connectome_manipulator.connectome_comparison.properties

Module for comparing connectomes based on synapse properties:

Structural comparison of two connectomes in terms of synapse properties per pathway, as specified by the config. For each connectome, the underlying properties maps are computed by the compute() function and will be saved to a data file first. The individual synapse properties maps, together with a difference map between the two connectomes, are then plotted by means of the plot() function.

connectome_manipulator.connectome_comparison.properties.compute(circuit, fct='np.mean', group_by=None, sel_src=None, sel_dest=None, per_conn=False, skip_empty_groups=False, edges_popul_name=None, **_)[source]

Computes a matrix of synapse property values between groups of neurons of a given circuit’s connectome.

Parameters:

  • circuit (bluepysnap.Circuit) – Input circuit

  • fct (str) – Function to apply, e.g., “np.mean”, “np.std”

  • group_by (str) – Neuron property name based on which to group connections, e.g., “synapse_class”, “layer”, or “mtype”; if omitted, the overall average is computed

  • sel_src (str/list-like/dict) – Source (pre-synaptic) neuron selection

  • sel_dest (str/list-like/dict) – Target (post-synaptic) neuron selection

  • per_conn (bool) – If selected, fct is applied to the average property value per connection (i.e., average value of all synapses belonging to a connection); otherwise, fct is applied to the synapses of all connections altogether

  • skip_empty_groups (bool) – If selected, only group property values that exist within the given source/target selection are kept; otherwise, all group property values, even if not present in the given source/target selection, will be included

  • edges_popul_name (str) – Name of SONATA egdes population to extract data from

Returns:

Dictionary containing the computed data elements; see Notes

Return type:

dict

Note

The returned dictionary contains the data elements that can be selected for plotting through the structural comparison configuration file, together with a common dictionary containing additional information. Each data element is a dictionary with “data” (numpy.ndarray of size <source-group-size x target-group-size>), “name” (str), and “unit” (str) items. Names of these data elements correspond to the synapse properties that are present in the given SONATA edges population. Usual properties may include for example:

  • “conductance”: Peak conductance

  • “decay_time”: Decay time constant

  • “depression_time”: Time constant for recovery from depression

  • “facilitation_time”: Time constant for recovery from facilitation

  • “u_syn”: Utilization of synaptic efficacy

  • …

connectome_manipulator.connectome_comparison.properties.plot(res_dict, common_dict, fig_title=None, vmin=None, vmax=None, isdiff=False, group_by=None, **_)[source]

Plots a properties matrix or a difference matrix.

Parameters:

  • res_dict (dict) – Results dictionary, containing selected data for plotting; must contain a “data” item with a properties matrix of type numpy.ndarray of size <#source-group-values x #target-group-values>, as well as “name” and “unit” items containing strings.

  • common_dict (dict) – Common dictionary, containing additional information; must contain “src_group_values” and “tgt_group_values” items containing lists of source/target values of the grouped property, matching the size of the properties matrix in res_dict

  • fig_title (str) – Optional figure title

  • vmin (float) – Minimum plot range

  • vmax (float) – Maximum plot range

  • isdiff (bool) – Flag indicating that res_dict contains a difference matrix; in this case, a symmetric plot range is required and a divergent colormap will be used

  • group_by (str) – Neuron property name based on which to group connections, e.g., “synapse_class”, “layer”, or “mtype”; if omitted, the overall average is computed