Why fNIRS?

What is fNIRS?

fNIRS is a non-invasive neuroimaging technique that uses light to record blood flow in the brain as a measure of neural activation.

Specifically, near infrared light (the spectrum of 650 – 900 nm) is transmitted into the skull. fNIRS takes advantage of this optical window – the range where the absorption of light is small, thus light is able to diffuse several centimetres through biological tissue to reach brain tissue.

Neuronal activity leads to changes in blood flow and oxygen consumption, known as neurovascular activity. This leads to an increase in oxygenated haemoglobin, and a decrease in deoxygenated haemoglobin – these forms absorb near infrared light differently. Thus, functional brain activity is measured through the changes in light absorption, which reflects changes in blood oxygenation levels.

Advances in this technology have led to the development of multichannel, wearable devices - this means a larger portion of the head can be monitored, and the data from fNIRS recordings can be stored on the unit itself, or sent wirelessly to a laptop.

Why fNIRS for cognitive neuroscience?

For over 20 years, cognitive neuroscientists have used fMRI (with PET, EEG and MEG) to make sense of the function of the human brain. So why do we need a new methodology?

The traditional methods of cognitive neuroscience provide strong experimental control, as isolated participants respond to a limited set of stimuli with a limited number of possible responses. It is now being recognized that these situations have limited ecological validity: the real world provides many competing stimuli and many possible responses, including other people who respond dynamically and in real time. Traditional methods are not well suited to studying these real world situations.

fNIRS is an ideal method for real world cognitive neuroscience because an fNIRS cap does not restrict movement or everyday behaviour.

fNIRS has advantages because it is:

• Portable
• Motion tolerant
• Quiet
• Low cost
• Suitable for all participants
• Better than fMRI in terms of its temporal resolution

These features mean fNIRS is the ideal method for taking cognitive neuroscience 'into the wild', moving beyond the constraints of traditional lab approaches.

What are the challenges?

Developing a new method means facing new challenges, which is why we need to bring together expertise in cognition, social interaction and engineering to make fNIRS work for us. Our work aims to understand how brain processes can explain behaviour, so we need to:

1) Make sense of our brain data:

We are working to develop new methods to help us analyse data recorded in real-world contexts, where experimental control is diminished and participants have free control over what stimuli to enage with.

One such method is known as AIDE (Automatic IDentification of Functional Events; Pinti et al., 2017). AIDE is an algorithm that can be used in a brain-first approach to analysis. This means looking for particular patterns in pre-processed fNIRS data, to detect the onsets and durations of functional events. These events can then be connected to participants' behaviour.

2) Make sense of human behaviour:

Advances in recording technology means we can create rich recordings of real world behaviour through the usage of motion capture, face capture, eyetracking, video/audio recordings etc. New analysis methods in this domain will be important future developments.

3) Put them both together:

To understand how brain processes can explain behaviour, fNIRS can be used in conjunction with behavioural measures as described above, and also other physiological measures such as heart rate or breathing rate, to provide a more comprehensive overview of coordinated changes in the body.

Data streams from fNIRS and behavioural recordings need to be processed and synchronised, so neural events can be linked to behaviour.

These findings can contribute towards models of cognition and social interaction.

4) Apply this in clinical contexts:

The flexibility of fNIRS to collect data in a range of contexts allows for the neuromonitoring of clinical populations. This makes it possible to examine social interactions in people with autism, and many other patient samples. For example, studies have illustrated differences in fNIRS responses in 4-month-old children at risk of autism (Lloyd-Fox et al., 2013), and adults with autism when processing facial stimuli (Jung et al., 2016).