Fig. 3: Modeling the ionic basis of uniform intrinsic properties.
From: Intrinsic neuronal properties represent song and error in zebra finch vocal learning
a Exemplar data for four neurons where manual adjustment of the five maximal conductances achieved a good fit between the corresponding biological recording (blue traces) and the model simulation (black traces). Model fits were judged based on the response to current pulses of 100 pA (left panel). Good fits resulted in good predictions in response to current injections not used in the fitting process, for example consider the predictions to step currents of 150 pA magnitude (middle panel), and chaotic current stimuli (right panel; chaotic stimuli not shown to maintain clarity). b An exemplar error manifold (see text) showing two parameters (gSK, gNa) of the five examined in an exhaustive parameter search. The manually determined fit (blue point) resides in the same concavity and near to the global minima. c A scatter plot of all 151 manually modeled neurons showing the two conductances that varied the most across the population, gNa and gSK. Different animals (N = 38), denoted by different color-symbols, occupy different regions in the 2D space. The dashed line encloses the 2D conductance space of the modeled data set, representing an estimate of the entire parameter space occupied by the population of normal adult zebra finches. The same dashed line, over the same scales, appears in Fig. 5c, 5f and 7c.
