Recovering signals in physiological systems with large datasets.

Area of Science:

• Physiology
• Systems Biology
• Signal Processing

Background:

• Physiological variables are often indirectly measured from noisy signals.
• Accurate estimation of these variables is crucial for physiological studies.

Purpose of the Study:

• Introduce two novel empirical methods for estimating true physiological signals from indirect, noisy measurements.
• Validate these methods computationally and experimentally.

Main Methods:

• Extension of Tikhonov regularization for large-scale problems using sequential updates.
• Least-squares estimation with assumed uniform input over a known time interval to improve problem conditioning.

Main Results:

• Methods validated using flow-through respirometry data with known CO2 infusion patterns.
• Demonstrated sub-second accuracy in recovering input signals.
• Applied to grasshopper respiratory data to analyze abdominal pumping and CO2 emission timing.

Conclusions:

• The developed methods accurately estimate physiological signals from noisy data.
• These techniques offer sub-second precision for input estimation in linear systems.
• Applicable to various physiological studies and biological systems analysis.