Neural information processing in the Drosophila motion vision pathway

Mishra, Abhishek (2023): Neural information processing in the Drosophila motion vision pathway. Dissertation, LMU München: Graduate School of Systemic Neurosciences (GSN)

Abstract

Detecting the direction of image motion is an essential component of visual computation. An individual photoreceptor, however, does not explicitly represent the direction in which the image is shifting. Comparing neighboring photoreceptor signals over time is used to extract directional motion information from the photoreceptor array in the circuit downstream. To implement direction selectivity, two opposing models have been proposed. In both models, one input line is asymmetrically delayed compared to the other, followed by a non-linear interaction between the two input lines. The Hassenstein-Reichardt (HR) model proposes an enhancement in the preferred direction (PD): the preferred side signal is delayed and then amplified by multiplying it with the other input signal. In contrast, the Barlow-Levick (BL) detector proposes a null direction (ND) suppression, whereby the null side signal is delayed and the other input is divided by it. The motion information is computed in parallel ON and OFF pathways. T4 and T5 are the first direction-selective neurons found in the ON and in the OFF pathway, respectively. Four subtypes of T4 and T5 cells exist each responding selectively to one of the four cardinal directions: front-to-back, back-to-front, upwards, and downwards, respectively. In the first manuscript, we found that both preferred direction enhancement and null direction suppression are implemented in the dendrites of all four subtypes of both T4 and T5 cells to compute the direction of motion. We, therefore, propose a hybrid model combining both PD enhancement on the preferred side and ND suppression on the null side. This combined strategy ensures a high degree of direction selectivity already at the first stage of calculating motion direction. Further processing, in addition to synaptic mechanisms on the dendrites of T4 cells, can improve the direction selectivity of the T4 cells' output signals. Such processing might involve: 1.) transformation from voltage to calcium, and 2.) from calcium to neurotransmitter release. In the second manuscript, we used in vivo two-photon imaging of genetically encoded voltage and calcium indicators, Arclight and GCaMP6f respectively, to measure responses in Drosophila direction-selective T4 neurons. Comparison between Arclight and GCaMP6f signals revealed calcium signals to have a significantly higher direction selectivity compared to voltage signals. Using these recordings we built a model which transforms T4 voltage responses into calcium responses. The model reproduced experimentally measured calcium responses across different visual stimuli using various temporal filtering steps and a stationary non-linearity. These findings provided a mechanistic underpinning of the voltage-to-calcium transformation and showed how this processing step, in addition to synaptic mechanisms on the dendrites of T4 cells, enhances direction selectivity in the output signal of T4 neurons. The two manuscripts included in this thesis are presented chronologically and were published in peer-reviewed journals.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Keywords: direction selectivity, Drosophila, imaging, nonlinear model, voltage to calcium transformation, computational modelling
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 570 Biowissenschaften, Biologie
• Fakultäten: Graduate School of Systemic Neurosciences (GSN)
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 13. April 2023
• 1. Bericht­erstatter:in: Borst, Alexander