Principles of Biosignal Acquisition and Processing (Short Notes)

Biosignal acquisition and processing is the chain that turns weak biological measurements into reliable digital representations for analysis, features, and clinical/biometric decisions. The process typically has two coupled stages: keyword analog front-end (AFE) and keyword digitization, followed by keyword preprocessing/denoising. AFE requirements often include high input impedance and low noise amplification because electrode signals are small and easily corrupted by interference and electrode-electrolyte effects. In practical ECG-style systems, the analog front-end frequently includes an instrumentation amplifier (INA) with keyword gain control and analog filtering prior to ADC. This analog processing can also target dominant disturbances such as keyword baseline wander before the signal is digitized.2

Key idea: if acquisition is wrong (wrong gain, missing anti-aliasing, poor electrode contact), preprocessing can only partly recover the signal; it is therefore principled to co-design the analog and digital steps.2

Footnotes

1. A Fully Differential Analog Front-End for Signal Processing from EMG Sensor in 28 nm FDSOI Technology - Discusses analog front-end needs like low-noise amplification, high input impedance, instrumentation amplification, and AFE preprocessing. ↩ ↩²

2. Analog front-end and algorithm co-design for efficient biosignal acquisition: And its application to cardiac signal monitoring - TU Delft Research Portal - Notes baseline wander as a major ECG acquisition challenge and motivates AFE high-pass filtering. ↩ ↩²

Block-level view (what to write in an exam short note)

A good short note typically covers: (i) measurement chain, (ii) sampling and aliasing control, and (iii) digital preprocessing choices for common biosignal noise.

Measurement chain: electrodes → AFE (amplify + filter) → anti-aliasing → ADC → preprocessing (denoise + correct) → features.

Sampling: sampling rate must satisfy the keyword Nyquist condition, and practical systems add an keyword anti-aliasing filter to suppress aliasing.2

Preprocessing: common digital operations include baseline correction (e.g., high-pass / detrending), keyword notch filtering for 50/60 Hz power-line interference, and bandpass filtering (e.g., Butterworth) to remove out-of-band noise.2

Footnotes

1. Sampling and Aliasing: The Nyquist Theorem Explained Visually | DSP Foundations | DSPRelated.com - Provides Nyquist condition intuition and emphasizes aliasing irreversibility and the role of anti-aliasing. ↩

2. Anti-aliasing filter - Wikipedia - Defines anti-aliasing filter purpose before sampling to satisfy Nyquist–Shannon constraints and prevent aliasing. ↩

3. Two-stage motion artefact reduction algorithm for electrocardiogram using weighted adaptive noise cancelling and recursive Hampel filter - Describes ECG pre-processing with band-pass + notch + adaptive cancellation; explains difficulty of motion artifacts. ↩

4. Baseline wander and power-line interference elimination of ECG signals using efficient signal-piloted filtering - PMC - States baseline wander and power-line interference are major noise elements and describes their frequency characteristics. ↩

1) Principles of biosignal acquisition (AFE + electrode considerations)

1. Weak biopotentials and electrode–skin interface: Biosignals are typically low amplitude and sensitive to contact quality and interference, motivating amplification with low noise and appropriate input impedance.2
2. Analog gain and filtering before digitization: An INA/programmable gain stage provides sufficient amplitude for ADC dynamic range while analog filters reduce out-of-band components and interference.2
3. Cope with dominant physiological/system artifacts: For ECG acquisition, keyword baseline wander is a major acquisition challenge caused by movement/respiration and can lead to saturation or waveform distortion; suppression often involves AFE high-pass filtering.
4. Anti-aliasing as a hard requirement: Because aliasing is irreversible after sampling, analog anti-alias filtering is used to restrict bandwidth to satisfy the Nyquist–Shannon sampling theorem over the band of interest.2

Typical architecture summary (one paragraph format)

“Biosignal acquisition uses electrodes connected to an analog front-end that performs low-noise amplification and analog filtering. This stage sets gain for ADC range and suppresses major disturbances (e.g., baseline wander). Before ADC, an anti-aliasing filter ensures that frequency content above the Nyquist frequency is attenuated to prevent aliasing.”2

Footnotes

1. A Fully Differential Analog Front-End for Signal Processing from EMG Sensor in 28 nm FDSOI Technology - Discusses analog front-end needs like low-noise amplification, high input impedance, instrumentation amplification, and AFE preprocessing. ↩ ↩² ↩³

2. Challenges of Physiological Signal Measurements Using Electrodes: Fundamentals to Understand the Instrumentation - IEEE Solid-State Circuits Magazine (listed within EMG AFE paper references) - Included in the same referenced discussion of electrode/sensor measurement challenges. ↩

3. Sampling and Aliasing: The Nyquist Theorem Explained Visually | DSP Foundations | DSPRelated.com - Provides Nyquist condition intuition and emphasizes aliasing irreversibility and the role of anti-aliasing. ↩ ↩²

4. Analog front-end and algorithm co-design for efficient biosignal acquisition: And its application to cardiac signal monitoring - TU Delft Research Portal - Notes baseline wander as a major ECG acquisition challenge and motivates AFE high-pass filtering. ↩

5. Anti-aliasing filter - Wikipedia - Defines anti-aliasing filter purpose before sampling to satisfy Nyquist–Shannon constraints and prevent aliasing. ↩ ↩²

2) Sampling and anti-aliasing principles (Nyquist + ADC)

Sampling theorem: To avoid ambiguous representation, ADC sampling rate should be greater than twice the highest frequency component of the analog signal of interest ($f_s > 2 f_{max}$). The frequency $f_s/2$ is the Nyquist frequency.2

Anti-aliasing filter: Practical systems place an anti-aliasing filter before the ADC to attenuate spectral components above the Nyquist limit, ideally restricting the sampled bandwidth to one unambiguous Nyquist zone.2

Exam-style points to include:

• Choose sampling rate based on expected biosignal bandwidth (and guard band).
• Add analog anti-aliasing filtering before ADC.
• Consider ADC resolution/dynamic range requirements driven by AFE gain and expected signal amplitude.

(Concept reference for anti-aliasing rationale and Nyquist-based design is widely described in sampling/anti-aliasing materials.)2

Footnotes

1. Anti-aliasing filter - Wikipedia - Defines anti-aliasing filter purpose before sampling to satisfy Nyquist–Shannon constraints and prevent aliasing. ↩ ↩² ↩³

2. Sampling and Aliasing: The Nyquist Theorem Explained Visually | DSP Foundations | DSPRelated.com - Provides Nyquist condition intuition and emphasizes aliasing irreversibility and the role of anti-aliasing. ↩ ↩² ↩³

3) Digital biosignal processing principles (preprocessing pipeline)

Digital preprocessing improves data quality so downstream steps (feature extraction, detection, classification) are accurate.

Common preprocessing steps (ECG/EEG-style generalization)

1. Detrending / baseline correction: Remove low-frequency baseline drift (e.g., due to respiration/motion). Baseline wander is often low-frequency (e.g., around $0$–$0.7$ Hz in ECG descriptions), so high-pass filtering or adaptive baseline cancellation is used.
2. Band-pass filtering: Remove out-of-band noise while keeping the physiological band. For ECG, a zero-phase bandpass such as Butterworth (example: $0.05$–$130$ Hz used in one motion-artefact pre-processing pipeline) is an illustrative approach.
3. Notch filtering for power-line interference (PLI): Mains interference is typically at 50 Hz or 60 Hz; notch (band-stop) filters suppress these narrowband components.2
4. Artifact handling (motion/muscle): Motion artefacts can overlap with signal spectra and may require adaptive techniques or reference channels (e.g., accelerometer-based cancellation). This is more difficult than simple filtering when spectra overlap.

Footnotes

1. Baseline wander and power-line interference elimination of ECG signals using efficient signal-piloted filtering - PMC - States baseline wander and power-line interference are major noise elements and describes their frequency characteristics. ↩ ↩²

2. Two-stage motion artefact reduction algorithm for electrocardiogram using weighted adaptive noise cancelling and recursive Hampel filter - Describes ECG pre-processing with band-pass + notch + adaptive cancellation; explains difficulty of motion artifacts. ↩ ↩²

3. Signal Processing Techniques for Removing Noise from Electrocardiogram Signal (JBER Signal-Processing.pdf) - Summarizes typical noise sources and mentions notch filtering for 50/60 Hz power-line interference. ↩

4) Noise sources and how they map to processing choices (writeable mapping)

Below is a compact “noise → principle → typical method” table suitable for short notes.

Footnotes

1. Baseline wander and power-line interference elimination of ECG signals using efficient signal-piloted filtering - PMC - States baseline wander and power-line interference are major noise elements and describes their frequency characteristics. ↩ ↩²

2. Signal Processing Techniques for Removing Noise from Electrocardiogram Signal (JBER Signal-Processing.pdf) - Summarizes typical noise sources and mentions notch filtering for 50/60 Hz power-line interference. ↩

3. Two-stage motion artefact reduction algorithm for electrocardiogram using weighted adaptive noise cancelling and recursive Hampel filter - Describes ECG pre-processing with band-pass + notch + adaptive cancellation; explains difficulty of motion artifacts. ↩ ↩²