Photoreceptor microsaccades in Drosophila: Three-dimensional charaterisation

My thesis is about the fruit fly (Drosophila melanogaster) photoreceptor microsaccades.
These are light-induced rapid micrometer-sized photoreceptor displacements caused by
their photomechanically contracting microvilli. I used two imaging methods to investigate
this phenomenon: synchrotron X-rays and the deep pseudopupil. I was motivated by our
group’s previous work, which showed that the microsaccades improve the Drosophila vision
beyond the optical limits of a static eye. I desired to characterise the microsaccades in 3-
dimensions.

First, our high-speed synchrotron X-ray experiments led to an unexpected discovery that
the X-rays themselves activate phototransduction similar to visible light. The X-ray induced
microsaccades were left-right mirror-symmetric and stronger where the two eyes’ visual
fields meet. In addition, the motion analysis along the receptor’s side-profile suggested
spatial specialisations. Second, I developed the first goniometric (angle-measuring) high-
speed deep pseudopupil (GHS-DPP) system for mapping photoreceptor microsaccades
across the left and right eyes. My results suggest that the microsaccades are tuned to
the optic flow and are developmentally set by the R1-R8 rhabdomere orientations, which
also follow the optic flow. Third, the localised GHS-DPP experiments using rhodopsin-
specific rescues showed that all R1-R8 rhabdomeres contributed towards the microsaccades,
irrespective how I immobilised the flies. Furthermore, light stimulation with sinusoidal and
frequency sweeps showed that the microsaccades could follow temporal contrast changes
up to 30 Hz (3 dB cut-off at 13 Hz), indicating that the microsaccades happen during light
contrast changes in the natural environment.

Overall, my results suggest that the photoreceptor microsaccades improve vision, accommodating the fly’s behavioural needs. The insect eyes have evolved over hundreds
of millions of years. My research helps to better understand some less obvious design
choices, broadening our knowledge about how visual sensory systems work dynamically.
The potential applications of this work range from man-made image sensors to stereo vision
algorithms, autonomous drones, consumer electronics and improved medical implants,
such as bionic eyes.