spycon utilities

spycon utilities#

class spycon.spycon_inference.SpikeConnectivityInference(params: dict = {<class 'str'>: <class 'object'>})#

Bases: object

Base class for spike connectivity inference. It will infer a fully connected graph.

Parameters:

params (dict) – Dictionary with necessary parameters as keywords. Keys should be strings.

infer_connectivity(times: ndarray, ids: ndarray, pairs: ndarray | None = None, parallel: bool = False, **kwargs) → SpikeConnectivityResult#

Full connectivity inference.

Parameters:

  • times (numpy.ndarray) – Spike times in seconds.

  • ids (numpy.ndarray) – Unit IDs corresponding to spike times.

  • pairs (numpy.ndarray) – Array of [pre, post] pair node IDs, for which inference will be done.

  • parallel (bool) – Whether the parallel version is used, if implemented.

Returns:

Connectivity result object.

Return type:

SpikeConnectivityResult

burst_detection(T, C, N=None, ISI_N=0.1)#

Burst detection algorithm. Credits to Adrian Bertagnoli.

[Citation needed]

burst_removal(T, C, N=None, ISI_N=0.1)#
class spycon.spycon_result.SpikeConnectivityResult(method: str, params: dict, nodes: ndarray, all_weights: ndarray, stats: ndarray, threshold: float, runtime: float)#

Bases: object

Class to store the results of a connectivity inference algorithm.

Parameters:

  • method (str) – Name of the method that was used to obtain the result.

  • params (dict) – Dictionary with the corresponding parameters.

  • nodes (numpy.ndarray) – An array of node labels with shape [number_of_nodes].

  • weights (numpy.ndarray) – An array of graded strengths of edges with shape [number_of_edges].

  • stats (numpy.ndarray) – A 2D array representing a fully connected graph with shape [number_of_edges, 3]. The columns are as follows: - The first column represents outgoing nodes. - The second column represents incoming nodes. - The third column contains the statistic used to decide whether it is an edge or not. A higher value indicates that an edge is more probable.

  • threshold (float) – Thresholding the stats, where a connection is defined as those where stats > threshold.

  • runtime (float) – Runtime of the inference algorithm.

set_threshold(new_threshold: float)#

Set threshold for the stats.

Parameters:

threshold (float) – Thresholding the stats, where a connection is defined as those where stats > threshold.

create_nx_graph(graph_type: str = 'binary') → DiGraph#

Create a NetworkX graph from the results.

Parameters:

graph_type (str) – Type of the graph to create: - ‘binary’: Creates an unweighted graph with the inferred connections. - ‘stats’: Creates a fully connected graph, with the decision stats as edge weights. - ‘weighted’: Creates a weighted graph, where weights are the inferred strengths. (Default)

Returns:

A NetworkX directed graph representing the connectivity.

Return type:

nx.DiGraph

draw_graph(graph_type: str = 'binary', ax: Axes | None = None, cax: Axes | None = None)#

Draw a NetworkX graph from the results.

Parameters:

  • graph_type (str) – Type of the graph to create: - ‘binary’: Creates an unweighted graph with the inferred connections. (Default) - ‘stats’: Creates a fully connected graph, with the decision stats as edge weights. - ‘weighted’: Creates a weighted graph, where weights are the inferred strengths.

  • ax (pyplot.Axes, optional) – Matplotlib axis in which the graph should be plotted. Default is None.

  • cax (pyplot.Axes, optional) – Matplotlib axis in which the colorbar should be plotted. Default is None.

Returns:

A NetworkX directed graph representing the connectivity.

Return type:

nx.DiGraph

create_connectivity_matrix(conn_type: str = 'stats') → ndarray#

Create the connectivity matrix from the edges.

Parameters:

conn_type (str) – Type of the connectivity matrix to create: - ‘binary’: Creates a binary connectivity matrix with the inferred connections. (Default) - ‘stats’: Creates a connectivity matrix, with the decision stats as edge weights. - ‘weighted’: Creates a weighted connectivity matrix, where weights are the inferred strengths.

Returns:

The connectivity matrix with shape [num_of_nodes, num_of_nodes].

Return type:

numpy.ndarray

save(name: str, path: str = '')#

Save the results object.

Parameters:

  • name (str) – Name of the result object.

  • path (str, optional) – Path to the saving location. Default is an empty string (‘’).

spycon.spycon_result.load_connectivity_result(name: str, path: str = '') → SpikeConnectivityResult#

Loads a result object.

Parameters:

  • name (str) – Name of the result object.

  • path (str, optional) – Path to where the result object is saved. Default is an empty string (‘’).

Returns:

The loaded result object.

Return type:

SpikeConnectivityResult