Connectomics and robotics research
What if a fly's neural architecture could guide a robot we build ourselves? openflymind explores that possibility using MaleCNS data, connecting a computational model to perception and movement in the physical world.
Current status: Project in development
Scale of the scientific resource
openflymind starts from a connectome published by other teams. The figures below describe that dataset, not what the project has already imported, run or integrated into hardware.
Data: FlyEM Male CNS Connectomemale-cns:v1.0 · FlyEM / HHMI Janelia Research Campus · University of Cambridge · MRC Laboratory of Molecular Biology · Google Research · Licence: CC BY 4.0. Use of public data does not imply institutional collaboration.
The question
A connectome reveals an architecture of connections. Our challenge is to investigate how that structure can become a model that receives signals, produces actions and responds to the consequences of its own movements. We aim to develop a pioneering demonstration of MaleCNS robotic embodiment, with explicit methods and evidence that others can examine and reproduce.
The method
We will use connectivity data to build a sparse computational representation. Model dynamics and sensory and motor interfaces will be documented experimental choices. The proposed first stage combines vision and movement, with touch and hearing extending the research agenda.
Proposed first stage: vision and movement. Touch and hearing come later as extensions, subject to validation.
Photographs of the real equipment will appear here once it exists. Until then, no renders posing as a prototype.
The body under construction
The proposed initial scope is vision and movement: a camera as input and movement actuators with physical limits and a safe stop as output; the form of the platform will be defined by the team. Touch and hearing appear as extensions, dependent on validating the earlier stages.
No hardware has been assembled yet. Sensors, processing board, power and actuators will only be given commercial part numbers once the team confirms them. What exists today is the planned architecture and the agenda of engineering decisions.
Progress
Next milestone
In progress
Selection of MaleCNS, formulation of the question and definition of evaluation criteria.
Dataset chosen and question formulated; evaluation criteria being drafted. Decisions count as documented only once recorded in the journal.
Last real update: September 22, 2026
Real states for each stage, with no completion percentages. A "tested" stage may have a negative result.
Foundations and scopeIn progress
Dataset chosen and question formulated; evaluation criteria being drafted. Decisions count as documented only once recorded in the journal.
Import and graphPlanned
Model dynamicsPlanned
Vision and movement in simulationPlanned
Robotic bodyPlanned
Physical closed loopPlanned
Sensory extensionsPlanned
Exploratory.
Reproduction and communicationPlanned
Research journal
The first notes are being reviewed by the team and will be published here.
Team
Project Lead · Student Researcher
Engineering Mentor · IEEE Senior Member
Join the investigation
Follow the questions, engineering decisions and evidence shaping openflymind. Each stage will document its context, its tests and what we still need to understand.
Email subscription is temporarily unavailable. In the meantime, you can follow the feed or the journal.
Have a reference, a hypothesis or an experimental suggestion? Contribute to the research.