**2. Techniques for assessment of brain function**

Compared to other clinical neurological methods such as electroencephalography (EEG), positron emission tomography (PET), fMRI, and MEG, NIRS was recently developed. EEG is a noninvasive method that directly measures the electrical activity of the brain via electrodes placed on the scalp. EEG has a high temporal resolution but poor spatial resolution. PET is an invasive method that uses radioactive tracers to measure the metabolism of the brain. PET provides high specific spatial resolution but poor temporal resolution. fMRI is a noninvasive method that measures the changes in blood oxygenation, and provides good spatial resolution but limited temporal resolution. MEG is a noninvasive method that measures the magnetic fields induced by electrical activity in the brain. MEG has good spatial and temporal resolution; however, MEG also requires the participants to remain very still during measurement.

NIRS is completely noninvasive, more patient-friendly, and can be performed very easily compared to other neurological techniques. It is relatively inexpensive compared to fMRI, making it possible to measure haemodynamic changes at a low cost. Furthermore, the NIRS instrument is portable; measurements can be obtained *in situ*, for example, at the bedside. In addition, the NIRS probe for assessing brain activity is very flexible, and does not require the patient to be still or maintain the head in any particular position (Minagawa-Kawai et al., 2008). These attributes are especially useful when the patients are children, as it enables them to feel relaxed even in an experimental setting. fMRI and MEG studies are difficult to perform in children, because these techniques require the patients to be still. Most fMRI studies in infants have been performed while they are asleep or are sedated inside the imaging tunnel (Lloyd-Fox et al., 2010; Minagawa-Kawai et al., 2008).

NIRS has a higher temporal resolution than fMRI. Haemodynamic changes at 10 Hz can be measured with commercially available products, whereas a typical fMRI instrument provides a temporal resolution of a few seconds. NIRS provides the concentration changes in both oxy-haemoglobin and deoxy-haemoglobin, whereas fMRI measures only deoxyhaemoglobin. NIRS is not as noisy as fMRI, and the NIRS signal is less affected by motion artefacts compared to signals of fMRI or MEG. Additionally, NIRS permits continuous and repeated measurement.

However, one limitation of NIRS is that it assesses brain function indirectly by measuring the haemodynamic changes in haemoglobin via an optical fibre probe attached to the scalp. The working principle of NIRS is that neural activation requires oxygen consumption; therefore, the haemoglobin concentration in the blood vessels depends on the neural activity. EEG, in contrast, assesses brain function directly by measuring the electrical potential of neurons via electrodes attached to the scalp. Additionally, although NIRS assesses brain activity via haemodynamic changes in the cerebral cortex, it is unable to examine deep brain activation. Specifically, NIRS allows the assessment of haemoglobin concentrations up to only 2 or 3 cm below the skull; therefore, the activation in the deep brain regions cannot be measured. Furthermore, NIRS has a low spatial resolution compared to fMRI or MEG; the spatial resolution of NIRS is 20–30 mm, whereas that of fMRI is 1–2 mm.
