Graphical user interface (GUI).

A screenshot of the web-based GUI accessed via the Chrome browser. The interface is organized into three main components: the left menu, the main workspace, and the right menu. It supports both light and dark themes.

Dendritic morphology and segmentation.

(A) Example of an L2/3 pyramidal neuron dendritic morphology from a .swc file. The selected section is highlighted in magenta; soma in orange, apical dendrites in blue, basal dendrites in green, and axon in yellow. (B) Detailed representation of the selected section. Top: Diameter of the selected section as a function of the section’s length, with circles marking segment centers. Middle: Equivalent circuit of the selected section shown as an RC circuit assuming a passive membrane. Bottom: Bar plot showing values of a user-selected parameter (surface area, µm2) for each segment. (C) Segmentation network graph representing the same cell as in (A) with d lambda parameters of 0.1 (left) and 0.05 (right); nodes represent segments, colored as in (A). (D) Visualization of the selected morphological parameter on the segmentation graph using a color code. The lasso mouse tool is shown, which allows the selection of specific segments. Statistical morphometric analysis can be performed for the selected part of the cell. (E) Histogram of segment areas for basal (green) and apical (blue) segments. (F) Example of dendritic geometry refinement for a somatostatin expressing (SST) interneuron. Left: Morphology of the SST neuron. Right: Original reconstructed diameter with plausible artifacts of reconstruction (top) vs. simplified diameter (bottom).

Standardization of ion channel models.

(A) Schematic of parsing and standardizing ion channel models from mod files. A CustomIonChannel class is automatically generated and instantiated, facilitating interaction between the GUI and the channel model. This class includes a standardization method that applies an algorithm to produce a StandardIonChannel instance. The standardized channel model can then be saved to a new mod file. (B) Kinetics of a voltage-gated sodium channel. Activation (blue) and inactivation (orange) curves for the steady-state value (left) and the time constant (right). The dashed line represents the original model, while the solid line depicts the model with standardized equations fitted to the original curves. (C) Visualization of GUI sliders for adjusting the five standardized parameters of the activation gating variable m. (D) Corresponding voltage (top) and current (bottom) traces, with the current trace showing sodium current only. Note that the ”missing” last spike can be restored by increasing the injected current amplitude by 0.01 pA.

Distributions of ion channels.

(A) Schematic of segment groups in a toy ball-and-stick model, each with a specific distribution as a function of distance from the soma. (B) Panel for creating a channel group. (C) Distribution parameters adjusted using the group’s widgets. (D) Graphs showing uniform Na channel conductance distribution (left) and a modified graph (right) where Na conductance in the selected region (dashed line) is decreased by 60%. Schematic electrodes indicate recording positions. Morphology from Park et al., 2019 [Park et al., 2019]. (E) Example of an exponential distribution for the HCN (Ih) channels (Inset - original morphology [Poirazi et al., 2003]). (F) Example of a calcium ”hot spot” (red) (Inset - original morphology [Hay et al., 2011]). (G) Top: Sodium-driven backpropagation-activated action potentials (BAPs). A current step of 160 pA is applied at the soma. Bottom: Expanded time scale for the two scenarios in (D), showing failure of BAP spike initiation (blue arrow) in the region with decreased sodium conductance. (H) Top: Distribution of maximal conductance of HCN channels as a function of distance from the soma. Bottom: Voltage sag produced by HCN channels. Current injected at the somatic (−200 pA, 200 ms) and then, after 300 ms, at the dendritic electrode at the distal apical trunk (697 µm from the soma). Dashed trace: blocking HCN channels, modeled as 80% decrease in channel conductance. (I) Top: Distribution of maximal conductance of calcium channels as a function of distance from the soma. Bottom: A dendritic calcium plateau potential triggered by dendritic step current injection (500 pA, 100 ms) at the calcium ”hot-spot”, leading to somatic firing. Somatic traces are shown in orange, dendritic in blue (and gray).

Kinetics and distribution of synapses.

(A) Schematic representation of distributing synaptic inputs. Three central segments of a distal apical branch are selected using the lasso tool, and synapses are added. p — proximal, d — distal. (B) Panel for associating a synapse group with the selected segments. The user must specify the type and number of synapses in a group. (C) The kinetics of synapses and input properties can be further adjusted using the group’s widgets. Note that not all available widgets are shown for visualization purposes. (D) Example responses evoked by activating 20 excitatory synapses placed within one branch as in (A). The regularity of inputs varies from synchronous activation to a random Poisson spike train. Note that the raster plot for input times is accessible in one of the workspace tabs. The examples demonstrate dendritic voltage responses in the presence or absence of NMDA conductances. (E) Experiment similar to Doron et al., 2017 [Doron et al., 2017], demonstrating the effect of inhibiting NMDA spikes. Top: One inhibitory GABAA synapse is placed in the middle of the section and its activation time varies as 0, 10, and 20 ms after excitatory synapse activation. Bottom: The synapse location varies from the most proximal to the most distal segment of the section, with the activation time kept at 20 ms. The same stimulation protocol as in (D) with synchronous activation is used for the excitatory inputs. Scale is the same as in (D). Distributed (F) and clustered (G) allocation of 40 excitatory AMPA-NMDA synapses, similar to an experiment from Poirazi et al., 2003 (Inset - original morphology, [Poirazi et al., 2003]). (H) Distributed synaptic inputs (10 Hz Poisson) nearly fail to evoke somatic action potentials. (I) High firing activity evoked by the same exact synapses clustered within five randomly selected branches.

Morphology reduction.

(A) Original morphology of L5 pyramidal neuron [Hay et al., 2011] and its segmentation graph. (B) Partially reduced morphology using the extended version of neuron_reduce. The extended version allows for the reduction of any selected branch, allowing to retain more apical branches, in contrast to (C). (C) Fully reduced morphology. All stem dendrites (children of the soma) are reduced to a single equivalent cylinder. (D-F) Voltage response of the three models to somatic current injection of 500 pA. Note the difference in the number of somatic APs between the three variations of the model. The partially reduced model (E) more accurately reproduces the response of the original model (D) compared to the fully reduced one (F).

Validation protocols.

Build-in validation protocols applied to the Hay et al., 2011 model [Hay et al., 2011]; See also Table 1). (A) Validation of passive membrane properties. Input resistance (42 MOhm) and membrane time constant (13 ms) measured by applying a step current injection (−50 pA) at the soma. (B) Voltage attenuation for somatic (−500 pA, left) and dendritic (−50 pA, right) step current injection at all bifurcation points along the path from a selected tip segment. (C) Detected somatic action potentials from stimulation with a positive step current (793 pA). (D) Single action potential indicated with an arrow in (C), with measured peak, amplitude, and half-width values. (E) Somatic frequency-current (f-I) curve constructed by applying current steps of increasing amplitude (100 pA step) at the soma. (F) Nonlinear integration of synaptic inputs in a tuft dendrite. Left: Expected vs. actual EPSP amplitude for 1 to 60 synchronously activated AMPA-NMDA synapses. Right: Actual EPSP waveforms. (G) Voltage sag ratio at the soma measured by applying a negative step current injection (−500 pA).

Validation protocols