Fly circuit exploration

What this site is. A connectome-mining exercise done with an LLM over two days in September 2026, audited afterwards. It reads the connectome and the literature and reports no new experiments. Each finding page opens with a box labelling every claim as already published, measured here in the wiring, simulated here, or proposed here and untested. Candidate neurons for vector memory were proposed by Maimon & Abbott 2026 and two have since been recorded (Janke 2025; Avritzer 2026); the contribution here is the wiring screen that ranks the unrecorded ones. Corrections are logged in the audit.