Exploring Cognitive Functions in Babies, Children & Adults with Near Infrared Spectroscopy


Your institution must subscribe to JoVE's Biology section to access this content.

Fill out the form below to receive a free trial or learn more about access:



Here we describe a data collection and data analysis method for functional Near Infrared Spectroscopy (fNIRS), a novel non-invasive brain imaging system used in cognitive neuroscience, particularly in studying child brain development. This method provides a universal standard of data acquisition and analysis vital to data interpretation and scientific discovery.

Cite this Article

Copy Citation | Download Citations | Reprints and Permissions

Shalinsky, M. H., Kovelman, I., Berens, M. S., Petitto, L. Exploring Cognitive Functions in Babies, Children & Adults with Near Infrared Spectroscopy. J. Vis. Exp. (29), e1268, doi:10.3791/1268 (2009).


An explosion of functional Near Infrared Spectroscopy (fNIRS) studies investigating cortical activation in relation to higher cognitive processes, such as language1,2,3,4,5,6,7,8,9,10, memory11, and attention12 is underway worldwide involving adults, children and infants 3,4,13,14,15,16,17,18,19 with typical and atypical cognition20,21,22. The contemporary challenge of using fNIRS for cognitive neuroscience is to achieve systematic analyses of data such that they are universally interpretable23,24,25,26, and thus may advance important scientific questions about the functional organization and neural systems underlying human higher cognition.

Existing neuroimaging technologies have either less robust temporal or spatial resolution. Event Related Potentials and Magneto Encephalography (ERP and MEG) have excellent temporal resolution, whereas Positron Emission Tomography and functional Magnetic Resonance Imaging (PET and fMRI) have better spatial resolution. Using non-ionizing wavelengths of light in the near-infrared range (700-1000 nm), where oxy-hemoglobin is preferentially absorbed by 680 nm and deoxy-hemoglobin is preferentially absorbed by 830 nm (e.g., indeed, the very wavelengths hardwired into the fNIRS Hitachi ETG-400 system illustrated here), fNIRS is well suited for studies of higher cognition because it has both good temporal resolution (~5s) without the use of radiation and good spatial resolution (~4 cm depth), and does not require participants to be in an enclosed structure27,28. Participants cortical activity can be assessed while comfortably seated in an ordinary chair (adults, children) or even seated in mom s lap (infants). Notably, NIRS is uniquely portable (the size of a desktop computer), virtually silent, and can tolerate a participants subtle movement. This is particularly outstanding for the neural study of human language, which necessarily has as one of its key components the movement of the mouth in speech production or the hands in sign language.

The way in which the hemodynamic response is localized is by an array of laser emitters and detectors. Emitters emit a known intensity of non-ionizing light while detectors detect the amount reflected back from the cortical surface. The closer together the optodes, the greater the spatial resolution, whereas the further apart the optodes, the greater depth of penetration. For the fNIRS Hitachi ETG-4000 system optimal penetration / resolution the optode array is set to 2cm.

Our goal is to demonstrate our method of acquiring and analyzing fNIRS data to help standardize the field and enable different fNIRS labs worldwide to have a common background.


Part 1: Prior to participant arriving to the lab

  1. Ensure that the room is free of extraneous articles that may be distracting to the participant.
  2. Set-up and load experimental protocol on the fNIRS Hitachi ETG-4000 system.
  3. Set-up your experimental paradigm. Experimental paradigms can be programmed with different presentation software, including Eprime, Presentation, Psyscope or a Matlab based psychology toolbox. Here we use Matlab based psychology toolbox.
  4. Timing is key for data analysis, thus the experimental paradigm must be perfectly timed with data collection. The fNIRS Hitachi ETG-4000 has triggering capabilities, allowing for the experimental paradigm to trigger the data collection or vice versa. Test triggering of presentation program from fNIRS Hitachi ETG-4000. Triggering can be done using parallel, serials, or USB ports. Here we show triggering via the parallel port.
  5. Prior to starting the fNIRS study it is important to conduct participant background screening. In the Petitto lab, we conduct background screening by having the participants or their parents fill out study-appropriate standardized questionnaires29.

Participant Arrives

  1. It is important to conduct the session and to treat the participants in a professional manner. The participant or the participants' parents/legal guardians must sign a consent form before the experiment begins. It is vital to thank the participant for their time in these important and exciting experiments.
  2. The participant is seated comfortably close to the fNIRS testing room. An infant participant may be seated on a parent's lap.

Part 2: Placing Optodes & Using the 10-20 system

Another component of the analysis method that enables consistent data interpretation is the standardization of fNIRS recording protocol. This entails optode placement, participant positioning, and triggering of stimulus presentation software. Both the accurate neuro-anatomical placement of probes and the confirmation of regions of interest (ROIs) are achieved by using the 10–20 system3,4,30. Further, stereotactic localization of the probe array was confirmed on the participant's skull by overlaying 3D tracking information from a Polhemus Fast trak system onto an anatomical MRI co-registration scan of the participant conducted with vitamin E capsules placed at each probe location3,4.Optimal participant positioning involved placing participants comfortably in a reclining chair, with fiber optics hanging loosely without contact with the body or chair.

  1. The following head measurements are taken with a tape measure and written down on the participants data sheet:
    • Nasion to Inion around
    • Nasion to Inions over top
    • Ear to Ear over top
  2. Surgical tape can be used to mark specific target locations. In this experiment we will mark Fp, T3/T4, F8/F7
  3. Optode arrays are placed on the participants' head with specific optodes anchored at 10-20 points as directed by the purposes of the experiment.

Part 4: Testing the Optode Array

  1. Introduction to the Hitachi ETG-4000 GUI Interface and probe testing.
  2. Testing signal: Once optodes are placed on the participants' scalp, the signal quality is tested. If an optode does not have a clear signal, researchers gently remove hair from the connection of the optode and the scalp. On occasion the optodes may need to be wiped with an alcohol swab.

Part 5: Running the experiment.

  1. At least two experimenters must be always present in the room; one observing the fNIRS Hitachi ETG-4000 real-time read out and the other observing the participant. Having a video camera focused on participant is highly recommended for post-hoc observations. An advantage of the fNIRS Hitachi ETG-4000 is that the video and fNIRS signal are synched and co-registered. A log containing all relevant information and files generated is kept.
  2. There are well-established methods of building experimental hemodynamic paradigms, namely Block design and Event-related designs. For a more complete description please see the recent review paper31.

Part 6: Analysis

Once all of the data have been collected, the participant is thanked for their time and willingness to participate and leaves the lab. As analysis is not done on the fNIRS Hitachi ETG-4000, as, instead, the data are exported to an analysis computer.

  1. Conversion from μV to hemoglobin concentrations. As raw attenuation values are collected in attenuation of laser strength (as measured in μV), these values must be converted to oxygenated and deoxygenated hemoglobin values. This is done using the modified Beer-Lambert equation.
  2. The application of the modified Beer-Lambert is conducted in two steps. Under the assumption that scattering is constant over the path length, first the attenuation for each wavelength (ΔAλ(t)) is calculated by comparing the optical density of light intensity during the task (Itask) to the calculated baseline of the signal (Ibaseline). The ΔA values for each wavelength and sampled time point (t) to solve the modified Beer-Lambert equation.
    Equation 1equation 1 Equation 2equation 2

λ1deoxy, λ1oxy, λ2deoxy and λ 2oxy are the constants for the extinction coefficients that measure the fraction of light lost to absorption per unit concentration distance in the tissue. The resultant Cdeoxy and Coxy values are the concentrations of deoxygenated and oxygenated hemoglobin for each t.

Part 7: Representative Results

Typical hemodynamic response results in several distinct characteristics. In the oxy-hemoglobin response, there is first a characteristic dip. This dip occurs as a region of neurons activates and depletes available oxygen. As blood flow increases, carrying oxygenated hemoglobin, the oxy-hemoglobin response rises rapidly above the initial baseline levels to a steady state level. When the region is no longer being activated, the oxy-hemoglobin response decays over the next 12-15 seconds and slowly drops back to baseline levels. There is occasionally an undershoot that occurs prior to the hemodynamic response returning to initial baseline levels.

Bad results are usually in the form of optodes not properly seated on the scalp or excessive movement. These types of noise – called, 'Flatling' – are evident in the signal as the microvolt values saturate and a number of different channels, both oxy- and deoxy- response move in a coordinated fashion.

DEMONSTRATION: Shaking the fiber optics.

Statistical Analyses: The extracted oxy and deoxy-hemoglobin values for each channel, for each participant and for each task can then be submitted to conventional statistical analysis, including t-tests, ANOVAs, correlations etc.

Subscription Required. Please recommend JoVE to your librarian.


In this study, we demonstrated the use of a novel, non-invasive fNIRS brain imaging technology to investigate human brain function in relation to human cognition and perception. fNIRS brain imaging may represent the future of non-invasive brain imaging, particularly with infant and child populations, that may one day be widely available in research labs, physicians' offices, and in the school systems allowing clinicians to apply basic scientific findings about the brain to their clinical practice.

Subscription Required. Please recommend JoVE to your librarian.


The authors have nothing to disclose.


This work was supported by grants to L.A.P. (P.I.):

National Institutes of Health R21 HD50558, awarded 2005-07; National

Institutes of Health R01 HD045822, awarded 2004-09; Dana Foundation Grant,

awarded 2004-06; Canadian Foundation for Innovation ("CFI" grant), awarded

2008-2012; The Ontario Research Fund Grant, awarded 2008-2012.


Name Company Catalog Number Comments
ETG-4000 Hitachi
Matlab Mathworks Psychology toolbox



  1. Quaresima, V. J. Biomed. Opt. 10, 11012-11012 (2005).
  2. Watanabe, E. Neurosci. Lett. 256, 49-52 (1998).
  3. Kovelman, I. NeuroImage. 39, 1457-1471 (2008).
  4. Kovelman, I. Brain and Language. (2008).
  5. Bortfeld, H. Developmental Neuropsychology. 34, 52-65 (2009).
  6. Petitto, L. A. The Cambridge Companion to Chomsky. Cambridge University Press. England. (2005).
  7. Berens, M. S. Society for Research in Child Development, (2009).
  8. White, K. S. Cognitive Neuroscience Society Annual Meeting, (2008).
  9. Dubins, M. Cognitive Neuroscience Conference, (2009).
  10. Dubins, M. H. Society for Research in Child Development. (2009).
  11. Dubins, M. H. NeuroImage. (2009).
  12. Ehlis, A. C. J. Biol. Psychol. 69, 315-331 (2005).
  13. Petitto, L. A. The Educated Brain. Fischer, K., Battro, A. Cambridge University Press. England. (2008).
  14. Pena, M. Proc Natl. Acad. Sci. U. S. A.. 100, 11702-11705 (2003).
  15. Baird, A. A. NeuroImage. 16, 1120-1125 (2002).
  16. Taga, G. Proc. Nat.l Acad. Sci. U. S. A. 100, 10722-10727 (2003).
  17. Wilcox, T. Dev. Science. 11, 361-370 (2008).
  18. Otsuka, Y. NeuroImage. 34, 399-406 (2007).
  19. Watanabe, H. NeuroImage. 43, 346-357 (2008).
  20. Kameyama, M. NeuroImage. 29, 172-184 (2006).
  21. Arai, H. Brain. Cogn.. 61, 189-194 (2006).
  22. Grignon, S. Cognitive and Behavioral Neurology. 21, 41-45 (2008).
  23. Boas, D. A. Neuroimage. 23, S275-S288 (2004).
  24. Aslin, R. N., Mehler, J. J. of Biomed. Opt.. 1-3 (2005).
  25. Plichta, M. M. NeuroImage. 35, 625-634 (2007).
  26. Schroeter, M. L. NeuroImage. 21, 283-290 (2004).
  27. Jobsis, F. F. Science. 198, 1264-1267 (1977).
  28. Villringer, A., Chance, B. Trends Neurosci. 20, 435-442 (1997).
  29. Kovelman, I. Bilingualism: Language & Cognition. 11, 203-223 (2008).
  30. Jasper, H. Electroenceph. Clin. Neurophysiol. 10, 370-371 (1958).
  31. Amaro, E. Brain Cogn. 60, 220-232 (2006).



  1. How much dŒs this equipment cost compared to a MEG or fMRI

    Posted by: Anonymous
    November 28, 2009 - 5:55 AM
  2. NIRS equipment is significantly cheaper than either MEG or fMRI (on the order of $100K for NIRS vs. $1million++ for MEG or fMRI).

    Posted by: Anonymous
    February 18, 2010 - 4:43 PM
  3. Hi Greg,

    As Michael responded you can get fNIRS machines for significantly cheaper then MRI. An added bonus is you do not need to have a full time tech as graduate & even undergraduate students can be taught how to properly handle the machine. And there is no fear of quenching.

    There are several companies along with Hitachi that sell these systems. The costs are usually directly correlated with the number of channels and other bells & whistles (e.g. Polhemus).

    Let me know if you need any more help!


    Posted by: Mark S.
    June 1, 2010 - 1:31 PM
  4. Awesome video! I wish my lab was so photogenic.

    Posted by: Anonymous
    February 18, 2010 - 4:48 PM
  5. Hi Sir/Mrs.,
    Just want to know what's the reliability and validity of Near Infrared Spectroscopy compared to fMRI and EEG in terms of spatial and temporal resolution.
    Also, is Near Infrared Spectroscopy useful in terms of the investigation of neural substrates(i.e.regions in hippocampus) involved in recognition memory? Would Near Infrared Spectroscopy have similar results as fMRI in terms of recognition memory?

    Thank You,

    Posted by: fahad a.
    June 1, 2010 - 11:54 AM
  6. Hi Fahad,

    fNIRS has better temporal resolution (10 samples / second) then fMRI (1 sample / ² seconds). The fNIRS also has a spatial resolution of ²-3cm. This means that the signal recorded by the fNIRS is typically from the cortical regions.

    In comparison to EEG fNIRS is slower however the spatial resolution is better. There are a number of companies as well as private labs that have & are developing combined EEG / fNIRS systems.
    For learning & memory paradigms the regions of interest are typically too deep and cannot be reliably recorded from using fNIRS.

    Please feel free to contact the two PIs of this project'
    Dr. Laura-Ann Petitto: www.utsc.utoronto.ca/~petitto/
    Dr. Ioulia Kovelman: http://sitemaker.umich.edu/childlanguage/home

    Posted by: Mark S.
    June 1, 2010 - 1:39 PM

Post a Question / Comment / Request

You must be signed in to post a comment. Please or create an account.

Usage Statistics