Data Loading

Overview

Neuralyzer supports loading diverse data types commonly used in systems neuroscience experiments. This section covers all data import methods and formats.

Data can be loaded through:

  1. Data Import Widget - Unified interface for importing all data types via Modules → Data Import
  2. JSON configuration files - Reproducible, scriptable data loading for batch workflows

Using the Data Import Widget

The Data Import Widget provides a unified interface for loading all supported data types. Access it via Modules → Data Import from the main menu.

Features

  • Type Selection: Choose the data type you want to import (Lines, Masks, Points, Analog, Events, Intervals, Tensors)
  • Format Selection: Each data type supports multiple file formats (CSV, HDF5, Binary, LMDB, etc.)
  • Preview: Many loaders show a preview of the data before importing
  • Scaling Options: Apply coordinate scaling for spatial data types
  • State Persistence: The widget remembers your last used directory and format preferences

Workflow

  1. Open the widget via Modules → Data Import
  2. Select your data type from the dropdown
  3. Choose the appropriate file format
  4. Configure format-specific options (columns, datasets, scaling)
  5. Click “Import” to load the data into the DataManager
Passive Type Awareness

When you select a data object in the DataManager, the Data Import Widget automatically switches to show the appropriate import options for that data type.

Media Data

Media provides the visual backdrop for data visualization and annotation.

Type Description Formats Reference
Video Frame-by-frame video files .mp4, .avi, .mov Video Loading
Image Stacks Sequence of image files .tif, .png, .jpg Image Loading
HDF5 Hierarchical data with image arrays .h5, .hdf5, .mat —
Media Import How-To Guides

Spatial Data

Spatial data represents locations and shapes over time, typically overlaid on video.

Type Description Common Sources Reference
Points (x, y) coordinates per frame DeepLabCut, SLEAP, manual annotation Point Loading
Lines Ordered sequences of points Whisker trackers, contour detection Line Loading
Masks Binary regions per frame Segmentation models, ROI tools Mask Loading

Time Series Data

Continuous and discrete signals recorded alongside behavior.

Type Description Common Sources Reference
Analog Time Series Continuous signals Electrophysiology, sensor data —
Digital Events Point events in time Spike times, behavioral markers —
Digital Intervals Time periods (start/end) Trial epochs, stimulus periods JSON Loading

Multi-dimensional Data

Type Description Common Sources Reference
Tensors N-dimensional arrays Neural network outputs, multi-channel data Tensor Loading

JSON Configuration Loading

For reproducible analysis workflows, define all data loading in a JSON file:

[
  {
    "filepath": "experiment.mp4",
    "data_type": "video",
    "name": "media"
  },
  {
    "filepath": "tracking.csv",
    "data_type": "point",
    "name": "nose"
  },
  {
    "filepath": "trials.csv",
    "data_type": "digital_interval",
    "name": "stimulus_on"
  }
]

See JSON Loading Reference for complete documentation on all supported formats and options.

How-To Guides

Step-by-step guides for common data import tasks:

Getting Started

  1. Single video analysis: Use File → Load Data to select your video
  2. Multi-modal experiments: Create a JSON configuration file referencing all your data files
  3. Reproducible workflows: Save your JSON configuration alongside your data for easy re-loading