A computational model for angular velocity integration in a locust heading circuit
Loading...
Files
Date
Publisher
Philipps-Universität Marburg
Supervisors
Abstract
Insect navigation and its neural underpinnings provide an excellent model for inves- tigating basal cognitive abilities. This thesis focuses on the integration of angular velocity cues into the compass-like encoding of heading direction in the central com- plex of the desert locust. The central complex serves as a navigation hub within the insect brain. While a core circuitry is shared across all insects, physiological data suggest a topographical difference in the internal heading representations of the desert locust and the fruit fly. Data supporting this notion are limited, and computational modelling can bridge such knowledge gaps by formulating and test- ing hypotheses in a theoretical framework. The overarching aim was to examine the viability of the neural activity observed in the locust as a heading signal.
Computational models of the desert locust heading circuit were based on available structural and functional data. Under-specified model parameters were determined via a machine learning-based approach. These models suggest that, like in the fruit fly heading circuit, recurrent loops are the motifs maintaining the heading signal. Updating the heading signal by integrating feed-forward angular velocity inputs as described in the fruit fly however is incompatible with the model of the locust head- ing circuit. Instead, multiplicative inputs modulating the circuit connectivity via neuromodulation at the synapse-level are proposed. An analysis of optic flow sen- sitivity of locust central complex neurons revealed responses to the simulation of rotational self-motion in neurons homologous to those conveying rotation informa- tion to the fruit fly heading circuit. Further examinations are required to determine the congruence of these results with the proposed neuromodulatory mechanism for integrating rotational self-motion cues.
This work supports the notion of distinct compass topographies in the heading circuits of fruit flies and desert locusts. The developed models generate testable hy- potheses guiding future anatomical, functional, and behavioural studies, contribut- ing to computational approaches to the fields of insect navigation and comparative cognition.
Review
Metadata
Contributors
Supervisor:
Dates
Created: 2024Issued: 2024-08-20Updated: 2024-08-20
Faculty
Fachbereich Psychologie
Publisher
Philipps-Universität Marburg
Language
eng
Data types
DoctoralThesis
DFG-subjects
navigationmachine learningdesert locustcomputational modelsky compasscentral complex
DDC-Numbers
500
show more
Pabst, Kathrin (0000-0001-8407-328X): A computational model for angular velocity integration in a locust heading circuit. : Philipps-Universität Marburg 2024-08-20. DOI: https://doi.org/10.17192/z2024.0235.