ANC language study: A multi-session, multi-echo fMRI dataset for individual level language research

We present the ANC language study (Austrian NeuroCloud language study) dataset, a high-quality, multimodal fMRI dataset using a precision fMRI approach, specifically designed to capture individual variability in the neural mechanisms of word processing. While group-level neuroimaging studies have provided key insights, they often obscure differences at the individual level. To address this, we collected approximately 161 hours of MRI data from 28 participants (23 completed all sessions) across four precision fMRI sessions. Each session included visual and auditory word discrimination tasks targeting phonological, semantic, orthographic, and control conditions. To enhance functional data quality, we employed multi-echo fMRI sequences, improving signal reliability and reducing noise. At the same time, the study emphasizes breadth: it includes both fixation-based and naturalistic movie-viewing resting-state scans, high-resolution T1- and T2-weighted anatomical images, and diffusion-weighted imaging. Beyond neuroimaging, the dataset includes behavioural measures of reading performance, assessments of cognitive ability, a mental health screening, and a pre-session state questionnaire. In this paper, we provide a comprehensive overview of the dataset, including image quality metrics derived from MRIQC for all imaging modalities. We further illustrate the dataset’s utility with first-level activation maps from three participants contrasting visual discrimination tasks with control. The ANC language dataset offers a comprehensive resource for investigating the neural architecture of word processing at the individual level, integrating functional, structural, and connectivity data.