Jürgensen, Anna-Maria ORCID: 0000-0002-7871-1887 (2023). Circuit motifs for sensory integration, learning, and the initiation of adaptive behavior in Drosophila. PhD thesis, Universität zu Köln.
PDF (Doctoral Thesis)
Anna_Thesis_3.pdf - Accepted Version Download (34MB) |
Abstract
Goal-directed behavior is crucial for survival in complex, dynamic environments. It requires the detection of relevant sensory stimuli and the formation of separable neuronal representations. Learning the contingencies of these sensory stimuli with innately positive or negative valent stimuli (reinforcement) forms associations, allowing the former to cue the latter. This yields cue-based predictions to upgrade the behavioral repertoire from reactive to anticipatory. In this thesis, the Trias of sensory integration, learning of contingencies, and the initiation of anticipatory behavior are studied in the framework of the fruit fly Drosophila olfactory pathway and mushroom body, a higher-order brain center for integrating sensory input and coincidence detection using computational network models representing the mushroom body architecture with varying degrees of abstraction. Additionally, simulations of larval locomotion were employed to investigate how the output of the mushroom body relates to behavior and to foster comparability with animal experiments. We showed that inhibitory feedback within the mushroom body produces sparse stimulus representations, increasing the separability of different sensory stimuli. This separability reduced reinforcement generalization in learning experiments through the decreased overlap of stimulus representations. Furthermore, we showed that feedback from the valence-signaling output to the reinforcement-signaling dopaminergic neurons that innervate the mushroom body could explain experimentally observed temporal dynamics of the formation of associations between sensory cues and reinforcement. This supports the hypothesis that dopaminergic neurons encode the difference between predicted and received reinforcement, which in turn drives the learning process. These dopaminergic neurons have also been argued to convey an indirect reinforcement signal in second-order learning experiments. A new sensory cue is paired with an already established one that activates dopaminergic neurons due to its association with the reinforcement. We demonstrated how different pathways for feedforward or feedback input from the mushroom body’s intrinsic or output neurons can provide an indirect reinforcement signal to the dopaminergic neurons. Any direct or indirect association of sensory cues with reinforcement yielded a reinforcement expectation, biasing the fly’s behavioral response towards the approach or avoidance of the respective sensory cue. We then showed that the simulated locomotory behavior of individual animals in a virtual environment depends on the biasing output of the mushroom body. In conclusion, our results contribute to understanding the implementation of mechanisms for separable stimulus representations, postulated key features of associative learning, and the link between MB output and adaptive behavior in the mushroom body and confirm their explanatory power for animal behavior.
Item Type: | Thesis (PhD thesis) | ||||||||||||||||||
Translated abstract: |
|
||||||||||||||||||
Creators: |
|
||||||||||||||||||
URN: | urn:nbn:de:hbz:38-718362 | ||||||||||||||||||
Date: | 2023 | ||||||||||||||||||
Language: | English | ||||||||||||||||||
Faculty: | Faculty of Mathematics and Natural Sciences | ||||||||||||||||||
Divisions: | Faculty of Mathematics and Natural Sciences > Department of Biology > Zoologisches Institut | ||||||||||||||||||
Subjects: | Natural sciences and mathematics | ||||||||||||||||||
Uncontrolled Keywords: |
|
||||||||||||||||||
Date of oral exam: | 5 May 2023 | ||||||||||||||||||
Referee: |
|
||||||||||||||||||
Refereed: | Yes | ||||||||||||||||||
URI: | http://kups.ub.uni-koeln.de/id/eprint/71836 |
Downloads
Downloads per month over past year
Export
Actions (login required)
View Item |