NeurOpto — Implantable Optical Neuromodulation for Treatment-Resistant Depression

The Problem

332 million people live with depression. 30% develop treatment-resistant depression — having failed two or more adequate pharmacotherapy trials. For these patients, every existing option shares one fatal flaw: no therapy can distinguish a depression-driving neuron from its neighbour.

Antidepressants — systemic, slow onset, no circuit specificity
TMS — superficial, imprecise, daily sessions over weeks
DBS — highly invasive, surgical risk, indiscriminate stimulation with off-target effects

The result: 100 million people globally with no adequate treatment.

The Science

Depression involves dysfunction in the nucleus accumbens (NAc) reward circuit. In healthy brains, D1 MSNs promote reward and positive affect. D2 MSNs promote aversion and behavioural suppression. In depression, D1 activity decreases and D2 dominance increases — driving anhedonia, negative bias, and loss of motivation.

Optogenetics solves this directly:

AAV gene therapy expresses light-sensitive opsins in specific neuron populations
470nm light activates D1 neurons via ChR2 — restoring reward processing
550nm light inhibits D2 neurons via eNpHR3.0 — reducing negative valence signalling
Dual-wavelength bidirectional control through a single implanted fiber optic probe

The System

Phase 1 — Gene Delivery (Week 2)

Robeauté's 1.8mm self-propelled surgical microrobot navigates curved 3D paths through brain tissue with sub-millimetre precision. AAV viral vectors are delivered directly to the NAc through a minimal skull burr hole — reaching targets inaccessible to straight-line stereotaxic approaches. The robot is here at this hackathon as a Gold Sponsor and has demonstrated safe intraparenchymal navigation in large animal models with first-in-human trials planned for 2026.

Phase 2 — Device Implantation (Week 4)

A dual-wavelength fiber optic lead is implanted via a minimally invasive endovascular approach through the jugular vein. A subcutaneous controller and battery unit sits in the chest wall — identical in form factor to existing cardiac rhythm devices.

Phase 3 — Biomarker Calibration (Week 6)

Following AAV expression, patient-specific depression biomarkers are identified from neural recordings. A neural classifier is trained offline on local field potential signatures of depressive versus healthy states. Weights are frozen and loaded onto the implanted low-power processor. No on-device learning required — hardcoded inference at implant-grade power consumption.

Phase 4 — Closed-Loop Deployment (Week 14+)

The device continuously monitors neural activity, detects depression-state biomarkers in real time, and automatically delivers the appropriate wavelength of light to restore circuit balance. Monthly clinic checkups allow recalibration as needed.

The Demo

We built a working closed-loop neural state detection system using a g.tec Unicorn BCI EEG headset and a custom real-time classification pipeline in Python. The system detects neural biomarkers and triggers a simulated optogenetic response — demonstrating the closed-loop detection-to-stimulation architecture that is the core innovation of NeurOpto.

The full system replaces scalp EEG with implanted fiber optic biosensing — same closed-loop architecture, cellular resolution, direct circuit access.

Why Now

Three converging developments make NeurOpto feasible today:

Robeauté's microrobot has demonstrated safe intraparenchymal navigation in large animal models — solving the precision delivery problem for deep brain targets
AAV gene therapy has achieved regulatory approval for neurological indications (Zolgensma, Luxturna) — establishing the safety and manufacturing pathway
Closed-loop neuromodulation has achieved FDA approval (Saluda EVOKE, Medtronic Percept) — proving the regulatory pathway for this device class

Market

Metric	Value
Global antidepressant market	$18B today
Annual market growth (CAGR)	5–7%
Treatment-resistant depression population	~100M globally
Patients with no remaining treatment options	~30% of diagnosed

FDA Breakthrough Device Designation — granted to closed-loop neuromodulation systems — would accelerate our regulatory pathway significantly.

Roadmap

Stage	Timeline	Milestone
Stage 1	~2029	Open-loop dual-wavelength benchtop prototype
Stage 2	~2031	Closed-loop optical biosensor validated
Stage 3	~2034	First human implant of closed-loop system
Stage 4	~2037	Expansion to schizophrenia, PTSD, bipolar disorder

Platform Potential

The cell-type specific closed-loop architecture is not limited to depression. The same platform — AAV delivery, fiber optic stimulation, neural biomarker classification — applies directly to any psychiatric disorder with established circuit-level dysfunction: schizophrenia, PTSD, bipolar disorder, OCD. NeurOpto is a platform, not a single indication.

Team — University of Oxford

Name	Role	Expertise
Mia Gemayel	2nd Year DPhil	Optogenetics, AAV delivery, opsin engineering
Mattan Pelah	2nd Year DPhil	Implanted closed-loop control systems, neural biomarkers
Helene Brinkschulte	MSc Quantum Sensing	NV diamond magnetometry, fiber optic biosensor design
Leanne Mabena	2nd Year Medical Student	Clinical pathway, regulatory strategy, surgical feasibility
Yuzhong (WeiWei) Luo	MEng Engineering Science	Full-stack engineering, ML pipeline, system integration