You can find the source code for this experiment at GitHub.
- Originally authored by: Theo Veenker
- Adapted by (heavily): Chris van Run
- For clients:
Purpose of this ZEP-based experiment is to see whether an infant participant can detect a difference between two types of auditory stimuli. The infant sits on the caregiver lap facing a wall on which a green light, an invisible speaker and a camera is mounted. On each side wall a red light (or other visual stimuli) and an invisible speaker are mounted. In this implementation there is a familiarization phase and a test phase.
For each trial in the test phase the infant's attention is drawn to one the side lights (blinking). When the infant looks at the blinking light a sequence of sound stimuli starts and a timer runs as long as the infant keeps looking at the blinking light. The trial ends when the infant looks away too long (or when a particular number of stimuli have been presented). In the familiarization phase that precedes the test phase a similar contingency procedure is used but only for the lights; the sound stimuli once started continue until all have been presented.
The researcher indicates a look start by pressing the RETURN key and a look
end by pressing the ESCAPE. This can alternatively be done via the BeexyBox B.
This version of the HPP experiment has been adapted to suit the requirements of the manyababies 1 experiment.
In this implementation the front and side lights will be presented via three computer monitors. Therefore this implementation requires a quad head videocard.
The familiarization-items presentation order is random. The test-items order is pseudorandomized such that there are never more than two of novel-type items (NOVEL) or familiar-type items (FAMIL) that appear in sequence.
Generate output in comma-separated files by running zepdb2extract. See the usage
description on how to use zepdb2exact (zepdb2extract -h).
Output tables:
- Table 1: stimuli details, looking time, not looking time, number of look aways (summary).
- Table 2: looking/no-looking epochs per trial including stimuli details and duration.
For information on running the experiment and extracting the experiment results please go the Zep website at http://www.beexy.nl/zep and check out the documentation section. There you'll also find explanations and instructions that help you understand and modify a Zep experiment.
Config Looking Times. For maximum looking-time settings see test/defs.zm and
familiarization/defs.zm.
Add stimuli. Add the WAV sound files to stimuli/sounds directory and edit the
stimuli lists at stimuli/EXAMPLE.csv
Edit stimuli lists. Every line represents one experiment trial. Each variable
separates by a semicolon (;).
idis a unique reference for human convenience. `trial_typeadult-directed speech (ADS), infant-directed speech (IDS) or training.sndfnis the soundfile that plays (e.g. 'test.wav').sound_directiondictates which direction the sound should come from and should be one of the following values:FRONT_SIDE/LEFT_SIDE/RIGHT_SIDE: self explanatoryLEFT_AND_RIGHT_SIDE: plays both in the left and right channel simultaneously.PSEUDO_RANDOM_SIDE: pseudo-random side with the restriction that one side not allowed more than twice in a row. Note that this does not count manually set sides.
Fine tune your hardware setup. As mentioned above a quad-head setup is
required (four monitors). Window- and head -linking settings is read from
modules/baby_windows3_settings.zm. You can find out the order of heads and
location by running zepdpyinfo in a terminal. Sound settings are in
modules/sound_settings.zm. You can find out the specifics of the sound card(s)
installed by running zepsndinfo in a terminal. This script assumes that the
order of channels are front left, front right, rear/side left, rear/side right.
Optimally, the setup should use a single graphical card that supports at least five monitor outputs. Using more than one graphical card leads to asynchronous visual blanks and minor visual artifacts. A stimulus-presentation computer should have this card installed along with an good-quality sound card. The stimulus-presentation computer should output to five monitors. One to the researcher, three to the participant (i.e. left, middle, right) and one to a video-capture card in a recording computer. The researcher-display signal and the video-capture signal should be duplicates. The recording computer can then apply a chroma-key to the display signal, overlay it on a video feed and save the result for off-line analysis. To get the multi-head setup to function under Kubuntu (Linux/GNU OS) we had to use the Nvidia's baseMosaic feature.
The critical hardware/software used in the UiL-OTS lab is as follows:
- 3 x Nvidia NVS 310 GPUs (though we recommend using a single NVS 810 instead)
- Asus Xonar Dx (sound card)
- BlackMagic Decklink (video-capture card)
- Logitech Webcam C920
- BeexyBox B (response box; alternatively one can use the keyboard)
- Open Broadcaster Studio (software)
- Zep 2.0.9
- Kemler-Nelson, D. G., Jusczyk, P. W., Mandel, D. R., Myers, J., Turk, A. & Gerken, L. (1995). The Head-turn Preference Procedure for testing auditory perception. Infant Behavior and Development 18, 111-116. doi
- KERKHOFF, A., DE BREE, E., DE KLERK, M., & WIJNEN, F. (2013). Non-adjacent dependency learning in infants at familial risk of dyslexia. Journal of Child Language, 40(1), 11-28. doi
This experiment script is released under the terms of the GNU General Public License (see http://www.gnu.org/licenses/gpl-2.0.html). It is distributed in the hope that it will be useful, but with absolutely no warranty. It is your responsibility to carefully study and test the script before using it with real participants.