Le Petit Prince multilingual naturalistic fMRI corpus

Area of Science:

• Cognitive Neuroscience
• Neuroimaging
• Psycholinguistics

Background:

• Neuroimaging studies of natural language comprehension traditionally used limited, single-language stimuli.
• This limitation restricted the generalizability of findings across different linguistic backgrounds.
• A need exists for multilingual, naturalistic datasets to explore cross-linguistic variations in language processing.

Purpose of the Study:

• To introduce the Le Petit Prince fMRI Corpus (LPPC-fMRI), a novel multilingual neuroimaging dataset.
• To facilitate research into the cognitive neuroscience of speech and language processing during naturalistic listening.
• To enable cross-linguistic comparisons of neural substrates for language comprehension.

Main Methods:

• Acquisition of multi-echo functional magnetic resonance imaging (fMRI) data from 49 English, 35 Chinese, and 28 French speakers.
• Participants listened to 'The Little Prince' audiobook in their native language.
• Provision of time-aligned speech annotation and word-by-word predictors derived from natural language processing tools.

Main Results:

• The LPPC-fMRI dataset exhibits high-quality timeseries data, characterized by good temporal signal-to-noise ratio.
• High inter-subject correlation was observed, indicating reliable neural responses across participants.
• Data-driven functional analyses confirmed the high quality and utility of the dataset.

Conclusions:

• The LPPC-fMRI is a valuable, annotated, multilingual resource for studying language processing in the brain.
• This dataset supports future research on both universal and language-specific neural mechanisms of language comprehension.
• It allows for detailed analysis across multiple perceptual and linguistic levels of language processing.