Speaker
Description
Time-Resolved Near-Infrared Spectroscopy (tr-NIRS) is a non-invasive optical technique that shows potential for bedside neuromonitoring of patients with or at risk of brain injuries. Current tr-NIRS analysis methods typically assume the head is optically homogeneous; however, such approaches are too sensitive to changes in the scalp, leading to inaccurate estimates of brain optical properties. Two-layer photon diffusion model inversion algorithms (TLPMI) are a class of algorithms that model the adult head as a two-layer medium to reduce the influence of the scalp on the estimation of cerebral optical properties. Implementations of TLPMI may include two detectors: a short-distance detector, used to estimate only the optical properties of the top-layer (scalp, skull, cerebrospinal fluid), and a long-distance detector, used to recover properties of the bottom layer (brain). In this study we will use two-detector TLPMI to estimate absolute optical properties in the brain, with a focus on the absorption coefficient. Using digital phantoms generated with both analytical and Monte Carlo methods, we will show how the brain layer absorption coefficient estimated with TLPMI is biased by detector configuration, mischaracterization of top-layer optical properties, fitting parameter tolerance, and medium heterogeneity. Hyperspectral data analysis will also be discussed. The consequences of the errors in TLPMI will be reviewed in reference to the recovery of clinically significant chromophores. This study will aid prospective researchers in effectively applying TLPMI and analyzing its results.
| Keyword-1 | Time-Resolved |
|---|---|
| Keyword-2 | Near-Infrared Spectroscopy |
| Keyword-3 | Non-Invasive |