4.3.5 fNIR Device Used in Thesis
This section will focus on the technical aspects of the continuous wave fNIR device used in this thesis, which is an fNIR Imager 100 continuous wave device (fNIR Devices, Potomac, MD, USA).
The fNIR device consists of a sensor pad fitted with 4 LED lights and 10 light sensors (Figure 2.1). The LED lights have a dual wavelength of 730nm and 850nm and the light sensors consists of silicon photodiodes with an integrated trans-impedance preamp. The arrangement of the light sources and the sensors results in data being gathered at 16 recording channels (or optodes). The sampling rate of the device is 2Hz, with a complete scan of all 16 optodes completed twice in each second (at both wavelengths). The LEDs activate sequentially (from left to right), with each of the four sensors surrounding the LED sampled in order to measure the light that has travelled from the detectors through the tissue in the head before reaching the detectors. The distance between the light source and the sensors is 2.5cm, meaning that the device is capable of recording haemodynamic activity at a depth of approximately 1.25cm, with a spatial resolution of 2-3cm2 (Gefen et al. 2014). The sensor pad is placed onto the forehead and is secured in place with a Velcro strap. The sensor pad is connected to a control box, which contains. That is in turn connected to a computer that has a software programme on it for operating the fNIR device and storing the data collected from it. The software programme (Cobi Studio) provides an interface through which it is possible to start and stop the device from recording, take baselines, as well as placing time stamps (called markers) into the data to note a point in time in which something important took place (e.g. a stimulus appeared).
The photons produced by the LEDs follow a characteristic banana shaped path through the target tissue and then back to a detector on the same approximate
plane (Figure 4.2). Most of the light transmitted from the light sources is absorbed or scattered but some makes its way to the detector sites. The attenuation of the photon density at the two wavelengths (as recorded at the detectors) provides the information (the spectroscopic signature) required to calculate the relative concentrations of Hb and HbO2.
Figure 4.1. Typical pathlengths taken by photons as they travel from an LED light source, through brain tissue and on to a light detector.
For this process to work, the sensor pad on this system needs to be in contact with the skin. As a result this fNIR device is primarily used to monitor hemodynamic activity in the frontal lobes, namely the DLPFC as placing the sensor on a participant’s forehead accesses these areas. The DLPFC is the most accessible area of the frontal lobes when using fNIR as it is not obstructed by the eyebrows and the eyes. As mentioned above, this fNIR device is capable of accessing regions 1-2cm deep, with a spatial resolution of 2-3cm2. As a result, this fNIR device is especially suitable for recording activity in the frontal lobes as they are a relatively short depth below the skin on the forehead (due to the shape of the brain) and the skull is relatively thin in this region. This allows the light from the LED emitters to better reach the cortex.
Before undergoing any analysis, the data are passed through a low-pass filter to remove physiological noise in the signal, including the signals from respiration and the cardiac cycle (Izzetoglu et al. 2005). As detailed earlier, to
calculate the changes in Hb and HbO2,fNIR uses a modified Beer-Lambert Law.
The combined changes in the intensity of the light received at the sensors (at 730nm and 850nm) are used to calculate the changes in Hb and HbO2 over time (absolute concentrations cannot be calculated with this device). This process first involves creating baseline readings for the relative concentrations of Hb and HbO2. The baseline also serves to factor in individual differences in the properties of participant’s skin, the shape of their heads and the shape of their brains. Further readings are then taken at different time points to calculate the change in Hb and HbO2 over time (e.g. during psychological tasks). Calculating absolute concentrations of Hb and HbO2 is not possible with this device. These changes in Hb and HbO2 can then be used to provide other measures of haemodynamic activity such as the change in blood volume and the change in blood oxygenation:
Δ Blood Volume = ΔHbO2 + ΔHb Δ Blood Oxygenation = ΔHbO2 – ΔHb
The procedure for using this fNIR device first involves properly placing the sensor pad on the participant’s forehead and securing it in place with the Velcro straps. A test must then be run to establish if the device is recording a clean signal.
If there is hair in the way of the light sources or detectors, or if the sensor pad is loose in certain places and is not in contact with the skin, the recorded signal will not be an accurate measure of haemodynamic activity. The signal can be checked on the computer that runs the custom software for the device. Once any adjustments have been made and the signal is at an acceptable level a baseline reading must be taken. This involves taking 10 seconds worth of haemodynamic recordings while the participant is in a state of rest. The purpose of the baseline is to establish relative concentrations of Hb and Hb02 at rest, which can be used to measure functional changes in Hb and HbO2 associated with psychological tasks.
Once the baseline is taken the psychological task can begin. As soon as the task begins the fNIR device is set to record any changes in haemodynamic activity relative to the baseline. The device is left recording until the end of the task when it is stopped from recording. The device is left to record any changes throughout
the duration of the task. As soon as the recording is stopped, several files containing data are created. One file contains raw data, another file contains data that have already had the changes in Hb and HbO2 over the tasks duration calculated, and another file contains information on any markers that have been used during the recording process. These values are provided for each of the 16 recording channels.