Scenario 4: End-to-End Research Workflow

Complete the full research workflow—from data loading to analysis and visualization—via the interactive ASA GUI (Attractor Structure Analyzer), without needing in-depth knowledge of model implementation details.

Note

It is recommended to prioritize using the ASA GUI. The ASA TUI is an earlier version and will no longer be maintained; it is provided only for transitional use.

Tutorials

Research Pipeline

Overview

This scenario demonstrates an end-to-end analysis pipeline based on both the ASA GUI and ASA TUI, targeting experimental neuroscientists and researchers. It provides a graphical interface to perform preprocessing, TDA, decoding, and result browsing.

Tutorial 1: End-to-End Analysis with ASA GUI

  • Complete preprocessing and analysis via a PySide6 graphical interface

  • Supports TDA / CohoMap / PathCompare / CohoSpace / FR / FRM / GridScore

  • Preview results in dedicated tabs and quickly open the output directory

Tutorial 2: End-to-End Analysis with ASA TUI (Legacy)

  • Interactive interface for data preparation, preprocessing, analysis, and result export

  • Supports dual input modes: ASA .npz and Neuron + Trajectory

  • Built-in support for TDA / CohoMap / PathCompare / CohoSpace / FR / FRM / GridScore

  • Results are automatically archived with logs and previews provided

Tutorial 3: Model Gallery TUI

  • Replicates the 5×3 analysis layout from canns-experiments/figure2

  • One-click generation of canonical visualizations for CANN1D / CANN2D / GridCell

  • Unified management of result previews and output directories

Who Should Use This Pipeline?

Highly Suitable For:

  • Experimental neuroscientists without extensive coding expertise

  • Rapid prototyping and exploratory analysis

  • Standardized processing of multiple datasets

  • Generating publication-quality figures

  • Teaching and demonstrations

Consider Manual Approaches When:

  • Implementing non-standard model architectures

  • Developing novel analysis methods

  • Requiring fine-grained control over each step

  • Extending pipeline functionality

Learning Path

Quick Start:

  1. Prepare ASA or Neuron + Trajectory data

  2. Launch ASA GUI and select input files

  3. Run with default preprocessing and analysis parameters

  4. Inspect outputs in the Results directory and adjust parameters as needed

Advanced Usage:

  • Batch process multiple sessions (by switching working directories and input files)

  • Fine-tune parameters for TDA, decoding, and visualization

  • Feed generated intermediate results into custom analyses

  • Integrate with existing experimental workflows

Prerequisites

  • Basic Python knowledge

  • Familiarity with your experimental data format (spike/x/y/t)

  • Ability to run commands in a terminal

Estimated Time

  • Tutorial 1: 30–40 minutes

  • Setting up with your own data: 15–30 minutes

  • Total: 70 minutes

Pipeline Features

The ASA GUI provides:

  • Interactive Workflow — GUI-based preprocessing and analysis

  • Automatic Data Validation — Checks input format and missing fields

  • TDA + Decoding — Persistent homology, phase decoding, and comparison

  • Visualization Suite — CohoMap / CohoSpace / FR / FRM / GridScore

  • Result Archiving — Automatic per-dataset directory organization

  • Logging and Caching — Execution logs and stage-wise cache reuse

Data Input Formats

Input support differs between GUI and TUI:

  • ASA GUI: only supports ASA .npz (spike / t; optional x / y)

  • ASA TUI (legacy): supports both ASA .npz and Neuron + Traj .npz

Next Steps

After completing this scenario:

  • Apply ASA GUI to your real experimental data

  • Build custom analyses on top of generated intermediate results

  • Extend functionality by referencing canns.pipeline.asa_gui (GUI) or canns.pipeline.asa (legacy TUI)

  • Contribute new analysis modules to the library