How to read a baby’s mind


By Priya Silverstein 
Edited by Jonathan and Rachael

Priya, a guest writer for The Brain Domain, is a second-year PhD student at Lancaster University. She spends half her time playing with babies and the other half banging her head against her computer screen.


Okay, I’ll admit that was a bit of a clickbait-y title. But would you have started reading if I’d called it ‘Functional Near Infrared Spectroscopy and its use in studies on infant cognition’? I thought not. So, now that I’ve got your attention…

Before I tell you how to read a baby’s mind, first I have some explaining to do. There’s this cool method for studying brain activity but, as one of the lesser used technologies, it’s a bit underground. It’s called fNIRS (functional Near Infrared Spectroscopy). Think of fNIRS as fMRI’s cooler, edgier sister. Visually, the two couldn’t look more different – with an MRI scanner being a human-sized tube housing a massive magnet that you might have seen on popular hospital dramas, and NIRS simply looking like a strange hat.

MRI.png   fNIRS_cover
Left: MRI scanner, Right: NIRS cap
Picture credit left: Aston Brain Centre, right: Lancaster Babylab

What these two methods do have in common is that they both measure the BOLD (Blood Oxygen Level Dependent) response from the brain. Neurons can’t store excess oxygen, so when they are active, they need more of it to be delivered. Blood does this by ferrying oxygen to the active neurons faster than to their lazy friends. When this happens, you get a higher concentration of oxygenated to deoxygenated blood in the more active areas of the brain.

Now, to the difference between fMRI and fNIRS. fMRI infers brain activity due to oxygenated and deoxygenated blood having different magnetic properties. When the head is put inside a strong magnetic field (the MRI scanner) changes in blood oxygenation, due to changes in brain activity, alter the magnetic field in that area of the brain. fNIRS on the other hand, uses the fact that oxygenated and deoxygenated blood absorb a different amount of light, as deoxygenated blood is darker than oxygenated blood. Conveniently, near-infrared light goes straight through the skin and skull of a human head (don’t worry, this is not at all dangerous and a participant would not feel a thing). So, shining near-infrared light into the head at a source location, and measuring how much light you get back at a nearby detector, gives a measurement of how much light has been absorbed by the blood in that area of the brain. Therefore, you get a measure of a relative change in oxygenated and deoxygenated blood in that area. All of this without the need for a person to lie motionless in a massive cacophonous magnet, with greater portability, and for about a hundredth of the price of an MRI scanner (about £25,000 compared to £2,500,000).

The source and detector are placed on the scalp, so that the light received at the detector is reflected light following banana-shaped pathways

Picture credit: Tessari et al., 2017

“That sounds amazing! Sign me up!” I hear you say. However, I must put a little disclaimer out. There are reasons why fMRI is still the gold standard for functional brain imaging. As fNIRS relies on the measurement of light that gets back to the surface of the scalp after being in the brain, it can’t be used to measure activity from brain areas more than about 3 cm deep. This is being worked on by using cool ways of organising sources and detectors on the scalp. However, it is not thought that fNIRS will ever be able to produce a whole-brain map of brain activity. Also, as fNIRS is looking at the centimetre level, rather than millimetre, its spatial resolution and accuracy of location is limited in comparison to fMRI. Despite this, if the brain areas you’re interested in investigating are closer to the surface of the head, and not too teensy tiny, then fNIRS is a great technology to use.

So, what has this all got to do with babies? Well, fNIRS has one vice, one Achilles heel. Hair. Yes, this amazingly intelligent technology has such a primitive enemy. If your participants are blonde or bald, you’ll probably be fine. But anything deviating from this can block light from entering the head, and therefore weaken the light reaching the brain and eventually getting back to the detectors. However, do you know who has little to no hair? Babies. Plus, babies aren’t very good at lying still, particularly in a cacophonous magnet. This is why fNIRS is especially good for measuring brain activity in infants.

fNIRS is used to study a variety of topics related to infant development.  One of the most studied areas of infant psychology is language development. Minagawa-Kawai et al (2007) investigated how infants learn phonemes (the sound chunks that words are made up of). They used fNIRS to measure brain activation in Japanese 3 to 28-month-olds while they listened to different sounds. Infants listened to blocks of sounds that alternated between two phonemes (e.g. da and ba), and then other blocks that alternated between two different versions of the same phoneme (e.g. da and dha). In 3 to 11-month-olds, they found higher activation in a brain area responsible for handling language for both of these contrasts. So, this means that infants were treating ‘da’ and ‘ba’ and ‘dha’ as three different phonemes. However, 13 to 28-month-olds only had this higher activation when listening to the block of alternating ‘ba’ and ‘da’. This means that the older infants were treating ‘da’ and ‘dha’ as the same phoneme. This is consistent with behavioural studies showing that infants undergo ‘perceptual narrowing’, whereby over time they stop being able to discriminate between perceptual differences that are irrelevant for them. This has been related to why it’s much easier to be bilingual from birth if you have input from both languages, than it is to try to learn a second language later in life.

Another popular area of infant psychology is how infants perceive and understand objects. Wilcox et al (2012) used fNIRS to study the age at which infants began to understand shapes and colours of objects. They measured brain activation while infants saw objects move behind a screen and emerge at the other side. This study used a live presentation, made possible by the fact that fNIRS has no prerequisites for a testing environment except to turn the lights down a bit.


The shape change (left), colour change (middle), and no change (right) conditions of Wilcox et al. (2012). Each trial lasted 20 seconds, consisting of two 10 second cycles of the object moving from one side to the other (behind the occluder) and back again.

These objects were either the same when they appeared from behind the screen, or they had changed in shape or colour. They found heightened activation in the same area found in adult fMRI studies for only the shape change in 3 to 9-month olds, but for both shape and colour changes in the 11 to 12-month-olds. This confirms behavioural evidence that infants are surprised when the features of objects have changed, and that babies understand shape as an unchanging feature of an object before they understand colour in this way. This study shows how you can use findings from adult fMRI and infant behavioural studies to inform an infant fNIRS study, helping us learn how the brain’s complex visual and perceptual systems develop from infancy to adulthood.

There’s a lot more to learn if you wish to venture into the world of infant fNIRS research; it’s a fascinating area filled with untapped potential. fNIRS can help us to measure the brain activity of a hard-to-reach population (those pesky babies), enabling us to ask and answer questions about the development of language, vision, social understanding, and more! Questions being investigated in the Lancaster Babylab (where I am doing my PhD) include:

  • Do babies understand what pointing means?
  • Are bilingual babies better at discriminating between sounds?
  • Why do babies look at their parents when they are surprised?

And beyond this, the possibilities are endless!

If you are intrigued by fNIRS and want to learn more, I’d recommend review papers such as the one by Wilcox and Biondi (2015), and workshops such as the 3-day Birkbeck-UCL NIRS training course.


Minagawa-Kawai, Y., Mori, K., Naoi, N., & Kojima, S. (2007). Neural Attunement          Processes in Infants during the Acquisition of a Language-Specific Phonemic Contrast. Journal Of Neuroscience, 27(2), 315-321.

Otsuka, Y., Nakato, E., Kanazawa, S., Yamaguchi, M., Watanabe, S., & Kakigi, R.   (2007). Neural activation to upright and inverted faces in infants measured by near infrared spectroscopy. Neuroimage, 34(1), 399-406

Tessari, M., Malagoni, A., Vannini, M., & Zamboni, P. (2015). A novel device for non-invasive cerebral perfusion assessment. Veins And Lymphatics, 4(1).

Wilcox, T., Stubbs, J., Hirshkowitz, A., & Boas, D. (2012). Functional activation of the infant cortex during object processing. Neuroimage, 62(3), 1833-    1840.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s