The ZuCo Benchmark on Cross-Subject Reading Task Classification with EEG and Eye-Tracking Data

Abstract

We present a new machine learning benchmark for reading task classification with the goal of advancing EEG and eye-tracking research at the intersection between computational language processing and cognitive neuroscience. The benchmark task consists of a cross-subject classification to distinguish between two reading paradigms: normal reading and task-specific reading. The data for the benchmark is based on the Zurich Cognitive Language Processing Corpus (ZuCo 2.0), which provides simultaneous eye-tracking and EEG signals from natural reading. The training dataset is publicly available, and we present a newly recorded hidden testset. We provide multiple solid baseline methods for this task and discuss future improvements. We release our code and provide an easy-to-use interface to evaluate new approaches with an accompanying public leaderboard: www.zuco-benchmark.com. HighlightsO_LIWe present a new machine learning benchmark for reading task classification with the goal of advancing EEG and eye-tracking research. C_LIO_LIWe provide an interface to evaluate new approaches with an accompanying public leaderboard. C_LIO_LIThe benchmark task consists of a cross-subject classification to distinguish between two reading paradigms: normal reading and task-specific reading. C_LIO_LIThe data is based on the Zurich Cognitive Language Processing Corpus of simultaneous eye-tracking and EEG signals from natural reading. C_LI