Effects of Active Vision and Internal State on Drosophila Visual Behaviour

For many people, fruit flies are simply an annoyance. However, as they perform their persistent straight-line flights, followed by quick switches in direction, they display a remarkable ability in both vision and agility. A key part of the fruit fly visual system that ensures that this behaviour occurs is the ability to stabilise locomotion within the world given perturbation, this is the optomotor response. Recent studies have introduced the existence of anti-optomotor behaviour in Drosophila. Under certain conditions, flies show a consistent, large, and sustained response in the opposite direction to rotational optic flow. However, the underlying reasons for this behaviour remain unclear.

We aimed to investigate the features and reasons for the anti-optomotor response. In addition we aimed to investigate the ability of flies to use active sensing mechanisms to allow for greater visual motion perception.

Our findings reveal that flies perceive rotational optic flow normally during anti-optomotor behaviour. However, they exhibit significant visual nystagmus of the head, which coincides with the occurrence of anti-optomotor behaviour. Moreover, this behavior is linked to high-frequency locomotory activity, and looming stimuli that increase the likelihood of directional changes also amplify the occurrence of anti-optomotor behaviour. Interestingly, restraining head movements does not eliminate the anti-optomotor response, although it does impair overall turning ability at lower stimulus wavelengths. Furthermore, our study reveals that fruit flies possess better visual acuity than previously believed, as they respond to 2° wavelength gratings, well below the 4-5° Nyquist limit. Notably, this behaviour is observable exclusively through head-optomotor responses, emphasizing the significance of active sensing mechanisms in fly vision.

The characteristics of anti-optomotor behaviour suggest that a saccadic-like behavioural pattern resembling course-changing movements is the fundamental cause of this phenomenon, although the exact neural circuitry involved is still unknown. Additionally, the presence of hyperacute behavioural responses lends support to the microsaccadic sampling hypothesis of animal vision and furthermore explains previous contradictory findings in regards to this phenomenon.