Information content of data types in time-domain optical tomography

Area of Science:

• Biomedical optics
• Medical imaging physics

Background:

• Time-domain optical tomography (TDOT) is crucial for non-invasive imaging.
• Quantifying information content in TDOT data is essential for optimizing imaging protocols and understanding limitations.
• Detectability of signals within complex backgrounds influences image quality and diagnostic accuracy.

Purpose of the Study:

• To quantify the information content of different data types in TDOT.
• To investigate the spatial detectability of signals in uniform and structured backgrounds.
• To establish a methodology for pre-reconstruction information content assessment.

Main Methods:

• Studied signal detectability within attenuation and reduced scatter coefficients.
• Analyzed detection performance in both uniform and structured background environments.
• Evaluated the information content of the full temporal waveform using photon arrival time and total counts.

Main Results:

• Demonstrated a complex spatial dependence of detectability maps on signal type, data type, and background.
• Identified mean photon arrival time and total counts as effective summaries of information content for lesion detection.
• Established a non-linear, pre-reconstruction methodology for quantifying information content.

Conclusions:

• Signal and background characterization are critical for accurate information content assessment in TDOT.
• The developed methodology allows for prior evaluation of information content without linearity assumptions.
• Findings provide a framework for optimizing TDOT acquisition and analysis for improved lesion detectability.