Dual-source dual-detector optical probe for improved depth discrimination in functional near-infrared spectroscopy

Near infrared spectroscopy (NIRS) has been shown to be an effective functional brain imaging modality and to have potential of detecting cerebral activity noninvasively. By delivering light in the near-infrared range (600--900 nm) to a scalp location and detecting the optical intensities from another scalp location at a distance of 1--4 cm, we can determine the concentration changes of Oxy-hemoglobin and Deoxy-hemoglobin that are the signatures of evoked brain activity. Since functional brain imaging depends on hemodynamic trends that are specific to the cerebral cortex in the practical NIRS measurement, it is necessary to reduce the interference of hemodynamic trends in the intervening extracerebral tissue (scalp and skull mainly). A previous study has employed one single-source and dual-detector configuration geometry, including one "near" (≈ 1 cm) source-detector separation and one "far" (≈ 3--4 cm) source-detector separation, to remove contributions from the extracerebral tissue. A least squares method was used to fit the temporal trace of the "far" signal with a scaled "near" signal (the scaling factor being the fitting parameter) and the residual was considered as the signals originated in the brain. Our method is based on the same concept but we introduced a second additional source close to the far detector so that we could take into account the contributions from the superficial extracerebral tissue near both detectors. Therefore, in contrast to previous methods, we assumed the hemodynamic trends in the extracerebral tissue are not homogeneous or layer-like. With the proposed source-detector arrangement we studied the brain hemodynamics elicited during a finger tapping protocol, our findings confirm the assumption that hemodynamic changes occurring in the extracerebral tissue can be heterogeneous. Also, we observed that our method for depth discrimination leads to a residual which is representative of cortex-specific hemodynamic trends and our findings are in agreement with known physiological responses to neural activation. In conclusion, our results demonstrated that it is important to use a dual-source and dual-detector configuration geometry for improved depth discrimination of functional NIRS during brain imaging.