MesoNet: minimal neuron experiments

tiny neuron tests 🙂 easier to follow one trace than the whole network. check which declared state variables actually appear in the equation

neuron-models/logistic_test.py: a single dimensionless neuron uses dv/dt = (0.1*v*(I-v))/tau, with tau=1 ms, Euler integration and initial I=2. Excitatory input increments v by 0.1 and sets I=2; inhibitory input decrements v by 0.1 and sets I=0. The script records v for 13 ms. The extra state q is changed by inputs but is not used in the displayed differential equation.

neuron-models/summation_test.py: a separate prototype declares v, u and v0, sets a 1 ms clock, then runs u += v0; v += u*dt in update_u(). Positive events arrive at 1–5 ms and negative events at 6–10 ms; StateMonitor records the response for 13 ms.

ANN and SNN neuron schematic

Neuron-response experiment

use: these scripts isolate the state-update behavior before embedding it in a large network. They are not interchangeable with the conductance-based LIF baseline.

review questions: does the summation model intentionally integrate a velocity-like state twice, and is v0 an increment per tick or a time derivative requiring a timestep factor? Test timestep changes explicitly. For the logistic model, distinguish continuous state response from a spike-generation rule; the shown experiment does not define a complete classification network.