Skip to content

network-viewer.md

Latest commit

 

History

History
131 lines (105 loc) · 6.12 KB

network-viewer.md

File metadata and controls

131 lines (105 loc) · 6.12 KB
Raw

Network Viewer

The network viewer, at the bottom of the Results tab, interactively explores the functional connectivity network from a completed MEA-NAP run, including optional cell-type overlays. It's the Python equivalent of MATLAB's runMEANAPviewer.m.

Using the network viewer

When the tab first opens the viewer shows a built-in example network so every control is usable immediately — no file needed. To explore your own data:

  1. Click Browse… and select a MEA-NAP output .mat file from the ExperimentMatFiles/ subfolder of an output directory, e.g. OutputData.../ExperimentMatFiles/<recording>_OutputData....mat. (Use example network reloads the synthetic demo at any time.)
  2. The network renders immediately. Recording metadata (name, DIV, group, active node count, edges shown) appears in the left panel.
  3. Adjust Network settings to update the plot in real time:
    • Lag — which functional connectivity lag value to view (e.g. 1000 ms, 2500 ms, 5000 ms).
    • Edge threshold by / Edge threshold — how the minimum edge weight is chosen (see Edge subsampling & thresholding).
    • Node size metric — drive node size by any node-level metric present in the file (defaults to node degree, ND).
    • Node color metric — color nodes by any node-level metric (betweenness centrality, node strength, z-score, ...), or leave as None for flat cyan nodes.
  4. (Optional) Click Load cell types from file… to overlay cell-type information — see below.

Node color uses the viridis colormap, with a colorbar legend on the right.

Display controls

The Display group changes how the network is drawn — none of these alter the underlying metrics, only the appearance:

  • Node layout — where nodes are placed:
    • Original (electrodes) — the physical electrode coordinates (default).
    • Circular, Spring (force), Kamada-Kawai, Spectral, Shell — positions derived from network topology (via networkx). These port MATLAB's getNodeCoords.m layout options and the circular plot type. Every derived layout is rescaled into the electrode bounding box so node, edge and legend sizing stay consistent.
  • Node size scale — a multiplier applied to every node (MATLAB's maxNodeSize); increase it to enlarge all nodes at once.
  • Node scaling — how the size metric maps onto node radius: Linear, Log2, Log10, Square, or Cube (port of MATLAB's nodeScalingMethod in getNodeSize.m).
  • Min edge width / Max edge width — the line width (in points) of the weakest and strongest drawn edges. The viewer keeps max ≥ min automatically.

Edge subsampling & thresholding

Dense networks can render as an unreadable hairball. Two independent controls limit what's drawn (neither changes node degree or any other metric — they are plotting-only, mirroring PlotIndvNetMet.m):

  • Max edges (Display group) — draw only the strongest N edges by absolute weight, keeping the rest hidden (port of limitEdgesForPlotting.m, HighToLow). 0 (shown as Unlimited) draws every edge.
  • Edge threshold by (Network settings) — how the threshold weight is chosen (port of getEdgeThreshold.m):
    • Absolute value — draw edges with weight ≥ the value you set (0–1).
    • Percentile — threshold at the Nth percentile of all adjacency entries, including zeros (MATLAB-faithful: on a sparse matrix a low percentile does nothing, so you typically need a high value).
    • Percentile (nonzero edges) — threshold at the Nth percentile of only the actual edge weights, so e.g. 80 % keeps the strongest ~20 % of edges — the intuitive behaviour.

The Edges shown field updates live so you can see the effect of each control.

Saving the plot

Click Save plot… (Display group) to export the current figure. Choose PNG (raster, saved at 600 dpi for publication-quality output) or SVG (vector, resolution-independent). If you omit the extension it is inferred from the selected file type. The export preserves the on-screen white background and trims surrounding whitespace.

Cell-type overlay

Cell types are rendered as concentric rings on each node, one line style per type, mirroring the MATLAB viewer.

Loading cell types:

  1. Prepare (or locate) a cell-type spreadsheet: each column is one cell type, each cell contains the 1-indexed channel number of a cell belonging to that type. Columns with no cells for a given type should be left blank. The PutativeCellType_*.xlsx files produced alongside MEA-NAP runs use this format.

    NeuN+ PV+ SST+
    68 25 110
    78 42 216
    117

  2. In the Cell types group, click Load cell types from file… and select the .xlsx or .csv file.

  3. A listbox appears with every cell type found — select one or more to filter the displayed network.

Filtering by cell type:

  • Selecting one type shows only nodes of that type.
  • Selecting multiple types shows only nodes belonging to all selected types (intersection — consistent with the MATLAB viewer).
  • Deselecting everything returns to showing all active nodes.

The concentric-circle legend at the bottom of the plot identifies which ring style corresponds to which cell type.

:::{admonition} Note on .mat cell-type data :class: warning MEA-NAP stores Info.CellTypes inside output .mat files as a MATLAB MCOS table object, which scipy.io can't decode. When the viewer detects this, it logs a message and prompts you to load the cell-type spreadsheet directly — the same .xlsx file originally supplied to the MATLAB pipeline. :::

Using the network plotting API from Python

The plotting code underneath this tab is a standalone module, meanap.network_plot, usable independently of the GUI. See the network-plotting notebook tutorial for a runnable, end-to-end walkthrough with real output baked in, or the API reference for the full function/class list.