Cardio Mantra

21/03/2025

🫀 Let’s Talk About the Heart – with Cardiologist Kyryll 🩺
In this video, Yuriy connected with his friend and cardiologist, Kyryll, to explore the vital role of the human heart.

They talked about where the heart is located, what it does, and why it's so crucial for our health. Kyryll also explained the main types of heart conditions, including conductivity disorders and ischemic heart disease, and how they affect the body.

It’s an insightful and accessible discussion that brings medical knowledge closer to everyone.

👉 Have questions about heart health? Drop them in the comments – they might be featured in a future episode.

18/01/2025

New enclosures arrived. So new devices are ready for testing.

15/01/2025

Vitalii Samotaiev is here.
Starting a series of articles about physiological signals filtration, applied to ECG signals.
Material will be interesting to cross-field experts, students and enthusiasts aimed in , , , , ,

I'm adding working to make testing easier.
The input data format is

Feel free to ask any questions and don't miss to follow us on Patreon

Get more from Cardio Mantra on Patreon

14/01/2025

Finally, the boards have arrived! I soldered a connector onto one of them for testing, and everything works perfectly. I’ll be able to test the ECG functionality once I assemble the device, but I don’t anticipate any issues there either.

What’s particularly satisfying is that replacing one of the power components with a more modern (and cheaper) alternative reduced power consumption in sleep mode by 2-2.5 times. Under ideal conditions and without accounting for battery self-discharge, this could mean 7-8 months of operation in sleep mode. There’s still work to be done on power consumption in active mode.

14/01/2025

Hello guys!

We are moving forward with hardware and thinking through all the users' aspects of use. We've decided that it is time for charging system. Usually we could use direct charging with electric equipment we have or some chips supported charging, but now the time for USB-C came since now it will be our standard connector for charging.

14/01/2025

Hey, I’m Yuriy! Yuriy Poteshkin
I’m an endocrinologist, healthcare manager, PhD holder, and I run two companies. During my postgraduate studies, I used ECG monitors for my research group, so I’m somewhat familiar with them—though not as much as Kyryll.
Currently, I practice as a doctor, but I don’t have much time for it. I manage a medical analytics business and have a personal PR presence, where I explain how the endocrine system works and how endocrine diseases should be managed on YouTube and TikTok. Since 2015, my company has developed numerous corporate health calculators and wellness systems, and we’ve produced thousands of medical content pieces.

Recently, I helped expand a single medical center into a network, focusing on building small clinics as a project manager. The project was a success, and I was ready for something new. As you can see, I’m a pretty active person, and I’ve been looking for a project where all my skills can be put to good use. This project with my talented friends feels like the perfect fit to create something truly useful, with the potential to help people stay healthy and prevent health issues.

14/01/2025

Hey, I'm Kyryll! Кирилл Рубаник

As a cardiologist with six years of hands-on experience in emergency departments and two years in clinical trials, ECG is at the heart of what I do every day. I’ve also worked extensively with telemedical recognition of heart problems and ECG monitoring. On top of that, I’m a self-taught programmer with a passion for working on microcontrollers. Combining my medical expertise and tech interests, I teamed up with my friends to create CardioMantra—a project focused on decoding the hidden language of the body’s signals.

Welcome to the channel and the community! Together, we’ll dive into understanding your body through the lens of modern technologies.

14/01/2025

Hey, Vitalii Samotaiev is here!

In short, I’m a Cloud Architect with over 13 years of hands-on expertise in AWS Cloud, IoT, Serverless, ML/AI, LLM, PCI, UAVs and hardware dev. Throughout my career, II worked for enterprises that constructed Dallas airport, recovered Pentagon after 911,

optimized code-compilation and infrastructure management tasks with own CI/CD

products for Pixar, Sony, AMD, Equinix, etc and increased by 85% mobile checkout for

biggest e-commerce platforms like PepBoys and 1-800-Flowers, etc.

All of this experience has given me the confidence to envision a product with the potential to make a real social impact and help us build a healthier, better society.

In general, I’ve been part of launching many startups and helped numerous companies stay at the forefront of modern technology. So, bringing together top experts in the field—who also happen to be my closest friends—was a natural next step. I just couldn’t wait any longer.

Together, we started CardioMantra!

Welcome to the channel and welcome to the club! We’re here to share knowledge, help you understand your body, improve your mental and physical health, and guide you into the world of biometric data.

12/01/2025

🤖 Filtering Physiological Signals with a Custom Neural Network 🤖

This script is a comprehensive pipeline for analyzing and processing electrocardiogram (ECG) signals. It has the following functionalities:

Signal Loading and Visualization:

Reads ECG signal data from files using the wfdb library.
Visualizes segments of the ECG signal, with optional annotations showing detected features like heartbeats.

Signal Preprocessing:
Cleans raw ECG signals using techniques such as baseline wander removal and noise filtering (neurokit2).
Resamples signals to a standard frequency (e.g., 125 Hz) for uniformity.

Feature Extraction:

Extracts time-domain, frequency-domain, and morphological features of the ECG using various methods and algorithms (e.g., heart rate variability, R-peak detection, entropy, etc.).
Computes Signal Quality Indices (SQIs) to assess the quality of the ECG signal.

Machine Learning:
Extracted features are used to train machine learning models (e.g., Random Forest) for classifying ECG signals into categories like "normal" or "abnormal."
Benchmarking Methods:

Implements metrics such as accuracy and AUC for evaluating the performance of different feature extraction and signal quality algorithms.

12/01/2025

💓 Detecting Outliers in ECG Data with Advanced Machine Learning 💓

Outliers in ECG signals can indicate irregularities or artifacts that impact analysis and diagnostics. To tackle this challenge, we've taken a comprehensive approach:

📌 Outlier Detection
We learn to systematically identify anomalies in ECG data, ensuring accurate signal interpretation.

📌 Training Three Custom Models
We developed and fine-tuned three distinct machine learning models, each designed to detect and classify outliers effectively.

📌 Comparative Analysis
Results from these models are rigorously compared to identify the most efficient and accurate approach for detecting ECG anomalies.

🧑‍💻 Dive into the details and explore the open-source implementation here: Outlier Detection Challenge – ECG Signals.

This work not only improves the understanding of ECG signal irregularities but also sets a foundation for developing scalable and robust diagnostic tools in healthcare.

12/01/2025

🩺 Understanding and Enhancing Physiological Signal Quality 🩺

Physiological signals, like ECGs, are often affected by various types of noise that can compromise their quality and reliability. To address this, we dive into:

🔍 Identifying the types of noise that impact these signals.
📊 Introducing the concept of noise quality as a measurable factor.
⚙️ Leveraging simple, mathematically derived quality metrics and advanced combinations of these metrics to train a custom machine learning model.
🏷 Using this model to label signals as normal or abnormal for better data classification and analysis.

📖 Learn more and explore the detailed implementation in this open-source notebook: Signal Quality Challenge – ECG Denoising.

This initiative is a step forward in improving the reliability of physiological signal analysis and developing smarter healthcare solutions.