Image Quality Analysis of High-Density Diffuse Optical Tomography Incorporating a Subject-Specific Head Model

High-density diffuse optical tomography (HD-DOT) methods have shown significant improvement in localization accuracy and image resolution compared to traditional topographic near infrared spectroscopy of the human brain. In this work we provide a comprehensive evaluation of image quality in visual cortex mapping via a simulation study with the use of an anatomical head model derived from MRI data of a human subject. A model of individual head anatomy provides the surface shape and internal structure that allow for the construction of a more realistic physical model for the forward problem, as well as the use of a structural constraint in the inverse problem. The HD-DOT model utilized here incorporates multiple source-detector separations with continuous-wave data with added noise based on experimental results. To evaluate image quality we quantify the localization error and localized volume at half maximum (LVHM) throughout a region of interest within the visual cortex and systematically analyze the use of whole-brain tissue spatial constraint within image reconstruction. Our results demonstrate that an image quality with less than 10 mm in localization error and 1000 m(3) in LVHM can be obtained up to 13 mm below the scalp surface with a typical unconstrained reconstruction and up to 18 mm deep when a whole-brain spatial constraint based on the brain tissue is utilized.