Abstract:
Functional near infrared spectroscopy (fNIRS) is a developing non-invasive technique used for measurement of oxygenation in the adult and newborn human brain during cognitive tasks. The oxygenation in the brain is expressed as the change in the concentrations of the major absorbers such as oxy- and deoxy-haemoglobin in blood as a response to the brain activation during light-tissue interactions. The absorption of these absorbers at different wavelengths are calculated by using a Modified Beer- Lambert Law (MBLL). Depending on the aim of the fNIRS measurement, at least two different wavelengths are selected from the optical window (600-900 nm) for the detection of the concentration changes of these absorbers in order to minimise the undesired effect of “crass-talk”. The probe of fNRIS containing a combination of various light sources (LD, LED, etc.) and detectors aims to pinpoint the activated regions of the brain relying on the theory of light path distribution known as “banana-shape” in literature. As a well proven technique, Monte Carlo Simulations which describes the photon migration multi layer media is used for modelling of head and fNIRS probe for investigation of the system. The most important disadvantage of the technique is the contamination of the brain signal with the signals received from the superficial layers of the head, namely; scalp and skull layers. For the decoupling of these signals, various signal regression and filtering techniques are used. In this thesis as a signal regression technique is presented, where a real fMRI data is used in Monte Carlo Simulations for assigning a near-far detector position which respectively contains the signal from the superficial layers only and the signals obtain from both the superficial layers and brain matter. The positions of these detectors were found to be around 17, 18 and 19 mm from the source for the head model with an average human skull thickness of 7 mm.|Keywords : MCML, fNIRS, Photon Migration in Biological Tissues.
