predictive Uncertainty in State-Estimation Drives Active Sensing

Osman Kaan Karagoz,Aysegul Kilic,Emin Yusuf AYDIN,Mustafa Mert Ankarali,Ismail Uyanik

DOI: https://doi.org/10.1101/2023.11.02.565312

2024-06-04

Abstract:Animals use active sensing movements to shape the spatiotemporal characteristics of sensory signals to better perceive their environment under varying conditions. However, the underlying mechanisms governing the generation of active sensing movements are not known. To address this, we investigated the role of active sensing movements in the refuge tracking behavior of , a species of weakly electric fish. These fish track the longitudinal movements of a refuge in which they hide by swimming back and forth in a single linear dimension. During refuge tracking, exhibits stereotyped whole-body oscillations when the quality of the sensory signals degrades. We developed a feedback control model to examine the role of these ancillary movements on the task performance. Here, we show that the proposed model generates fish trajectories that are statistically indistinguishable from the actual fish, implying that active sensing movements are regulated to minimize the predictive uncertainty in state estimation.

Animal Behavior and Cognition

What problem does this paper attempt to address?

The paper primarily explores the active sensing mechanisms in weakly electric fish (using Eigenmannia virescens as an example) during refuge tracking behavior and attempts to establish a feedback control model to explain how these animals minimize the predictive uncertainty of state estimation through active sensing movements.

The research background points out that animals shape the spatiotemporal characteristics of sensory signals through active sensing movements to better perceive their environment. However, the fundamental mechanisms regulating these active sensing movements remain unclear. To address this issue, researchers studied the role of active sensing movements in weakly electric fish during refuge tracking behavior. These fish track the longitudinal movement of a refuge by swimming back and forth along a single dimension. During the refuge tracking process, when the quality of sensory signals declines, these fish exhibit characteristic whole-body oscillations.

The paper proposes a feedback control model to investigate the impact of these auxiliary movements on task performance. The results show that the trajectories generated by the model are statistically indistinguishable from the actual fish trajectories, indicating that active sensing movements are regulated to minimize the predictive uncertainty in state estimation.

In summary, the paper aims to address the question of how mathematical modeling can be used to understand the mechanisms by which weakly electric fish improve the accuracy of state estimation through active sensing movements during refuge tracking behavior.