Skip to content

ppradyoth/synaptic-wetware

Repository files navigation

🧠 Synaptic Wetware

A state-of-the-art interactive simulator for Organoid Intelligence (OI) biocomputing β€” built to be both scientifically rigorous and accessible to anyone.

Grew brain cells on a chip. Taught them to play Pong. Built a dashboard for it.

Live Demo GitHub License: MIT Built with Vite


What Is This?

Synaptic Wetware is an interactive web dashboard simulating what happens inside a real organoid intelligence biocomputer lab β€” where living human brain cells grown on silicon chips are used as biological processors.

Every number in this simulator is grounded in real neuroscience:

  • The voltage traces use real Hodgkin-Huxley and Izhikevich neuron models
  • Burst detection uses the MaxInterval algorithm (same as MEA-NAP / meaRtools)
  • The Pong training loop replicates the actual DishBrain Free Energy Principle feedback
  • The ethics panel follows the Baltimore Declaration on Organoid Intelligence
  • Benchmark figures come from published Nature and Frontiers papers

πŸ†• New to this topic? Read EXPLAINER.md β€” a plain-English guide from zero to expert, written for fresh grads and curious people.


Features

🌱 Stem-Cell Grow Room

  • L-system axon branching pathfinding on an HTML5 Canvas
  • Nerve Growth Factor (NGF) injection triggers growth surges
  • Live metrics: cell count, synaptic density, myelination factor

⚑ Electrophysiology MEA (8Γ—8 Grid)

  • 64 interactive micro-electrode channels β€” click to stimulate
  • Real-time burst detection (MaxInterval): burst frequency, mean IBI, network synchrony
  • Rolling oscilloscope trace
  • Spike event counter (4-second rolling window)

πŸ”¬ Dual Neuron Physics Models (toggle in sidebar)

Model Equations Accuracy Cost
Izhikevich 2 ODEs Biologically plausible spike shapes Fast (200 steps/tick)
Hodgkin-Huxley 4 ODEs (Na⁺/K⁺/leak channels) Nobel Prize-level accuracy (1963) Full (1000 steps/tick)

Both models run on all 64 electrodes simultaneously using Euler integration inside the simulation loop.

πŸ•ΉοΈ Cognitive Conditioning (Training Playground)

  • DishBrain Pong β€” biological feedback loop replicating the Cortical Labs experiment
  • Logic Gates (AND/OR/XOR) β€” wetware logic learning
  • Learning efficiency modulated by dopamine, GABA, and synaptic density

πŸ§ͺ Incubator Life-Support

  • Temperature, Glucose, Oβ‚‚, Dopamine, GABA sliders
  • Cell death triggered by starvation
  • Chemical depletion from cell metabolism
  • Seizure detection when network synchrony > 70%

βš–οΈ Ethics Monitor (always-visible sidebar widget)

  • IIT-Ξ¦ proxy gauge β€” simplified Integrated Information Theory measurement
  • Sentience risk score (0–100) β€” calibrated from synchrony Γ— cell count Γ— synaptic density
  • Baltimore Declaration compliance checklist β€” 5 items, live-updating
  • Welfare event log β€” auto-logs threshold crossings
  • Welfare levels: Safe β†’ Monitor β†’ Review Required β†’ Halt Protocol

πŸ“Š Silicon vs. Biotech Benchmarks

  • Power draw comparison (organoid ~10Β΅W vs H100 cluster ~700W)
  • Synaptic density (3D biological vs 2D silicon)
  • Carbon footprint
  • Market projections

πŸ“– About / Explainer Tab

  • 11-section interactive accordion explaining the science
  • Covers: organoids, MEA, DishBrain, AI vs biology, transformers vs neurons, backprop vs Hebbian learning, ethics, and where this is all going
  • Written for a 15-year-old with no background

Tech Stack

Layer Technology
Framework React 18 + TypeScript
Build Vite 6
Styling Vanilla CSS (glassmorphism, custom animations)
Fonts Google Fonts β€” Outfit + JetBrains Mono
Icons Lucide React
Physics Custom Hodgkin-Huxley + Izhikevich Euler integration
Deploy Vercel

No Tailwind. No external charting libraries. All visualisations are raw HTML5 Canvas or CSS.


Run Locally

git clone https://github.com/ppradyoth/synaptic-wetware.git
cd synaptic-wetware
npm install
npm run dev

Open http://localhost:5173


Project Structure

src/
β”œβ”€β”€ hooks/
β”‚   └── useWetwareSim.ts        # Core simulation engine
β”‚                               #   β€” HH + Izhikevich models
β”‚                               #   β€” MaxInterval burst detection
β”‚                               #   β€” Ethics metrics (IIT-Ξ¦ proxy)
β”‚                               #   β€” Raster plot spike events
β”œβ”€β”€ components/
β”‚   β”œβ”€β”€ GrowRoom.tsx            # Axon branching canvas
β”‚   β”œβ”€β”€ ElectrophysiologyGrid.tsx  # MEA + oscilloscope + burst metrics
β”‚   β”œβ”€β”€ TrainingPlayground.tsx  # Pong + logic gates
β”‚   β”œβ”€β”€ IncubatorControls.tsx   # Life support dashboard
β”‚   β”œβ”€β”€ Benchmarks.tsx          # Silicon vs wetware comparisons
β”‚   β”œβ”€β”€ EthicsPanel.tsx         # Baltimore Declaration compliance widget
β”‚   └── About.tsx               # Plain-English explainer accordion
β”œβ”€β”€ App.tsx                     # Lab shell β€” model toggle, sidebar, routing
β”œβ”€β”€ index.css                   # Bio-cybernetic design system
└── main.tsx
EXPLAINER.md                    # Deep-dive guide for fresh grads

Scientific Basis

Feature Real-World Reference
DishBrain Pong training Kagan et al., Nature Electronics (2022)
Baltimore Declaration ethics Bhanu et al., Nature (2024)
FinalSpark neuroplatform finalspark.com, Frontiers in Neuroscience (2024)
Hodgkin-Huxley model Hodgkin & Huxley, Journal of Physiology (1952) β€” Nobel Prize 1963
Izhikevich model Izhikevich, IEEE Trans Neural Networks (2003)
MaxInterval burst detection Pasquale et al., Journal of Neuroscience Methods (2010)
IIT / Ξ¦ Tononi et al., BMC Neuroscience (2008)
Energy efficiency figures FinalSpark whitepaper (2024), Cortical Labs CL-1 datasheet

Credits

Conceived, researched, and built by @ppradyoth β€” with the assistance of Antigravity, an agentic AI coding assistant by the Google DeepMind Advanced Agentic Coding team.


License

MIT β€” do whatever you want with it. If you publish research using this simulator, a citation would be appreciated.

About

🧠 Organoid Intelligence Biocomputer Simulator β€” HH + Izhikevich neuron models, MEA burst detection, DishBrain Pong, Baltimore Declaration ethics monitor. Built by Antigravity (Google DeepMind).

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages