ASME-speller: 30-class Auditory BCI EEG Data and Stimuli

64-channel EEG recordings with 2 EOG channels collected during experiments for the ASME-speller, a 30-class auditory brain-computer interface. The dataset includes the auditory stimuli used in the paradigm and was contributed by Simon Kojima to Harvard Dataverse, with a last update recorded on 2026-05-05.

Use Cases

• Train machine learning models for BCI classification based on EEG signals from 64 channels.
• Analyze neural correlates of auditory stream segregation mentioned in the paradigm.
• Benchmark new auditory BCI spellers against the QWERTY layout-based ASME-speller.
• Study the relationship between EOG artifacts and cognitive tasks during BCI use.

Strengths

• Includes 64-channel EEG data, which provides high spatial resolution for neural signal analysis.
• Contains the specific auditory stimuli used, allowing for exact experimental replication.
• Data supports a 30-class BCI speller, indicating a relatively complex classification task.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale model training.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Harvard Dataverse

Collection Method

Collected during experiments for the ASME-speller research.

Freshness

Last updated 2026-05-05 14:43:19; freshness should be verified.