Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2024
|
| Online Access: | https://doi.org/10.5281/zenodo.13363443 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866902290148884480 |
|---|---|
| author | Morales-Cervantes, Antony Herrera, Victor Zamora-Mendoza, Blanca Nohemí Flores-Ramírez, Rogelio Guevara, Edgar |
| author_facet | Morales-Cervantes, Antony Herrera, Victor Zamora-Mendoza, Blanca Nohemí Flores-Ramírez, Rogelio Guevara, Edgar |
| contents | <p> </p> <p>This repository contains the complete dataset and Python scripts used for analyzing functional Near-Infrared Spectroscopy (fNIRS) data to identify neural correlates of long COVID. The analysis integrates advanced machine learning techniques to classify patients as either long COVID or control based on fNIRS data. The repository includes implementations of four machine learning models: K-Nearest Neighbors (KNN), Random Forest, Support Vector Machine (SVM), and XGBoost.</p> <h2>Contents</h2> <ul> <li><strong>Dataset</strong>: The fNIRS data used in this study, including preprocessed features extracted from time series data for each patient.</li> <li><strong>Scripts</strong>: Python scripts that perform data preprocessing, apply the Synthetic Minority Over-sampling Technique (SMOTE) for class balancing, and train the machine learning models. Each script is designed to handle the cross-validation process, model training, evaluation, and generation of performance metrics.</li> <li><strong>Models</strong>: Implementations of KNN, Random Forest, SVM, and XGBoost, optimized for classifying long COVID patients.</li> <li><strong>Results</strong>: Detailed outputs including accuracy, sensitivity, specificity, PPV, and NPV for each model, across multiple iterations of cross-validation.</li> </ul> <h2>Methodology</h2> <p>The data were preprocessed to remove noise and artifacts, and then statistical features were extracted from the fNIRS time series data. SMOTE was applied to address class imbalance. The dataset was split into training and testing sets, ensuring representativity of both classes in each iteration. The models were trained using a 5-fold cross-validation approach to ensure robust performance.</p> <h2>Purpose</h2> <p>This project aims to enhance the understanding of long COVID's neurological impacts by leveraging portable neuroimaging techniques combined with machine learning. The results provide insights into the most effective models for classifying long COVID patients, potentially aiding in better diagnosis and management of the condition.</p> <h2>How to Use</h2> <ol> <li><strong>Clone the Repository</strong>: Download the repository to your local machine.</li> <li><strong>Install Dependencies</strong>: Ensure you have Python and necessary libraries installed as listed in <code>README</code> from data_processing file.</li> <li><strong>Run the Scripts</strong>: Use the provided scripts to preprocess the data, train the models, and evaluate their performance.</li> <li><strong>Explore the Results</strong>: Analyze the outputs to understand the performance of each model and the implications for long COVID classification.</li> </ol> <h2>Data availability statement</h2> <p>The dataset that contains functional near-infrared spectroscopy (fNIRS) data from seventeen patients, twelve of whom are healthy, is available via GitHub: <a href="https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis">https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis</a>. The data is located in the <code>Long_COVID_fNIRS_Raw_Data</code> folder and includes preprocessed features extracted from time series data for each patient.</p> <p>The code to analyze fNIRS signals with machine learning to assist in identifying neural correlates of long COVID is available on GitHub: <a href="https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis">https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis</a>. The code is located in the <code>data_processing</code> folder. This repository contains all necessary scripts for data preprocessing, model training, and evaluation, as well as performance metrics and results.</p> <h2>Ethics statement</h2> <p>The studies were conducted on human participants who signed informed consent according to the Declaration of Helsinki (with registration number 77-21) at the Neurology Department of the Central Hospital "Dr. Ignacio Morones Prieto" in Mexico from October 2021 to October 2022.</p> <h2>Authors' contributions</h2> <p>Edgar Guevara contributed to data collection. All authors were involved in statistical analyses, data interpretation, and planning and designing the study. All authors approved the publication of the data.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_13363443 |
| institution | Zenodo |
| language | |
| publishDate | 2024 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Long-COVID-fNIRS Data Processing and Model Analysis Morales-Cervantes, Antony Herrera, Victor Zamora-Mendoza, Blanca Nohemí Flores-Ramírez, Rogelio Guevara, Edgar <p> </p> <p>This repository contains the complete dataset and Python scripts used for analyzing functional Near-Infrared Spectroscopy (fNIRS) data to identify neural correlates of long COVID. The analysis integrates advanced machine learning techniques to classify patients as either long COVID or control based on fNIRS data. The repository includes implementations of four machine learning models: K-Nearest Neighbors (KNN), Random Forest, Support Vector Machine (SVM), and XGBoost.</p> <h2>Contents</h2> <ul> <li><strong>Dataset</strong>: The fNIRS data used in this study, including preprocessed features extracted from time series data for each patient.</li> <li><strong>Scripts</strong>: Python scripts that perform data preprocessing, apply the Synthetic Minority Over-sampling Technique (SMOTE) for class balancing, and train the machine learning models. Each script is designed to handle the cross-validation process, model training, evaluation, and generation of performance metrics.</li> <li><strong>Models</strong>: Implementations of KNN, Random Forest, SVM, and XGBoost, optimized for classifying long COVID patients.</li> <li><strong>Results</strong>: Detailed outputs including accuracy, sensitivity, specificity, PPV, and NPV for each model, across multiple iterations of cross-validation.</li> </ul> <h2>Methodology</h2> <p>The data were preprocessed to remove noise and artifacts, and then statistical features were extracted from the fNIRS time series data. SMOTE was applied to address class imbalance. The dataset was split into training and testing sets, ensuring representativity of both classes in each iteration. The models were trained using a 5-fold cross-validation approach to ensure robust performance.</p> <h2>Purpose</h2> <p>This project aims to enhance the understanding of long COVID's neurological impacts by leveraging portable neuroimaging techniques combined with machine learning. The results provide insights into the most effective models for classifying long COVID patients, potentially aiding in better diagnosis and management of the condition.</p> <h2>How to Use</h2> <ol> <li><strong>Clone the Repository</strong>: Download the repository to your local machine.</li> <li><strong>Install Dependencies</strong>: Ensure you have Python and necessary libraries installed as listed in <code>README</code> from data_processing file.</li> <li><strong>Run the Scripts</strong>: Use the provided scripts to preprocess the data, train the models, and evaluate their performance.</li> <li><strong>Explore the Results</strong>: Analyze the outputs to understand the performance of each model and the implications for long COVID classification.</li> </ol> <h2>Data availability statement</h2> <p>The dataset that contains functional near-infrared spectroscopy (fNIRS) data from seventeen patients, twelve of whom are healthy, is available via GitHub: <a href="https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis">https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis</a>. The data is located in the <code>Long_COVID_fNIRS_Raw_Data</code> folder and includes preprocessed features extracted from time series data for each patient.</p> <p>The code to analyze fNIRS signals with machine learning to assist in identifying neural correlates of long COVID is available on GitHub: <a href="https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis">https://github.com/morales-cervantes/Long-COVID-fNIRS-Analysis</a>. The code is located in the <code>data_processing</code> folder. This repository contains all necessary scripts for data preprocessing, model training, and evaluation, as well as performance metrics and results.</p> <h2>Ethics statement</h2> <p>The studies were conducted on human participants who signed informed consent according to the Declaration of Helsinki (with registration number 77-21) at the Neurology Department of the Central Hospital "Dr. Ignacio Morones Prieto" in Mexico from October 2021 to October 2022.</p> <h2>Authors' contributions</h2> <p>Edgar Guevara contributed to data collection. All authors were involved in statistical analyses, data interpretation, and planning and designing the study. All authors approved the publication of the data.</p> |
| title | Long-COVID-fNIRS Data Processing and Model Analysis |
| url | https://doi.org/10.5281/zenodo.13363443 |