Fast phonetic learning in infants

1. The experiment
The experiment is described in detail in the following publication:
Wanrooij, Boersma & Van Zuijen (2014). Frontiers in Psychology: Language Sciences, 5, article 77.

2. The dataset
This dataset consists of EEG-recordings (44 files in BDF-format) and a table (4 files = the same table in four formats: PDF, TXT, CSV and XLSX).


2.1. The EEG recordings
The 44 files with EEG recordings are EEG recordings of infants (= 2 files for each of the 22 infants). The two files per infant consist of an “a”-file (i.e., the filename ends with “a”) and a “b”-file (i.e., the filename ends with “b”). The “a”-file represents the first half of the recordings; the“b”-file the second half.

The two files per infant show the infant’s EEG during a discrimination test after distributional vowel training. In the discrimination test, the mismatch response (MMR) was measured in an oddball paradigm.


The files were recorded with a Biosemi Active Two system (Biosemi Instrumentation BV, Amsterdam, The Netherlands), and downsampled from 8 kHz to 512 Hz (with Biosemi Decimator 86).

2.2. The table
The table presents the following data per infant participant: the identification number (ID), the age group (infant), the experimental condition as specified by the Distribution Type (unimodal or bimodal) and the Standard Stimulus ([ε] or [æ]), and the results of our measurements as represented in the mean amplitude of the mismatch response (in microvolt).

The ID corresponds to the name of the BDF-files.