Photoplethysmograph quality estimation through multichannel filtering.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Diagnostics

Background:

• Physiological waveform quality is crucial for accurate medical condition detection.
• Existing methods often rely on complex, ad-hoc feature extraction.
• A robust, generalizable signal quality index is needed.

Purpose of the Study:

• To introduce a novel point-by-point signal quality index (SQI) for physiological waveforms.
• To evaluate the SQI's performance using adaptive multichannel prediction.
• To assess the SQI's utility in identifying high-quality photoplethysmograph (PPG) signals.

Main Methods:

• Developed a new SQI based on adaptive multichannel prediction.
• Applied the SQI to photoplethysmograph (PPG) waveforms.
• Validated SQI against signal-to-noise ratio (SNR) and human quality assessments.
• Utilized receiver-operating-characteristic (ROC) curve analysis for performance evaluation.

Main Results:

• The novel SQI demonstrated a monotonic relationship with SNR and human quality assessments.
• An area under the ROC curve of 0.863 was achieved, indicating strong classification performance.
• A SQI threshold of 0.8 reliably identified good quality PPG waveforms.

Conclusions:

• The proposed SQI offers an effective, morphology-independent method for assessing physiological signal quality.
• This index shows significant promise for improving the reliability of PPG-based medical diagnostics.
• The SQI provides a quantitative measure for real-time signal quality monitoring.