Quantification of blood flow index in diffuse correlation spectroscopy using long short-term memory architecture

Area of Science:

• Biomedical Optics
• Medical Imaging
• Signal Processing

Background:

• Diffuse correlation spectroscopy (DCS) is a noninvasive optical technique used to measure blood flow.
• Blood flow index (BFI) derived from DCS correlates with absolute blood flow, making it valuable for physiological monitoring.
• Accurate and rapid quantification of BFI is crucial for real-time clinical applications.

Purpose of the Study:

• To investigate and assess the utility of a long short-term memory (LSTM) neural network architecture for quantifying BFI in DCS.
• To evaluate the performance of the LSTM model in terms of accuracy and computational speed compared to existing methods.
• To determine the potential of the proposed LSTM-based approach for continuous, real-time blood flow monitoring.

Main Methods:

• Development and application of a long short-term memory (LSTM) architecture.
• Acquisition of normalized intensity autocorrelation function data from phantom and in vivo experiments.
• Utilizing LSTM for the quantification of blood flow index (BFI) from DCS measurements.

Main Results:

• The proposed LSTM architecture demonstrated improved accuracy in BFI quantification.
• Faster computational times were achieved using the LSTM model.
• Experimental validation using phantom and in vivo data supported the efficacy of the LSTM approach.

Conclusions:

• The LSTM architecture is a viable and effective tool for accurate BFI quantification in DCS.
• This approach offers significant advantages in terms of speed and accuracy for blood flow assessment.
• The developed method holds promise for continuous, real-time monitoring of blood flow in various physiological conditions.