Skip to content

Analysis and plugins

Analyses run on a selection in any trace, spike or data widget, so they work the same on simulated and recorded data.

Run an analysis

  1. Drag to select a neuron × time region.
  2. Right-click it and choose an analysis.

Built-in: Histogram, Histogram 2D, Spike count per neuron, Pearson correlation. Installed plugins are listed below them.

Open analysis in a new window (checkbox at the top of the menu, on by default) opens each result in its own window. Unticked, results go to the Custom Analysis panel (View → Show Custom Analysis), where you can select again and run a follow-up analysis.

Bundled plugins

Plugin Purpose
Mean activity / Advanced: Mean activity Mean trace (or spike activity) of the selection; the advanced one has controls.
EPSP amplitudes Detects EPSPs and plots their amplitudes (also used by the STDP routines).
Estimate membrane parameters Passive membrane parameters from voltage and injected current.
Replay MSE (cycle vs reference) MSE and correlation of replay cycles against the first selected cycle.
ELiSe target MSE / ELiSe learning curves ELiSe model metrics.

For the ELISE Sequence Learning Network example, the input editor toolbar also shows Analyse ELiSe Replay (target MSE) (replay MSE and correlation of firing_rate_r against the teaching pattern from opsin_exc, saved as .csv) and Run ELiSe Protocol (learning curves). They are enabled once their inputs exist.

Plugin folders

Plugins are discovered from:

  • src/neuroseq/resources/analysis_plugins/ (bundled), and
  • a per-user folder: <user data dir>/neuroseq/analysis_plugins (e.g. ~/.local/share/neuroseq/analysis_plugins on Linux), or the folder in NEUROSEQ_USER_PLUGIN_DIR. The chat assistant writes plugins here after you confirm them.

Every *.py not starting with _ is a plugin. A user plugin whose file name or PLUGIN_ID matches a bundled one is skipped with a warning. New or edited plugins appear the next time the menu opens, without a restart.

Write a plugin

A plugin module defines:

  • PLUGIN_ID: stable internal identifier (used in logs),
  • PLUGIN_NAME: menu label,
  • PLUGIN_ICON: optional QtAwesome icon (e.g. mdi6.chart-line),
  • render_analysis(widget, data, neuron_slice, time_slice): draws the result.

Details: Developer Guide: Analysis plugins.