Direct comparison reveals algorithmic similarities in fly and mouse visual motion detection

Area of Science:

• Neuroscience
• Comparative Biology
• Computational Vision

Background:

• Vertebrates and invertebrates possess neural circuits for visual motion encoding.
• Previous studies noted computational similarities between mouse and fruit fly motion circuits, but lacked direct experimental comparison.
• Molecular and morphological differences exist between species, necessitating algorithmic-level analysis.

Purpose of the Study:

• To directly compare visual motion encoding at an algorithmic level between mouse and fruit fly.
• To investigate similarities and differences in analogous neurons: mouse ON starburst amacrine cells (ON SACs) and *Drosophila* T4 neurons.
• To identify conserved and divergent strategies in neural motion processing.

Main Methods:

• Utilized matched visual stimuli for both species.
• Focused analysis on the mouse ON SAC and *Drosophila* T4 neurons.
• Performed algorithmic-level comparison of neural responses.

Main Results:

• Mouse ON SACs and *Drosophila* T4 neurons exhibit similar spatiotemporal receptive field structures.
• Both neuron types show comparable sensitivity to spatiotemporal correlations and tuning to drifting gratings.
• Significant differences were observed in responses to apparent motion stimuli.
• Both neurons displayed responses to summed sinusoids deviating from existing models.

Conclusions:

• Despite distinct molecular and morphological characteristics, mouse and fruit fly neurons share fundamental principles of visual motion encoding.
• Apparent motion processing represents a key divergence point between these species.
• The observed deviations in summed sinusoid responses indicate gaps in current motion processing models, requiring further investigation.