connectome_manipulator.model_building.delay

Module for building synaptic delay models

connectome_manipulator.model_building.delay.build(dist_bins, dist_delays_mean, dist_delays_std, dist_delay_min, bin_size_um, **_)[source]

Fits a linear distance-dependent synaptic delay model of type LinDelayModel to the data.

Parameters:

  • dist_bins (numpy.ndarray) – Distance bin edges, as returned by extract()

  • dist_delays_mean (numpy.ndarray) – Delay mean for all bins, as returned by extract()

  • dist_delays_std (numpy.ndarray) – Delay std for all bins, as returned by extract()

  • dist_delay_min (float) – Overall delay minimum, as returned by extract()

  • bin_size_um (float) – Distance bin size in um

Returns:

Fitted linear distance-dependent delay model

Return type:

connectome_manipulator.model_building.model_types.LinDelayModel

connectome_manipulator.model_building.delay.extract(circuit, bin_size_um, max_range_um=None, sel_src=None, sel_dest=None, sample_size=None, edges_popul_name=None, CV_dict=None, **_)[source]

Extracts distance-dependent synaptic delays between samples of neurons.

Parameters:

  • circuit (bluepysnap.Circuit) – Input circuit

  • bin_size_um (float) – Distance bin size in um

  • max_range_um (float) – Maximum distance range in um to consider

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

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

  • sample_size (int) – Size of random subsample of data to extract data from

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

  • CV_dict (dict) – Optional cross-validation dictionary, containing “n_folds” (int), “fold_idx” (int), “training_set” (bool) keys; will be automatically provided by the framework if “CV_folds” are specified

Returns:

Dictionary containing the extracted data elements

Return type:

dict

connectome_manipulator.model_building.delay.plot(out_dir, dist_bins, dist_delays_mean, dist_delays_std, dist_count, model, **_)[source]

Visualizes extracted data vs. actual model output.

Parameters:

  • out_dir (str) – Path to output directory where the results figures will be stored

  • dist_bins (numpy.ndarray) – Distance bin edges, as returned by extract()

  • dist_delays_mean (numpy.ndarray) – Delay mean for all bins, as returned by extract()

  • dist_delays_std (numpy.ndarray) – Delay std for all bins, as returned by extract()

  • dist_count (numpy.ndarray) – Number of data elemets in each bin, as returned by extract()

  • model (connectome_manipulator.model_building.model_types.LinDelayModel) – Fitted linear distance-dependent delay model, as returned by build()