FLYHUB WHITEPAPER · V1.0

The internet for a 100,000-neuron brain.

FlyHub is a speculative research interface for discovering which visual patterns most strongly guide Drosophila attention and courtship-associated behavior.

The site does not claim that flies experience media, attraction or preference in a human sense. It turns measurable behavior into a familiar interface so a closed-loop experiment can be understood at a glance.

01 · Abstract

Modern Drosophila neuroscience can map neural circuits with extraordinary detail, but a connectome alone does not explain how a living animal chooses, attends or reacts. FlyHub proposes a behavioral layer: show controlled visual stimuli, measure the fly's response, and let an algorithm select the next stimulus.

02 · The research question

Can an adaptive system discover a visual pattern that reliably produces stronger pursuit and orientation than naturally occurring targets? The question links sensory neuroscience, animal behavior and machine learning. To a human observer, the winning pattern may be only a moving dot. To the fly's visual system, it may be unusually salient.

03 · The closed-loop experiment

01Generate stimulus
02Display to fly
03Track behavior
04Calculate score
05Adapt parameters

A fly walks on a spherical treadmill while a high-refresh display presents a target. A camera estimates position, heading and wing posture. The system converts these measurements into an engagement score and adjusts speed, size, contrast or trajectory for the next trial.

04 · Behavioral metrics

“Likes,” “views” and recommendations are interface metaphors. A real study would retain raw measurements, controls, trial metadata and uncertainty—not only a single score.

05 · Scientific value

The platform could help researchers search a large stimulus space without hand-designing every trial. It could reveal supernormal stimuli, compare individual differences, test how internal state changes perception, and connect circuit-level hypotheses with observable behavior.

FlyHub also acts as a public communication tool. By borrowing the grammar of a recommendation platform, it makes one idea intuitive: an algorithm can learn from an animal's actions without translating those actions into language.

06 · Ethics and limitations

Animal behavior must not be casually anthropomorphized. Strong pursuit is evidence of a motor response, not proof of pleasure, desire or a human-like subjective experience. Any real implementation should use the minimum number of animals, minimize restraint and stress, define stopping criteria, and follow applicable institutional animal-care review.

Current prototype: This website uses simulated records only. It is an interface and experimental concept, not a report of completed findings.