An Open, Diverse and Multi-Purpose P300 Brain-Computer Interface Dataset

The P300-based brain-computer interface (BCI) is one of the most widely researched non-invasive BCI. We present an open P300 BCI speller dataset that provides data in an enriched and standardised format, with study metadata, demographic data (if available) and BCI data elements that align with developing IEEE P2731 Working Group standard for BCI data. Our BCI dataset is curated from internal data from our prior BCI studies and transformed from a proprietary format to an open, standardised format. Our dataset has a large sample size for big data analysis with 327 participants, including 47 individuals with amyotrophic lateral sclerosis, a target BCI end user population that is under-represented in public BCI datasets. We demonstrate the multi-purpose use of our dataset with results from P300 classification, BCI spelling and BCI error detection.