WiFi CSI Dataset of Patients with Gait-Affecting Neurological Disorders

This webpage provides access to our WiFi Channel State Information (CSI) dataset, which includes 234 sample walks from 114 participants, consisting of patients with five main gait-affecting conditions and healthy individuals. It also provides sample code for denoising and spectrogram generation, along with detailed information on data collection.

If you have any questions or comments regarding this dataset, please contact us at mostofi-lab@ece.ucsb.edu.

Acknowledgments: This work was supported in part by NSF CNS award 2226255 and in part by ONR award N00014-23-1-2715.

Credits and Usage

If you use this dataset for your work, please cite the following paper:

A. Parsay*, M. Torun*, , Philip R. Delio, and Y. Mostofi, "Gait Disorder Assessment Based on a Large-Scale Clinical Trial: WiFi vs. Video vs. Doctor's Visual Inspection," accepted to appear, IEEE Internet of Things Journal, 2025.[paper] (*equal contribution)

Data Collection Details

Collection method: Data were collected at a neurology clinic using two laptops equipped with Intel 5300 WLAN cards (IEEE 802.11n) configured as transceivers. Channel state information (CSI) was then extracted using CSI Tool.

Main categories: The categories represented within this dataset are as follows:

Parkinson's disease (20 participants)
Neuropathy (10 participants)
Post-stroke (16 participants)
Arthritis (hip or knee, 8 participants)
Dementia (14 participants)
Other conditions (8 participants)
Multiple conditions (5 participants)
Healthy (33 participants)

The dataset includes WiFi received signal measurements as each participant walked in the vicinity of the link. It further includes the gait condition, as well as an assessment of gait disorder severity (mild, moderate, or severe).

Data Collection Area: The data were collected in the area illustrated in the figure above (from [1]), located between two medical clinics. The collection was conducted during the regular operational hours of the clinic while the staff and patients were doing their usual activities inside the clinics. However, we ensured that no one else walked or moved in the area during data collection from a participant.

Use Case Example: For an example application of the dataset, please refer to the following paper.

[1] A. Parsay*, M. Torun*, Philip R. Delio, and Y. Mostofi, "Gait Disorder Assessment Based on a Large-Scale Clinical Trial: WiFi vs. Video vs. Doctor's Visual Inspection," accepted to appear, IEEE Internet of Things Journal, 2025.[paper] (*equal contribution)

If you have any questions or comments regarding this dataset, please contact us at mostofi-lab@ece.ucsb.edu.

Download Dataset

Access to the dataset: To request access to the dataset, sample scripts for denoising and spectrogram generation, demographic information, and readme files, please email us at mostofi-lab@ece.ucsb.edu including your affiliation and intended use of the dataset. Upon approval, you will be provided a link to access the materials.

Rights

WiFi CSI Dataset of Patients with Gait-Affecting Neurological Disorders by Alireza Parsay, Mert Torun, Philip R. Delio and Yasamin Mostofi is licensed under Creative Commons Attribution-NonCommercial 4.0 International

Disclaimer: This dataset, its annotations, any related scripts, and all accompanying documents are provided on an "AS IS" basis, without any form of warranty, whether express or implied. We do not offer any assurances regarding the dataset's accuracy, completeness, or reliability. Our liability is limited strictly to what is required by law, with no obligations beyond mandatory statutory duties. Originally compiled for our own research purposes, this dataset and its scripts have not been subjected to rigorous quality checks for suitability in other applications or contexts. Therefore, users should proceed with caution and understand that they are fully responsible for how they choose to use this dataset, including interpreting its content and managing any risks that may arise from its use.

Ethical Compliance Statement: Our Institutional Review Board (IRB) committee has reviewed and approved this research.

Data Usage Agreement: By downloading the dataset, you acknowledge that you have read, understood, and agreed to the Rights and Disclaimer on this webpage. You further affirm that this dataset will be used exclusively for academic research purposes and that you shall not use, distribute, or exploit the dataset for commercial purposes, financial gain, product development, or any other unauthorized activities. Additionally, you agree that you will only share this dataset with individuals or entities that also acknowledge and agree to these same terms, ensuring compliance with the specified restrictions and ethical usage guidelines.