intan.applications package

GUI applications and launchers for EMG analysis and trial labeling.

Lightweight package exports for intan.applications.

Make imports lazy/guarded so importing the package doesn’t pull optional GUI, SciPy, or pandas dependencies during CLI or test runs.

Submodules

intan.processing._emg_viewer_gui

Graphical User Interface (GUI) for loading and visualizing EMG data.

The EMGViewer provides an interactive interface for exploring high-density EMG signals from Intan .rhd recordings. It supports: - Channel selection - Segment visualization - Preprocessing - Trigger-based analysis (if available)

Typically launched via _emg_launcher.py.

save_model_obj(obj, path)

Save a model object with joblib or pickle.

load_model_obj(path)

Load a model object with joblib or pickle.

save_trials_csv(trials, path)

load_trials_csv(path)

save_features_npz(features, path)

load_features_npz(path)

class EMGViewerTk(root)

Bases: object

build_layout()

build_acquisition_tab()

build_trials_tab()

build_filtering_tab()

build_training_tab()

add_feature_directory()

add_training_directory()

apply_visualization_filters(signal)

Filtering for visualization using Filtering Tab controls.

copy_display_to_training_filters()

load_feature_segments()

load_segment_and_visualize()

load_file()

Load an RHD, NPZ, or CSV file and plot the first channel.

update_training_labels_from_directory()

Helper function to refresh the training labels listbox with all the labels detected from the current directories

remove_training_directory()

remove_selected_label()

visualize_selected_segment()

run_pca_visualization()

refresh_labels_list()

Clear and rebuild the labels list from all currently listed directories.

update_training_labels_from_filename(fname)

Extract label from filename and add to label list if not already present. Assumes filenames like: participant_label_0.npz

run_trial_segmentation()

Segment the EMG data into trials based on manual indices or notes files. Saves each segment as a compressed .npz file (with ‘emg’ data and ‘label’) in an ‘emg’ subfolder.

enable_indexing()

scroll_plot(*args)

on_click(event)

save_table()

parse_channel_range(text)

load_table()

delete_selected()

update_channel(event=None)

plot_channel(event=None)

extract_features_from_directories()

Iterate over each selected directory of segmented trials, apply filters, extract features, and aggregate the results. Saves individual feature files and builds a dataset.

extract_td_features(segment)

Computes a variety of unique features from segmented EMG data. Input shape expects (channels x samples).

apply_training_filters(emg)

Applies selected filters to the EMG data. Handles both 1D (single channel) and 2D arrays.

add_scalebars(ax, scale_time=5, scale_voltage=10)

insert_channel_labels(ax, time_vector, num_channels, num_labels=2, font_size=8)

insert_vertical_labels(ax)

build_training_dataset()

waterfall_gui_plot(channel_indices)

on_closing()

EMGViewer

alias of EMGViewerTk

intan.processing._emg_trial_selector

Graphical tool for interactively labeling trial events on EMG recordings.

This GUI allows researchers to: - Load and visualize EMG signals from .rhd files - Select individual channels - Click to mark trial onset points - Assign labels to each indexed event - Append new recordings for multi-session review - Export trial events to a timestamped CSV or TXT file

Clicking the signal while “Set Trial Index” is enabled will add a labeled marker. This tool is useful for supervised training of gesture classifiers, post-hoc annotation, or protocol validation in EMG experiments.

class EMGTrialSelector(root)

Bases: object

Tkinter-based application for manual EMG trial indexing.

emg_data

EMG signal matrix (channels × samples)

Type:

np.ndarray

time_vector

Time vector aligned with EMG samples

Type:

np.ndarray

sampling_rate

Sampling rate of amplifier

Type:

float

current_channel

Channel index currently displayed

Type:

int

indexing_enabled

If True, allows user to click to insert marker

Type:

bool

load_file()

Load EMG data from a .rhd file and initialize the GUI with the first channel.

append_file()

Append EMG data from another .rhd file to the current data.

sample_index_to_timestamp(index)

Convert a sample index to a timestamp string.

Parameters:

index (int) – Sample index to convert.

Returns:

Formatted timestamp string (HH:MM:SS).

Return type:

str

save_table()

Save the trial markers to a text file with sample index and timestamp.

delete_selected()

Delete selected rows from the table.

extract_edges()

clear_indices()

Clear all rows in the table and remove drawn marker lines.

update_channel(event=None)

Update the current channel based on the selection from the dropdown.

enable_indexing()

Enable the indexing mode to allow trial marking on the plot.

on_click(event)

Handle mouse click events on the plot to mark trial onset points.

Parameters:

event (matplotlib.backend_bases.Event) – The mouse event.

plot_channel()

Plot the currently selected EMG channel.

on_closing()

Launch the packaged desktop applications.

launch_emg_viewer()

Launch the PyQt viewer when available, otherwise use the Tk fallback.

launch_emg_trial_selector()

Launch the EMG trial selector GUI.