fibertract

Peripheral nervous system substrate for neuromorphic systems.

Twin primitive to neuropool. Where neuropool provides spiking neural substrate (brain tissue), fibertract provides connective tissue (peripheral nerves).

Core Concepts

Fiber tracts are labeled lines. The tract kind defines what signals on that tract mean. A signal on a nociceptive tract IS pain. A signal on a proprioceptive tract IS body position. Same data format, different qualia -- determined by which tract carries it.

Asymmetric encoding. Motor tracts (efferent, brain -> body) carry Signal -- the brain's native ternary format {polarity: i8, magnitude: u8}. Sensory tracts (afferent, body -> brain) carry i32 -- wide dynamic range with Weber-law quantization for biologically realistic discrimination.

Use-dependent adaptation. Tracts don't learn (no plasticity). They adapt: conductivity improves with use, jitter decreases with practice, strength grows with sustained effort, endurance builds over time. Idle tracts atrophy. This is connective tissue, not neural substrate.

Architecture

FiberBundle ("left_hand", "gaze", "vocal_tract", ...)
  +-- Vec<FiberTract>
        |-- MotorSkeletal   [Signal]  brain -> body (amplify)
        |-- Proprioceptive  [i32]     body -> brain (position)
        |-- NociceptiveFast [i32]     body -> brain (sharp pain)
        |-- Interoceptive   [i32]     body -> brain (fatigue)
        +-- ...

Fiber Types

Physical Properties (all u8, no floats)

Weber-Law Quantization

Sensory signals are quantized using Weber's law -- just-noticeable difference is proportional to stimulus magnitude:

Transmission Pipeline

Motor (brain -> body):

Signal -> Gain -> Conductivity loss -> Fatigue -> Jitter -> Recruitment -> Elasticity -> Output

Sensory (body -> brain):

Raw i32 -> Weber quantize -> Gain -> Conductivity -> Fatigue -> Jitter -> Sensitivity -> Elasticity -> Output

Body Profiles

Pre-configured LimbProfile presets for biologically appropriate fiber bundles:

Usage

use fibertract::{FiberBundle, FiberTract, FiberTractKind, LimbProfile};

// Build a hand from preset profile
let profile = LimbProfile::hand("left");
let mut hand = profile.build();

// Chemical modulation (affects entire bundle)
hand.apply_adrenaline(200);  // fight-or-flight: boost motor, gate pain

// Adapt over time (call each tick)
use fibertract::adapt::{adapt_bundle, AdaptationConfig};
let config = AdaptationConfig::default();
adapt_bundle(&mut hand, &config);

Pain System

Pain events carry perceived intensity (after chemical gating), onset sharpness, duration, and habituation state. The brain uses salience scoring to allocate attention:

use fibertract::{PainEvent, PainSource};

let event = PainEvent {
    bundle_name: "left_hand".into(),
    source: PainSource::Sharp,
    intensity: 200,
    onset: 220,
    duration_ticks: 5,
    habituating: false,
};

assert!(event.is_urgent());      // high intensity + high urgency source
assert_eq!(event.salience(), _);  // weighted score for attention

Features

• serde (default) -- enables serialization for all types via serde + serde_json

Dependencies

• ternary-signal -- the Signal type ({polarity, magnitude})
• log -- structured logging
• rand -- jitter noise generation

License

MIT OR Apache-2.0