Neural Noise

UET175: EEG motor imagery dataset in Vietnamese stroke patients

2 min read

  • UET175 is an open EEG dataset of motor imagery tasks collected from Vietnamese stroke patients.1
  • The dataset supports BCI rehabilitation research and algorithm benchmarking for motor restoration.1
  • Population-specific (stroke) data enables EEG-based BCI and neuroprosthetics pipeline development.1 1

Weekly enrichment (2026-07-20)

  • UET175 was published in Frontiers in Neuroscience in 2025 (19:1580931) and comprises EEG from 30 post-stroke patients recorded at Hospital 175 in Ho Chi Minh City, Vietnam, between October 2024 and January 2025.2
  • The completed dataset contains 220 recording sessions across the 30 subjects, with each session holding at least one recording run.2
  • Signals were acquired with the saline version of the commercial-grade Emotiv EPOC Flex headset at a 128 Hz sampling rate, using 22 of the device’s 32 supported channels, streamed via Lab Streaming Layer in double-precision format.23
  • The cohort spanned ages 43-78 years (mean 63.85, SD 12.85) with a 60% male / 40% female split.2
  • Lesion characteristics skewed toward the right hemisphere, with damage to the corona radiata being the most common finding (present in 26.7% of medical reports); the most frequent comorbidities were hypertension (80%) and type 2 diabetes mellitus (33.3%).2
  • The authors position UET175 as filling a gap: most existing motor-imagery datasets (e.g., BCI Competition sets) use expensive medical-grade equipment, whereas UET175 targets affordable commercial devices to enable a VR-BCI system for stroke rehabilitation.2
  • Data are organized per patient with a JSON metadata file (year of birth, gender, medical conditions, diagnoses, assessments, session count) and per-session runs, each run stored as a raw-signal CSV plus session-setup JSON, action-label text, and event-timestamp text files.3
  • The dataset is openly released on GitHub for reuse in BCI and neuroprosthetics research.3
  • For context, a separate 2025 open dataset from Liu et al. captured lower-limb motor imagery in 27 stroke patients using a 64-channel NeuSen W system at 1000 Hz across longitudinal sessions (4,260 MI trials); UET175 is distinguished by its low-cost consumer device and its specifically Vietnamese post-stroke population.4

Footnotes

  1. https://news.google.com/rss/articles/CBMilAFBVV95cUxNT3Flb0ptajV1VElJaWZlMHRUNVIxSUctT3JqUU5fNV8zcWxpc0g1RDlWQlhoU0ZKQ3NTX0t0VFZWSC1vTjFNRTg5RHBTdlNmNS1GLW1RdzNpX2FxeV9vbUlKX050YnRDSFRFYUl6RFlVbkFsS1E1VF9TV1ozdXdyQzNhT2dwUkNrbkw5cFpzVHY3TmpI?oc=5 ↩ ↩2 ↩3 ↩4

  2. https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2025.1580931/full ↩ ↩2 ↩3 ↩4 ↩5 ↩6

  3. https://github.com/nmk-k66-uet/UET175 ↩ ↩2 ↩3

  4. https://doi.org/10.1038/s41597-025-04618-4 ↩

Graph View