Skip to content

Izhikevich 20 Firing Patterns Validation

Validates SCIzhikevichNeuron against the 20 electronic neuron types from Izhikevich (2003) Table 1 and Izhikevich (2004) extended patterns.

Reference

  • Izhikevich, E.M. (2003). Simple model of spiking neurons. IEEE TNN 14(6):1569-72.
  • Izhikevich, E.M. (2004). Which model to use for cortical spiking neurons? IEEE TNN 15(5):1063-70.

Model

v' = 0.04v² + 5v + 140 - u + I
u' = a(bv - u)
if v >= 30 mV: v ← c, u ← u + d

Validated Patterns (Kaggle run 2026-03-28)

Pattern (a, b, c, d) I Spikes/500ms ISI (ms) CV
RS 0.02, 0.2, -65, 8 10 12 45.0 0.123
IB 0.02, 0.2, -55, 4 10 16 31.2
CH 0.02, 0.2, -50, 2 10 37 13.1
FS 0.1, 0.2, -65, 2 10 50 10.0
TC (burst) 0.02, 0.25, -65, 0.05 0 0
TC (tonic) 0.02, 0.25, -65, 0.05 5 46 10.8
RZ 0.1, 0.26, -65, 2 0 1
LTS 0.02, 0.25, -65, 2 10 35 14.5
Tonic Spiking 0.02, 0.2, -65, 6 14 19 26.8
Phasic Spiking 0.02, 0.25, -65, 6 0.5 1
Tonic Bursting 0.02, 0.2, -50, 2 15 58 8.5
Phasic Bursting 0.02, 0.25, -55, 0.05 0.6 1
Mixed Mode 0.02, 0.2, -55, 4 10 16 31.2
SFA 0.01, 0.2, -65, 8 30 20 24.9
Class 1 0.02, -0.1, -55, 6 30 15 33.1
Class 2 0.2, 0.26, -65, 0 5 53 9.5
Spike Latency 0.02, 0.2, -65, 6 7 9 55.6
Subthreshold Osc 0.05, 0.26, -60, 0 2 25 20.3
Accommodation 0.02, 1.0, -55, 4 10 84 6.0
Inhibition-Induced -0.02, -1.0, -60, 8 80 1

Qualitative Checks

  • RS has low CV (0.123 < 0.15) — confirmed regular spiking
  • FS fires faster than RS (50 > 12) — confirmed
  • All suprathreshold patterns produce spikes

Test Files

  • tests/test_izhikevich_20_patterns.py — pytest suite
  • benchmarks/results/phase1_validation_results.json — measured data