A Corpus of Joint EEG and Self-Paced Reading of Natural Dutch Texts

Sara Møller Østergaard, Bruno Nicenboim

Keywords: SPR, EEG, corpus

Abstract: We present an open-access corpus that combines electroencephalography (EEG) with self-paced reading (SPR) on eight natural Dutch texts of roughly 800 words, offering word-specific neural and behavioral measures free from parafoveal confounds. Seventy-one participants read seven texts in central SPR and one in rapid serial visual presentation, enabling precise alignment of EEG signals with word onsets. Bayesian hierarchical analyses replicated standard findings: higher word probability and frequency sped up reading times, while lower probability increased N400 amplitude, and, unexpectedly, higher frequency also increased the N400. The resource supports detailed investigation of reading and sentence processing in ecologically valid contexts and is well-suited for computational cognitive modeling of language comprehension. It is freely available on DataverseNL from February 2026 for further research.

Submission Number: 61