Multimodal Tools for Cognitive and Neuroergonomics Research

The Lab Streaming Layer (LSL) is a system for the unified collection of measurement time series in research experiments that handles both the networking, time-synchronization, real-time access as well as the centralized collection, viewing and disk recording of the data (via the Extensible Data Format).

 

EEGNet is a collection of Convolutional Neural Network (CNN) models for EEG signal processing and classification, written in Keras and Tensorflow. The aim of this project is to provide a set of well-validated CNN models for EEG signal processing and classification, facilitate reproducible research and enable other researchers to use and compare these models as easy as possible on their data.

 

BLASST

BLASST (Band Limited Atomic Sampling With Spectral Tuning) uses a sophisticated algorithm based on matching pursuit with Gabor atoms to remove statistically anomalous spectral power in a narrow band of frequencies.  The algorithm uses an iterative approach to set thresholds based on convergence criterion derived from surrounding frequencies.

Blinker

Blinker automatically extracts blinks from EEG channels, ICs or EOG signals. Blinker identifies individual blinks in signals, identifies which signals have the highest quality blinks, inserts event markers for blinks and computes blink properties.

DETECT

Detect is a toolbox that allows users to quickly train a classifier based on multiple signal classes (usually artifacts). Detect then applies the classifier to unlabeled data. Detect has visualization tools that allow users to scroll through the data and adjust boundaries of identified artifacts. The resulting labeled data can be used to quickly perform manual artifact identification or to check results from automated methods.

EEG-Annotate

EEG-Annotate is an integrated event-identification toolbox that uses machine learning to train classifiers based on labeled event data and then to locate potential similar events in continuous EEG signals. EEG-Annotate employs notions of timing slack and various adaptive thresholding as well as sophisticated transfer learning classifiers.  EEG-Annotate includes tools for training, labeling, analyzing, and visualizing results.

EEGNet

EEGNet is a collection of Convolutional Neural Network (CNN) models for EEG signal processing and classification, written in Keras and Tensorflow. The aim of this project is to provide a set of well-validated CNN models for EEG signal processing and classification, facilitate reproducible research and enable other researchers to use and compare these models as easy as possible on their data.

HED Validator

HED (Hierarchical Event Descriptors) is a platform/vocabulary for annotating events in data. Researchers use tools available on various platforms to annotate events. Tools are available to validate the HED tags and to perform analysis based on HED tags.  For example, researchers might want to extra all epochs in a repository that are time-locked to target events. Considerable effort was expended to annotate SANDR with HED.

Lab Streaming Layer

The Lab Streaming Layer (LSL) is a system for the unified collection of measurement time series in research experiments that handles both the networking, time-synchronization, real-time access as well as the centralized collection, viewing and disk recording of the data (via the Extensible Data Format).

PREP Pipeline

PREP does automated early-stage preprocessing of EEG.  Specifically, PREP removes line noise, computes a robust (artifact-independent) average reference, and identifies bad channels. PREP also produces an extensive report about signal quality.

Spindler

Spindler is an automated tool for detecting oscillatory narrow-band signal bursts in EEG and other physiological measurements. Spindler has been applied to measure alpha spindles in driving data and sleep spindles in overnight recordings.  Spindler uses the geometry of parameter curves that it constructs on the fly to select thresholds appropriate for the analysis.