@phdthesis{Ernst1999,
  author    = {Ernst, Roman},
  title     = {Visuelle Mustererkennung und Parameterextraktion bei Drosophila melanogaster},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-1156},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {1999},
  abstract  = {In operanten Konditionierungsexperimenten im Flugsimulator werden vier Parameter gefunden die Drosophila melanogaster aus visuellen Mustern extrahieren kann: Musterfl{\"a}che, vertikale Position des Musterschwerpunkts, Verteiltheit und Musterausrichtung in horizontaler und vertikaler Richtung. Es ist nicht auszuschliessen, dass die Fliege weitere Musterparameter extrahieren kann. Spontane Musterpr{\"a}ferenzen und konditionierte Pr{\"a}ferenzen zeigen unterschiedliche Zusammenh{\"a}nge mit den Musterparametern. Aus r{\"a}umlich getrennten Musterelementen zusammengesetzte Muster werden von der Fliege wie ein Gesamtmuster behandelt. Retinaler Transfer wird auch bei der Pr{\"a}sentation von Mustern an zwei verschiedenen vertikalen Trainingspositionen nicht beobachtet. Muster werden generalisiert, wenn die Schwerpunkte korrespondierender Muster zwischen Training und Test ungef{\"a}hr an der gleichen Position liegen aber keine retinale {\"U}berlappung von Trainings- und Testmustern besteht. Retinotopie des Musterged{\"a}chtnisses liegt in diesem Fall nicht auf der Ebene der Bildpunkte, jedoch m{\"o}glicherweise auf der Ebene des Parameters 'Musterschwerpunkt' vor. Fliegen k{\"o}nnen nicht trainiert werden bestimmte Musterpaare zu diskriminieren die sich nur durch die vertikale Position ihres Musterschwerpunktes unterscheiden. Dennoch bevorzugen sie beim Lerntest mit anderen Mustern mit korrespondierenden Schwerpunktspositionen die zuvor nicht bestrafte Schwerpunktsposition. F{\"u}r die Modellierung der Extraktion von Musterschwerpunkt und Musterfl{\"a}che wird ein einfaches k{\"u}nstliches neuronales Filter pr{\"a}sentiert, dessen Architektur auf einem Berechnungsalgorithmus f{\"u}r den gemeinsamen Schwerpunkt mehrerer Teilelemente beruht.},
  subject      = {Taufliege},
  language  = {de}
}
@phdthesis{Nguyen2023,
  author    = {Nguyen, Tu Anh Thi},
  title     = {Neural coding of different visual cues in the monarch butterfly sun compass},
  doi       = {10.25972/OPUS-30380},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-303807},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2023},
  abstract  = {Monarch butterflies are famous for their annual long-distance migration. Decreasing temperatures and reduced daylight induce the migratory state in the autumn generation of monarch butterflies. Not only are they in a reproductive diapause, they also produce fat deposits to be prepared for the upcoming journey: Driven by their instinct to migrate, they depart from their eclosion grounds in the northern regions of the North American continent and start their southern journey to their hibernation spots in Central Mexico. The butterflies cover a distance of up to 4000 km across the United States. In the next spring, the same butterflies invert their preferred heading direction due to seasonal changes and start their northward spring migration. The spring migration is continued by three consecutive butterfly generations, until the animals repopulate the northern regions in North America as non-migratory monarch butterflies. The monarch butterflies' migratory state is genetically and epigenetically regulated, including the directed flight behavior. Therefore, the insect's internal compass system does not only have to encode the butterflies preferred, but also its current heading direction. However, the butterfly's internal heading representation has to be matched to external cues, to avoid departing from its initial flight path and increasing its risk of missing its desired destination. During the migratory flight, visual cues provide the butterflies with reliable orientation information. The butterflies refer to the sun as their main orientation cue. In addition to the sun, the butterflies likely use the polarization pattern of the sky for orientation. The sky compass signals are processed within a region in the brain, termed the central complex (CX). Previous research on the CX neural circuitry of the monarch butterflies demonstrated that tangential central complex neurons (TL) carry the visual input information into the CX and respond to a simulated sun and polarized light. However, whether these cells process additional visual cues like the panoramic skyline is still unknown. Furthermore, little is known about how the migratory state affects visual cue processing. In addition to this, most experiments studying the monarch butterfly CX focused on how neurons process single visual cues. However, how combined visual stimuli are processed in the CX is still unknown. This thesis is investigating the following questions: 1) How does the migratory state affect visual cue processing in the TL cells within the monarch butterfly brain? 2) How are multiple visual cues integrated in the TL cells? 3) How is compass information modulated in the CX? To study these questions, TL neurons from both animal groups (migratory and non-migratory) were electrophysiologically characterized using intracellular recordings while presenting different simulated celestial cues and visual sceneries. I showed that the TL neurons of migratory butterflies are more narrowly tuned to the sun, possibly helping them in keeping a directed flight course during migration. Furthermore, I found that TL cells encode a panoramic skyline, suggesting that the CX network combines celestial and terrestrial information. Experiments with combined celestial stimuli revealed that the TL cells combine both cue information linearly. However, if exposing the animals to a simulated visual scenery containing a panoramic skyline and a simulated sun, the single visual cues are weighted differently. These results indicate that the CX's input region can flexibly adapt to different visual cue conditions. Furthermore, I characterize a previously unknown neuron in the monarch butterfly CX which responds to celestial stimuli and connects the CX with other brain neuropiles. How this cell type affects heading direction encoding has yet to be determined.},
  subject      = {Monarchfalter},
  language  = {en}
}