Introduction to Biomedical Engineering (IBME)

Biopotential Instrumentation, Analog Front-End Circuits & Real-Time Embedded ECG Telemetry

University of Tehran  |  Faculty of Electrical and Computer Engineering
Course: Introduction to Biomedical Engineering (IBME)  |  Semester: Fall 2023 (پاییز ۱۴۰۲)
Instructor: Dr. Majid Badiee Rostami (دکتر مجید بدیعی رستمی)
Author: Alireza Najafi Motiei (علیرضا نجفی مطیعی)  |  Student ID: 810100224

---

📋 Table of Contents

• Executive Overview
• System Architecture
• Module 1: Biosignal Processing (VCG & EEG Sleep Staging)
• Module 2: Analog Front-End (AFE) Biopotential Circuits
• Module 3: Real-Time Hardware ECG Acquisition & HRV Telemetry
• Clinical Findings & Empirical Benchmarks
• Repository Organization
• How to Run & Reproduce
• Technical Reports
• Citation
• Author & License

---

🔬 Executive Overview

This repository houses the complete, publication-grade engineering curriculum portfolio for Introduction to Biomedical Engineering (IBME) at the University of Tehran (Department of Electrical and Computer Engineering, Fall 2023), instructed by Dr. Majid Badiee Rostami.

The codebase bridges three fundamental pillars of modern biomedical technology:

1. Theoretical Biosignal Modeling & Decomposition: Derivation of 3D Vectorcardiography (VCG) spatial dipole loops and automated EEG sleep stage classification (NREM 1--3, REM, and Wake) using Fourier spectral power ratios.
2. Analog Front-End (AFE) Microelectronic Design: SPICE/Multisim design and AC frequency analysis of biopotential amplification circuits, including a variable-gain ECG bandpass amplifier (0.1--100 Hz), a high-Q 50 Hz active Twin-T notch filter for powerline hum suppression, and a high-gain (~75,000) low-noise EEG amplifier.
3. Physical Embedded Hardware Telemetry & Real-Time HRV: In-vivo Lead-I biopotential acquisition interfacing an Analog Devices AD8232 front-end integrated circuit with an Arduino microcontroller, streaming real-time telemetry to MATLAB for zero-phase Butterworth filtering, Pan-Tompkins style QRS detection, and Heart Rate Variability (HRV) analysis in both time and frequency domains (Welch PSD).

---

🏛️ System Architecture

                                  PHYSICAL DOMAIN
+---------------------------------------------------------------------------------+
|                                                                                 |
|   Human Subject          Ag/AgCl Electrodes         AD8232 Analog Front-End     |
|  [ Lead-I: RA, LA ] ---> [ Shielded Leads ] ---> [ INA + High-Pass + RLD Stage ]|
|                                                          |                      |
+----------------------------------------------------------|----------------------+
                                                           v Biopotential (0-3.3V)
                                                 +--------------------+
                                                 | Arduino MCU        |
                                                 | 10-Bit ADC (fs~95Hz)|
                                                 | LO+ / LO- Monitors |
                                                 +--------------------+
                                                           |
                                                           v Serial Telemetry (9600 bps)
                                                 +--------------------+
                                                 | MATLAB Processing  |
                                                 +--------------------+
                                                           |
                    +--------------------------------------+--------------------------------------+
                    v                                      v                                      v
          [ Baseline Wander Removal ]            [ QRS & R-Peak Detection ]             [ Spectral HRV Decomposition ]
          Zero-Phase Butterworth (0.5-40 Hz)     Dynamic Refractory Thresholding        Welch Periodogram PSD (VLF/LF/HF)

---

🧠 Module 1: Biosignal Processing (VCG & EEG Sleep Staging)

Located in 01-biosignal-processing-vcg-eeg/.

1. Vectorcardiography (VCG) Loop Synthesis

The electrical activity of the myocardium can be approximated as a time-varying 3D equivalent cardiac dipole: $$�ec{D}(t) = �egin{bmatrix} V_x(t) & V_y(t) & V_z(t) \end{bmatrix}^T$$ Using the Frank lead network, orthogonal biopotentials $V_x$, $V_y$, and $V_z$ are extracted and integrated to reconstruct the continuous frontal, sagittal, and transverse vector loops. The standard clinical 12-lead ECG is mathematically derived via linear transformation projections.

VCG Orthogonal Leads (Time Domain)	3D Cardiac Spatial Vector Loop

2. Polysomnographic EEG Sleep Stage Classification

Continuous EEG brainwave recordings from central (C3/C4) and occipital (O1/O2) channels are analyzed via Discrete Fourier Transform (DFT). Spectral energy is computed across standard neurophysiological bands:

• Delta ($\delta$): $0.5 - 4.0 ext{ Hz}$ — Characteristic of slow-wave sleep (NREM Stage 3).
• Theta ($ heta$): $4.0 - 8.0 ext{ Hz}$ — Prominent during light sleep transition (NREM Stage 1).
• Alpha ($�lpha$): $8.0 - 13.0 ext{ Hz}$ — Characteristic of relaxed wakefulness with closed eyes (Stage 0).
• Beta ($�eta$): $13.0 - 30.0 ext{ Hz}$ — Associated with alertness and Rapid Eye Movement (REM) sleep.

Awake State (Stage 0 Alpha Rhythm)	Light Sleep (NREM Stage 1 Theta Dominance)

---

⚡ Module 2: Analog Front-End (AFE) Biopotential Circuits

Located in 02-analog-front-end-multisim/.

Designed and verified in NI Multisim to evaluate small-signal AC response, gain-bandwidth products, and noise margins for microvolt/millivolt biopotentials.

1. ECG Biopotential Amplifier

• Topological Structure: Input instrumentation buffer stage, active 2nd-order bandpass filter ($0.1 ext{ Hz} - 100 ext{ Hz}$), and an inverting op-amp summing stage.
• Gain Staging: Continuously adjustable gain up to $7{,}000$ ($76.9 ext{ dB}$).
• CMRR: $> 90 ext{ dB}$ to eliminate common-mode noise on the thorax.

Figure: Complete multi-stage ECG instrumentation amplifier schematic in Multisim.

Figure: AC Frequency response (Bode magnitude and phase) demonstrating flat passband between 0.1 Hz and 100 Hz.

2. Active Twin-T 50 Hz Notch Filter

Rejects severe $50 ext{ Hz}$ powerline hum without attenuating adjacent QRS diagnostic energy.

• Topology: Parallel symmetrical low-pass ($R-R-2C$) and high-pass ($C-C-R/2$) T-networks.
• Selectivity ($Q$): Active bootstrapped feedback from an operational amplifier buffer sharpens the notch to achieve $> -40 ext{ dB}$ rejection at exactly $50 ext{ Hz}$.

Figure: Symmetrical active Twin-T 50 Hz notch filter circuit.

3. High-Gain EEG Biopotential Amplifier

EEG signals measured at the scalp have amplitudes of merely $10 - 100\ \mu ext{V}$.

• Architecture: Cascaded three-stage low-noise amplifier with AC coupling to prevent DC polarization saturation from the electrode-skin interface.
• Total Gain: $�pprox 75{,}000$ ($97.5 ext{ dB}$) with a passband up to $200 ext{ Hz}$.

EEG Amplifier Schematic	Frequency Response (Bode Plot)

---

🩺 Module 3: Real-Time Hardware ECG Acquisition & HRV Telemetry

Located in 03-hardware-ecg-acquisition-hrv/.

1. Hardware Interconnection & Telemetry Setup

The experimental hardware platform pairs an Analog Devices AD8232 analog front-end board with an Arduino microcontroller.

AD8232 Pin	Arduino Pin	Signal Type	Description
OUTPUT	A0	Analog Input	Conditioned biopotential signal ($0 - 3.3 ext{ V}$)
3.3V	3.3V	DC Supply	Regulated low-noise power rail
GND	GND	Reference	Common system ground
LO+	Pin 4	Digital Input	Leads-off comparator output: Positive electrode detached
LO-	Pin 7	Digital Input	Leads-off comparator output: Negative electrode detached

Hardware Breadboard Interface	Einthoven Lead-I Electrode Placement

2. Firmware Implementation

• ad8232_ecg_sampler.ino: Samples ADC channel A0 at $f_s �pprox 95 ext{ Hz}$, checks digital lead-off status, and streams data via 9600 bps UART.
• heart_rate_bpm.ino: Embedded moving-window circular buffer with dynamic thresholding to calculate instantaneous BPM with a $300 ext{ ms}$ physiological refractory blanking period.

3. Digital Signal Processing & HRV Analysis Pipeline

• Baseline Wander Removal: 2nd-order zero-phase Butterworth bandpass filter ($0.5 - 40.0 ext{ Hz}$) via filtfilt to prevent phase distortion.
• R-Peak Extraction: Adaptive amplitude thresholding with a $450 ext{ ms}$ blanking window to prevent false triggers on tall T-waves.
• Time-Domain HRV Metrics: Extraction of Normal-to-Normal ($NN$) intervals: $$ ext{SDNN} = \sqrt{rac{1}{N-1}\sum_{i=1}^N (RR_i - \overline{RR})^2}$$ $$ ext{RMSSD} = \sqrt{rac{1}{N-1}\sum_{i=1}^{N-1} (RR_{i+1} - RR_i)^2}$$
• Frequency-Domain Spectral HRV: Welch periodogram power spectral density to quantify the Sympathovagal balance index: $$ ext{LF/HF Ratio} = rac{\int_{0.04}^{0.15} S_{RR}(f),df}{\int_{0.15}^{0.40} S_{RR}(f),df}$$

Raw vs. Filtered ECG Waveform	Welch PSD of R-R Intervals (HRV)

---

📊 Clinical Findings & Empirical Benchmarks

The in-vivo 30-second recording session from healthy subject yielded the following physiological metrics:

Metric	Measured Value	Standard Clinical Reference	Diagnostic Interpretation
Mean Heart Rate	$78.42 ext{ BPM}$	$60 - 100 ext{ BPM}$	Normal resting sinus rhythm
Mean R-R Interval	$765.10 ext{ ms}$	$600 - 1000 ext{ ms}$	Stable cardiac pacing
SDNN (Total Variability)	$42.18 ext{ ms}$	$> 30 ext{ ms}$ (at rest)	Healthy autonomic neurocardiac regulation
RMSSD (Parasympathetic)	$31.45 ext{ ms}$	$20 - 50 ext{ ms}$	Robust vagal tone and parasympathetic activation
pNN50	$18.60%$	$> 3%$	High beat-to-beat variability and vagal responsiveness
LF/HF Ratio	$1.42$	$1.0 - 2.0$	Balanced sympathovagal autonomic homeostasis

---

📁 Repository Organization

Biomedical-Instrumentation-and-Signal-Processing/
├── 01-biosignal-processing-vcg-eeg/        # Biosignal modeling & sleep stage classification
│   ├── IBME_CA1_part1.m                   # VCG 3D loops & 12-lead ECG derivation
│   ├── IBME_CA1_part2_stage0.m            # EEG Awake State (Stage 0) Alpha rhythm analysis
│   ├── IBME_CA1_part2_stage1.m            # EEG NREM Stage 1 Theta band dominance
│   ├── IBME_CA1_part2_stage2.m            # EEG NREM Stage 2 Sleep spindles & K-complexes
│   ├── IBME_CA1_part2_tofind.m            # Unknown patient sleep phase identification
│   ├── IBME_CA1_part2_tofind2.m           # REM vs Deep Sleep slow-wave classification
│   ├── matlab.mat                         # Polysomnographic EEG clinical dataset
│   └── v1.mat ... v6.mat                  # Precordial ECG recording vectors
│
├── 02-analog-front-end-multisim/           # SPICE circuit simulations (.ms14)
│   ├── ECG.ms14                           # Multi-stage ECG instrumentation amplifier
│   ├── EEG.ms14                           # Ultra-high gain (~75,000) low-noise EEG amplifier
│   └── Notch_filter.ms14                  # Active 50 Hz Twin-T notch filter
│
├── 03-hardware-ecg-acquisition-hrv/        # Embedded hardware telemetry & HRV analysis
│   ├── firmware/                          # Microcontroller code
│   │   ├── ad8232_ecg_sampler.ino         # ADC biopotential telemetry with lead-off detection
│   │   └── heart_rate_bpm.ino             # Real-time QRS thresholding & BPM counter
│   ├── matlab/                            # Signal processing scripts
│   │   ├── ecg_live_capture.m             # Serial COM port streaming & CSV logger
│   │   ├── qrs_detect_and_hrv_metrics.m   # Butterworth filter, QRS detection & HRV stats
│   │   └── hrv_spectral_welch_psd.m       # Frequency-domain FFT & Welch periodogram PSD
│   └── data/                              # Recorded biometric datasets
│       └── ecg_recorded_lead1.csv         # Raw in-vivo Lead-I biopotential recording
│
├── docs/media/                            # High-resolution architectural figures
│   ├── circuits/                          # Multisim schematics and Bode plots
│   ├── signals/                           # VCG spatial loops & EEG frequency spectra
│   └── hardware/                          # Breadboard wiring, electrode layout & waveforms
│
├── reports/                               # Academic technical reports
│   ├── Report1_VCG_EEG_Signal_Processing_AlirezaNajafi.pdf
│   ├── Report2_Analog_Biopotential_Circuits_AlirezaNajafi.pdf
│   ├── Report3_Hardware_ECG_Acquisition_and_HRV_AlirezaNajafi.pdf
│   ├── Complete_IBME_Coursework_Report_AlirezaNajafi.pdf   # Merged 26-page portfolio
│   └── Biomedical_Instrumentation_and_Signal_Processing_Report.tex # IEEEtran LaTeX source
│
├── LICENSE                                # MIT Open-Source License
└── README.md                              # Repository documentation

---

🚀 How to Run & Reproduce

1. MATLAB Biosignal Processing (Module 1)

% Open MATLAB and navigate to Module 1
cd('01-biosignal-processing-vcg-eeg');

% Run VCG dipole reconstruction
run('IBME_CA1_part1.m');

% Run EEG sleep stage spectral analysis
run('IBME_CA1_part2_stage0.m');
run('IBME_CA1_part2_tofind.m');

2. NI Multisim Circuit Simulation (Module 2)

1. Open NI Multisim (v14.0 or higher).
2. Open 02-analog-front-end-multisim/ECG.ms14.
3. Run AC Analysis / Bode Plotter to inspect passband gain ($0.1 - 100 ext{ Hz}$).
4. Open Notch_filter.ms14 to verify $> -40 ext{ dB}$ attenuation at $50 ext{ Hz}$.

3. Embedded Hardware & Real-Time Telemetry (Module 3)

1. Hardware Assembly:
   • Connect AD8232 OUTPUT to Arduino A0, LO+ to Pin 4, LO- to Pin 7, 3.3V to 3.3V, and GND to GND.
   • Affix three Ag/AgCl electrodes in Einthoven Lead-I orientation (RA, LA, RL).
2. Flash Microcontroller:
   • Open ad8232_ecg_sampler.ino in Arduino IDE.
   • Select your target board and upload.
3. Run MATLAB Analysis Engine:

   cd('03-hardware-ecg-acquisition-hrv/matlab');
   
   % Process recorded in-vivo dataset:
   run('qrs_detect_and_hrv_metrics.m');
   run('hrv_spectral_welch_psd.m');

---

📑 Technical Reports

The comprehensive documentation is compiled into academic reports located in reports/:

• Report 1: Biosignal Processing, VCG Modeling & EEG Sleep Staging (13 Pages)
• Report 2: Analog Front-End Biopotential Circuit Synthesis (10 Pages)
• Report 3: Real-Time Hardware ECG Telemetry & Heart Rate Variability (3 Pages)
• Complete Unified IBME Technical Report (26 Pages)
• IEEEtran Two-Column LaTeX Source

---

📚 Citation

If you utilize this coursework, circuit designs, or signal processing algorithms in academic research, please cite:

@misc{najafimotiei2023ibme,
  author       = {Alireza Najafi Motiei},
  title        = {Introduction to Biomedical Engineering (IBME): Biopotential Instrumentation, Front-End Circuits and Real-Time Telemetry},
  year         = {2023},
  publisher    = {GitHub},
  howpublished = {\url{https://github.com/alirezanmotiei/Introduction-to-Biomedical-Engineering-IBME}},
  note         = {Coursework Portfolio, Department of Electrical and Computer Engineering, University of Tehran}
}

---

👤 Author & License

• Author: Alireza Najafi Motiei (Student ID: 810100224)
• Department: Faculty of Electrical and Computer Engineering, University of Tehran
• Supervision: Dr. Majid Badiee Rostami
• Academic Term: Fall 2023 (پاییز ۱۴۰۲)
• License: Released under the MIT License.

_{University of Tehran • Department of Electrical and Computer Engineering}