brain_worker.py

Not to be confused with brain_render_worker.py

brain_worker.py

A background QThread dedicated to handling computationally expensive brain logic that would otherwise block the UI. It operates on cached snapshots of brain state, processes tasks from a thread-safe queue, and emits results back to the main thread via signals. This separation keeps the simulation responsive even during complex Hebbian learning calculations or neurogenesis evaluations.

Task Queue System:

• Uses thread-safe Queue for task dispatch
• Processes three task types: neurogenesis, hebbian, state_update
• Supports pause/resume for game state changes

Neurogenesis Checks:

• Evaluates stress/novelty/reward triggers against configurable thresholds
• Emits neurogenesis_result signal with creation recommendations
• Handles emergency stress neuron creation when anxiety exceeds 90

Hebbian Learning:

• Selects top-k neuron pairs based on co-activation scores
• Includes anti-loop mechanisms (randomization, cooldown penalties on recently-used pairs)
• Can create new connections between previously unconnected co-active neurons
• Emits weight updates back to main thread for application

State Decay Processing:

• Applies temporal decay toward baseline for non-input neurons
• Adds small random noise for organic feel
• Propagates connection effects through the network