Simulated traces were generated in NEURON 7.3 with Python 2.7 as interpreter (Hines et al., 2009 (link)). To generate the action potential waveform used for the validation of the slope algorithm, a small current injection (100 pA for 2 s) was injected into a single compartment with a specific membrane capacitance of Cm = 1 μF cm−2, a leak conductance of 0.1 mS cm−2 and active sodium (35 mS cm−2) and potassium (9 mS cm−2) peak conductances, as described by Wang and Buzsáki (1996 (link)).
To validate event detection algorithms, excitatory postsynaptic currents (EPSCs) were generated in a ball-and-stick model with a somatic diameter and length of 20 μm, a dendritic length of 500 μm and a dendritic diameter of 5 μm. Specific membrane capacitance Cm was 1 μF cm−2, specific membrane resistance Rm was 25 kΩ cm2, and specific axial resistivity Ra was 150 Ω cm. Excitatory synaptic conductance changes had a bi-exponential time course with τonset = 0.2 ms, τdecay = 2.5 ms, a peak amplitude of 1 nS and a reversal potential of 0 mV. Dendritic locations of synaptic conductance changes were distributed on the dendrite according to a normal distribution with a center at 400 μm distance from the soma and a standard deviation of 12 μm. Time constants and amplitudes of synaptic conductance changes were varied by multiplying with a random number drawn from a normal distribution with mean 1 and standard deviation 0.3 for time constants and 0.1 for amplitudes. Onset times of synaptic conductance changes were simulated as a Poisson process to yield a mean EPSC frequency of 5 Hz as described by Schmidt-Hieber and Häusser (2013 (link)).
Free full text: Click here