Hume

Hume is an open-source MATLAB program for scoring and analyzing sleep polysomnography and EEG building on the EEGLAB toolbox. Hume allows researchers to unify their sleep EEG workflow across different acquisition systems. If it can be loaded into EEGLAB, Hume can work with it. Hume supports event-marking, artifact-rejection, the generation of sleep statistic reports, inter-scorer reliability comparisons, and storing output in off- and on-line databases (work in progress).

Hume grew out of sleepSMG, developed in partnership with Stephanie Greer PhD. After reaching adolescence it was re-branded Hume after the philosopher; or, perhaps it was for the Elvish word for sleep?

Hume Scoring.png

Key Features

Sleep scoring

Hume is designed for displaying sleep polysomonography and EEG data in a user-friendly flexible manner. Traditional sleep staging is integrated right into the viewing window. Users can define custom viewing montages (scales, colors, gridlines, timebase) to fit their own data and scoring preferences.

Sleep statistics

Traditional commercial software packages lack customizable research-grade sleep statistics, or hide how data are calculated. Hume exports a full slate of traditional sleep statistics including interval analyzes (hours, quarters, thirds, halves, cycles), transition analyzes. See an example report here.

Inter-rater reliability

Consistent scoring requires verifying inter-rater reliability. Hume can compare sleep staging across raters and export a series of statistics (epoch-by-epoch, percent agreement, Cohen's Kappa, latency agreement, REM period agreement).

Event-marking and artifact rejection

Annotate records for manual-detected events (REMs, sleep spindles) and Hume will calculate their occurrence by sleep stage. Whole-page artifact rejection is easily accomplished while scoring using the "Artifact" button (or pressing "x"). These codes travel with the data and can be integrated into EEG post-processing procedures.

Storing and querying data

Hume has preliminary support for SQL databases (either internal or external) and allow for storing of sleep scores across the laboratory and generating cross-study reports. [Work in progress, if interested, ask me!]