15.1 Sources and types of noise in biosignals

Biosignals often contain unwanted noise from various sources, including power line interference, motion artifacts, and baseline wander. These disturbances can significantly impact signal quality, making analysis and interpretation challenging for bioengineers and healthcare professionals.

Different types of noise, such as white, pink, and colored noise, affect biosignals in unique ways. Understanding these noise characteristics is crucial for developing effective signal processing techniques to improve signal-to-noise ratios and extract meaningful information from biosignals.

Sources and Types of Noise in Biosignals

Sources of biosignal noise

Top images from around the web for Sources of biosignal noise

• 

  GI - The baseline wander correction based on the improved ensemble empirical mode decomposition ... View original

  Is this image relevant?

• 

  Characteristics of biosignals - WikiLectures View original

  Is this image relevant?

• 

  Powerline noise elimination in biomedical signals via blind source separation and wavelet ... View original

  Is this image relevant?

• 

  GI - The baseline wander correction based on the improved ensemble empirical mode decomposition ... View original

  Is this image relevant?

• 

  Characteristics of biosignals - WikiLectures View original

  Is this image relevant?

1 of 3

Top images from around the web for Sources of biosignal noise

• 

  GI - The baseline wander correction based on the improved ensemble empirical mode decomposition ... View original

  Is this image relevant?

• 

  Characteristics of biosignals - WikiLectures View original

  Is this image relevant?

• 

  Powerline noise elimination in biomedical signals via blind source separation and wavelet ... View original

  Is this image relevant?

• 

  GI - The baseline wander correction based on the improved ensemble empirical mode decomposition ... View original

  Is this image relevant?

• 

  Characteristics of biosignals - WikiLectures View original

  Is this image relevant?

1 of 3

• Power line interference introduces 50 Hz or 60 Hz sinusoidal signal depending on power line frequency (North America, Europe)
• Motion artifacts cause abrupt changes in signal amplitude and introduce low-frequency noise due to patient movement or electrode displacement (muscle contractions, cable movement)
• Baseline wander results in slow, non-stationary drift in signal's baseline caused by respiration, body movements, or changes in electrode-skin impedance (breathing, sweating)

Types of signal noise

• White noise has flat power spectral density across all frequencies containing equal power within fixed bandwidth at any center frequency (thermal noise, quantization noise)
• Pink noise (1/f noise) has power spectral density inversely proportional to frequency with more power at lower frequencies and less at higher frequencies (flicker noise, biological systems)
• Colored noise has non-uniform power spectral density such as brown noise (1/f^2 noise) and blue noise where power increases with frequency (Brownian motion, high-frequency emphasis)

Noise impact on biosignals

• Reduced signal-to-noise ratio (SNR) obscures desired signal components making them difficult to distinguish leading to decreased accuracy in signal analysis and interpretation (low-amplitude EEG, fetal ECG)
• Noise affects performance of various signal processing methods such as filtering, feature extraction, and pattern recognition (adaptive filters, wavelet denoising)
• Noise introduces artifacts that may be mistaken for physiologically relevant signal components leading to incorrect diagnoses or misinterpretation of patient's condition (false arrhythmia detection, muscle artifacts in EEG)

Frequency characteristics of noise

• Power line interference occurs at 50 Hz or 60 Hz overlapping with frequency content of some biosignals such as electromyography (EMG) and electroencephalography (EEG)
• Motion artifacts and baseline wander typically have low-frequency content (< 1 Hz) overlapping with slow-varying biosignals such as electrocardiography (ECG) and respiration signals
• White noise has flat power spectral density potentially affecting all frequencies of biosignal
• Pink noise has more power at lower frequencies overlapping with frequency content of many biosignals (heart rate variability, EEG)
• Colored noise may overlap with various frequency ranges of biosignals depending on its specific characteristics (muscle noise in EMG, high-frequency noise in ECG)