Skip to content

About

MATLAB app for gap-in-noise detection: adaptive threshold and psychometric function. In use with patients.

Resources

Stars

0 stars

Watchers

1 watching

Forks

Repository files navigation

GapDetection

A MATLAB application for gap-in-noise detection, a psychoacoustic measure of auditory temporal resolution. It estimates a listener's gap threshold with an adaptive procedure, then measures the full psychometric function around it. Developed at the Hearing Institute (Institut Pasteur, Paris) and used with patients with auditory neuropathy spectrum disorder.

Stimulus design

The test

Each trial presents 800 ms of white noise containing a silent gap at a random position. The listener reports whether they heard a gap; answers and reaction times are recorded.

  1. Rough threshold (adaptive). Up to 30 trials; gap duration decreases after a detection and increases after a miss. The run stops early if the gap exceeds 100 ms, reaches 0 ms, or the same minimum is reached twice.
  2. Psychometric function. Gap durations around the rough threshold are presented in random order, 10 repetitions each (100 trials), independent of previous answers. The output is the proportion of detections as a function of gap duration.

Normal-hearing listeners typically detect gaps of about 4 ms; listeners with auditory nerve disorders often need 20 ms or more.

Stimulus defaults (all adjustable in the GUI)

Parameter Default
Noise duration 800 ms, white noise
Presentation level 65 dB
Sampling rate 48 kHz
Onset/offset ramps (noise) 20 ms, Hanning
Onset/offset ramps (gap) 2 ms
Protected area 30 ms from stimulus and gap edges, so gaps are never truncated
Gap duration range 1–60 ms (psychometric stage)

Ramps avoid spectral splatter (audible clicks) at every transition.

Usage

  1. Open MATLAB (developed and tested on R2023a) in this folder.
  2. Run guiGap.
  3. Select the audio output device and calibrate each ear: the app plays a reference stimulus and asks for the level measured with a sound level meter.
  4. Enter the participant identifier and run the two stages.

Results are saved as CSV, one file per stage: YYYY-MM-DD_HH-MM-SS_<id>_threshold.csv and YYYY-MM-DD_HH-MM-SS_<id>_psycho.csv, with every stimulus parameter, the answer and the response delay for each trial.

Code structure

File Role
guiGap.mlapp Graphical interface
GapPresentation.m Stimulus synthesis and playback
FindThreshold.m Adaptive rough-threshold stage
PlotPsychometric.m Psychometric stage and plot
CalibrationGap.m Per-ear level calibration
AudioTool.m Audio device discovery

Authors

Marta Campi, Grégory Gérenton, Eva Thurot — Hearing Institute, Institut Pasteur, Paris. Contact: gregory.gerenton@pasteur.fr

License

GNU General Public License v3.0 — see LICENSE.

About

MATLAB app for gap-in-noise detection: adaptive threshold and psychometric function. In use with patients.

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages