| Summary: | Periodic changes of environmental signals are sufficient to synchronise circadian rhythms across species. Circadian time, then, is a concept tethered to a diverse spread of different sensory modalities. In spite of this fact, circadian systems have historically been studied in a unimodal fashion investigating the processing of singular cues, while keeping others constant. My research sought to challenge this dogma via exploration of multisensory cue combination in the circadian clock of Drosophila melanogaster. Systematic behavioural analysis in wild type flies showed that misalignments between light and temperature (two potent environmental cues) produced abnormal profiles of circadian locomotor activity. Further molecular investigation revealed this behavioural disruption was associated with a breakdown of molecular rhythms in central clock neurons. Both the behavioural and molecular phenotypes observed during sensory conflict depended on the circadian photoreceptor, cryptochrome. Outside the central clock network, the circadian system of fruit flies forms an extensive network of peripheral oscillators. A luciferase reporter assay showed that photic signals play a more prominent role in peripheral clocks, compared to the core clock network in the brain. Here, molecular rhythms displayed continued light preference during sensory conflict, which again depended on cryptochrome. To further explore the implications of multisensory processing, with particular focus on the blurred boundary between clock input and output, circadian gene expression was evaluated in the ‘Johnston’s Organ’ - a key mechanosensory appa- ratus in Drosophila. These preliminary data suggest the existence of a previously unidentified peripheral clock in the fruit fly ear. Finally, a statistical model of the circadian clock was developed using a novel graphical architecture based on the hidden Markov model framework. This model was capable of inferring the phase of an underlying clock from both simulated and experimental locomotor datasets. More broadly, learning the parameters of this model from the data produced a probabilistic representation of the system, including its phase response dynamics.
|
|---|
