Instrumentation and methods for frequency-domain and multimodal near-infrared spectroscopy

Abstract

In this thesis, instrumentation for a frequency-domain (FD) near-infrared spectroscopy (NIRS) device and for multimodal brain measurements was implemented. Different techniques were applied to human and newborn lamb brain studies. A method to detect light leakage in NIRS measurements was developed. The FD instrument, developed at the Helsinki University of Technology was extended, by implementing 16 pseudo-differential preamplifiers for the analog-to-digital converters with a low noise and excellent interchannel isolation. An instrumentation for a digital signal-processor based lock-in amplifier was also developed. Methods for increasing the number of wavelengths and source positions in the imaging instrument were studied. A second-generation source system with a fast fiber-optic switch and four high-power laser diodes with a low noise temperature-stabilizing electronics was implemented. The imaging device was placed into a cabinet to enable its portability. New detection and source fiber terminals were developed for multimodal brain studies. The different versions of the imaging instrument were applied to four human brain measurements. In a breath-holding and hyperventilation study, the effects of source-detector distance (SDD) and measurement wavelength on the contrast of NIRS responses and the frequency content of signals were studied. Hemodynamic changes in the human brain related to the changes in sleep stages were detected. The multimodal NIRS and electroencephalography measurement setup was implemented and used to study the effects of baseline blood flow changes on the visually evoked hemodynamic and neuronal responses. The feasibility of NIRS as a part of multimodal monitoring setup to detect cerebral hemodynamic changes induced by iloprost and nitric oxide in the preterm lamb brain was also demonstrated. The linearity of the FD measurement parameters as a function of SDD on the human forehead was studied. The regression of phase measurement was observed to be sensitive to light leakage from source to detectors, much more than the regression of modulation amplitude or average intensity measurement. Utilizing this observation, a method to detect light leakage based on the pathlength measurement was developed. The contrast and depth sensitivity of NIRS signals were shown to decrease in measurements where light leakage occurs.