Computational neuroscience · Replication and mechanism comparisons
Fly Connectome: Testing a Model Under Perturbation
A connectome is a starting point, not the answer.
Replicate a connectome-driven whole-brain fly model, probe instability with background input, and compare changes to neuron signs, coupling and synaptic mechanisms.
Research question
Does connecting neurons produce plausible neural activity? Starting from an existing whole-brain model, this study adds sensory background input within a physiological range and asks whether responses remain plausible as input changes and is withdrawn.
Approach
- Replicate the original model on FlyWire v630 and v783, measuring firing and activity that persists after input withdrawal.
- Trace instability along the olfactory circuit and compare neuron signs, local coupling, input weights and short-term synaptic depression.
- Shuffle the added coupling locations to distinguish a specific structural effect from more general damping.
Where it stands
Replication and several perturbation comparisons are complete. Some modified configurations suppress self-sustaining excitation, but shuffled coupling locations also stabilize the model. Stability therefore cannot be attributed to a specific gap-junction topology. Firing levels after modification and the broader biological interpretation need further work.
Personal replication and comparison study based on the whole-brain fly model by Shiu and colleagues and the FlyWire connectome.