🪼 NEURAL HYDRA 004 (JELLYFISH): Simulation of Proto-Consciousness with 5000 Neurons

🪼 NEURAL HYDRA 004 (JELLYFISH): Simulation of Proto-Consciousness with 5000 Neurons
I built a complete simulation of a jellyfish's nervous system and let it "think".

🧠 NEURAL NETWORK
I do not use classic artificial neural networks (like ChatGPT or DALL-E). Instead, I implemented a Spiking Neural Network (SNN) – a biologically faithful model that operates in the same way as real neurons in your brain.
Each of the 5000 neurons uses the Leaky Integrate-and-Fire (LIF) model:

• It has a membrane potential (resting state: -70 mV)
• It collects signals from other neurons
• When it exceeds the threshold (-55 mV), it "fires" a spike
• After firing, it has a refractory period (it cannot fire again immediately)

🔬 ARCHITECTURE

• 500 sensory neurons
  ─ 150 photoreceptors (light detection)
  ─ 200 chemoreceptors (food/toxin detection)
  ─ 150 mechanoreceptors (touch)

• 4300 interneurons
  ─ Processing and integration of signals

• 200 motor neurons
  ─ 16 tentacles × ~12 neurons

The neurons are connected by synapses with varying weights – some are excitatory (amplifying the signal), while others are inhibitory (dampening it). In total, there are approximately 150,000 synaptic connections.

💉 NEUROMODULATION (Hormonal System)
Real organisms do not only have neurons – they also have hormones that change how neurons respond. We implemented 5 "hormones":

• Stress – increases sensitivity, speeds up reactions
• Hunger – amplifies chemoreceptors, motivates searching
• Energy – affects movement speed
• Pain – triggers escape reflexes
• Excitement – overall system activation

These hormones modulate the threshold and sensitivity of the neurons in real-time.

🌊 EMERGENT BEHAVIOUR
Crucially, we did not programme any specific behaviour. Everything emerges from:

• Chemotaxis – movement along a chemical gradient towards food
• Negative phototaxis – avoidance of light
• Escape response – rapid fleeing from predators
• Homeostasis – maintaining energy balance

The jellyfish "realises" on its own that when it is hungry and senses a chemical gradient, it makes sense to move in that direction.

💭 CONSCIOUSNESS?
We added a layer of "thoughts" – the LLM generates an internal monologue based on the state of the neural network and hormones.
If we were to refine the simulation to the level of individual molecules... would the jellyfish have a subjective experience?

Better resolution: https://youtu.be/TUfeXumjzW8

Originally published on Facebook — link to post