Skip secondary menu

Workshop on Near Infrared Spectroscopy (NIRS) for Infant Cerebral Function Monitoring

Summary of NINDS and NICHD Workshop on Near Infrared Spectroscopy (NIRS) For Infant Cerebral Function Monitoring

May 18,1999
Neuroscience Center, NIH

Table of Contents


On May 18, 1999, the NINDS and the NICHD held a workshop to discuss the use of near infrared spectroscopy (NIRS) for cerebral function monitoring in infants. Near infrared spectroscopy is a technology using a beam of light in the near infrared range (700-1000 mm) which passes through brain tissue, and measurements of the absorption and scattering of the photons are made. The consensus of the participants was that the technology has now been developed sufficiently that NIRS can be used to quantitatively measure cerebral oxygenation, as well as cerebral blood volume, continuously at the bedside. It can also be used to monitor cerebral blood flow (CBF), functional activation, and regional oxygenation at the bedside, although not continuously.

The primary question for discussion was whether sufficient progress has been made in NIRS to undertake clinical validation studies.


Derangements in cerebral blood flow, and as a consequence, oxygen delivery, are considered to be critical in the genesis of much of the neuropathology caused by insults during the perinatal period. Major conditions associated with perinatal brain injury include hypoxic ischemic injury and hemorrhagic ischemic cerebral injury for term infants, and intraventricular hemorrhage (IVH), and periventricular white matter injury in preterm infants. If we apply NIRS to the study of neonatal neurological problems, we may increase our understanding of mechanisms which contribute to perinatal brain injury.

Applications of NIRS

Hypoxic ischemic injury, which occurs in about .5 to .75 per thousand deliveries, may include parasagittal injury, involvement of the basal ganglia, white matter, and focal or multifocal cerebral injury. Clinical factors which indicate high risk for these injuries are depression at birth, requirement for resuscitation, and severe fetal academia (cord pH <7.0). However, these markers are not highly specific, and most children who have them do not have brain damage. Using NIRS in infants in whom these risk factors are present, to determine whether there are changes in blood flow or volume, or oxygen saturation, may allow identification of which high-risk infants are likely to progress to hypoxic-ischemic encephalopathy.

Treatment with novel strategies such as hypothermia or antioxidants aimed at prevention of ongoing injury following hypoxia-ischemia are now being explored. A cerebral function monitor using NIRS would be extremely helpful in determining the effect of protective strategies such as these on variables which can be measured. Additional potential uses for NIRS include focal cerebral infarction, and subdural hemorrhage, where identifying changes in blood flow and oxygenation at the time of clinical symptoms (usually apnea and/or seizures) may allow for intervention to prevent ongoing injury. In subdural hematomas, conventional imaging often does not show obvious changes within the brain, yet seizures occur.

Several invasive procedures, such as ECMO for respiratory failure and cardiac surgery under hypothermic bypass, could benefit from using NIRS to monitor brain function during and after the procedure. NIRS may also be applied to determine the mechanism of brain injury due to infection.

The very small premature infant is at high risk for brain injury. Intraventricular hemorrhages may be related to elevations in venous pressure. There may be changes in blood flow, volume, or oxygen saturation which precede the evolution of white matter injury leading to cystic periventricular leukomalacia (PVL), or to nonobstructive ventriculomegaly, or to hyperechogenicity of the white matter. NIRS may enable us to find out what changes distinguish those who progress to overt injury such as cyst formation vs. those where there is resolution of the lesions.

Regional imaging of oxygenation will be extremely useful and is feasible with NIRS. However, whereas oxygenation can be monitored continuously, regional imaging and blood flow monitoring are more difficult and need to be performed by trained personnel. NIRS can now only detect relatively large changes in metabolism.

The pathophysiology of hypoxic injury indicates a biphasic pattern of response over time, and the second wave of energy depletion may be even more damaging than the first. Excitotoxic injury plays a major role. NIRS may be useful in detection of risk while there is still time for prevention of injury. A decision must be made as to the timing of monitoring the infant in order to track the effects of injury.

Monitoring of cerebral blood flow using NIRS may be useful in conjunction with a metabolic marker which correlates with the cellular metabolic state. PVL results from preterm insult, whereas basal ganglia lesions are more common in the term infant. Watershed damage in the parasagital area occurs with moderate insult, and brainstem injury with profound anoxic insult. Regional vulnerability depends on substrate delivery, metabolism of cells, and the types of cells undergoing development at the time of injury.

Potential applications of a NIRS instrument in the neonate include as a diagnostic device, a treatment monitor, and as a prognostic device. For prognosis, regional brain measurements will be needed because of selective vulnerability of the brain to insult. Possible target populations including term infants with respiratory failure or pulmonary hypertension, apnea, seizures, perinatal infection, and cardiac conditions. Extremely preterm infants on those requiring high frequency ventilation would also be appropriate subjects.

The available types of NIRS monitors utilize continuous wave, phase modulation, or time resolved techniques. The continuous wave instruments have limitations and do not appear to be sufficiently effective for measuring cerebral oxygen saturation. Phase modulation multi wavelength systems calibrate more accurately in critical saturation ranges. Time resolved systems also calibrate well.

Regional imaging can be accomplished by measuring the presence or absence of activity after a simple stimulus such as moving the infant's foot. CBF, functional activation, and regional oxygenation can be imaged at the bedside, but require minimal manipulation which can be performed intermittently. Normal infants responding to touch show activation of the somatosensory cortex, with regional increases in 02. Cerebral oxygenation and total hemoglobin can be determined continuously at the bedside without the need to touch the infant, once the probes have been placed on the infants head.

In addition, NIRS can be used to detect changes in the light scattering. The effusion of potassium during hypoxic events causes a light scattering effect. Light scattering changes may precede an abrupt increase in brain water content which may mark the onset of an ischemic depolarization and indicate that the energy level is low enough that ionic gradients cannot be maintained. Tissue injury may be reflected in changes in light scattering caused by lipid and water content shifts. The changes in scattering and absorption coefficients resulting from brain insult can be determined by NIRS techniques.

It would be useful to look at NIRS in conjunction with the EEG., which measures global and regional electrical brain activity. At this time, it is premature to couple EEG with NIRS in a multimodal monitor, but this should be an option in the future. EEG has high sensitivity but low specificity, and cannot distinguish cerebral dysfunction from permanent damage.


In order to determine the benefits of NIRS as a continuous bedside monitor in the neonate, several steps are needed: Continuation of work on instrument development and refinement, standardization of NIRS measurements, animal and infant pilot studies, and finally clinical validation trials.

Instrument Development:

In the last few years, the technology of NIRS has been rapidly developing. Probes for infant heads have been developed, but improvements are still needed. It is likely that within the year, devices with reasonably workable probes for frequency domain instruments will be available. The probe must be easily applied to a neonate, and should not carry any high voltage signals or electrical currents. Non-metallic probes should be available for potential MRI or NIRS studies.

Software controlling data acquisition must be easy to use, reliable and stable. Animal models must be incorporated into ongoing instrument development programs. Instruments must be relatively low cost, and easy to operate. The instrument should be modular and interfaceable to other measuring devices.

Standardization of Measurements:

In order to move forward to clinical validation, comparison of available NIRS devices is necessary. Minimum confidence levels for established parameters will need to be rigorously defined. The range of normal values will need to be established for measured parameters. By multi-device comparisons using animal models, confidence in measurements can be established.

Pilot Data:

In order to determine which instruments to use for further trials, a few devices which operate on fundamentally different principles can be compared in animals or in a pilot clinical study with a short term outcome such as seizures or MRI findings. In an animal on heart/lung bypass, complete control over blood flow and oxygenation can be attained. The exact determination of how great a change in oxygenation is necessary to be predictive of outcome will probably have to be approximated from animal models.

Although both animal models and infants have been studied with NIRS, once the best available device (s) are chosen, more pilot data from animal and clinical studies are needed before validation trials can be designed. Optional placement of optodes for the best signal needs to be established. More data is needed on the ranges expected for healthy vs sick neonates. The frequency and severity of undetected desaturation events in the NICU needs to be estimated from observational data, and the circumstances under which these occur need to be understood.

Using animal models, studies can focus on how measures might be expected to vary under various pathological conditions. Questions for regional imaging can be addressed, including what areas can be imaged, how well, and how can the data be analyzed quantitatively. The possibility of diagnosing disturbances of autoregulation of CBF can be explored.

Clinical Validation Trials:

Although MRS is extremely useful to look at spectroscopic changes in the brain, in order to determine metabolic activity and neuronal integrity, further studies are still needed comparing data obtained after insult with long term outcome. Thus there is no "gold standard" technology with which NIRS can be compared. The only way to perform clinical validation studies to demonstrate that NIRS measurements are highly predictive of outcome will be to actually carry out the outcome studies, following infants to at least age two or three years. Evidence is needed that decreased oxygenation, regional or global, decreased functional activation, and decreased ability to autoregulate blood flow, are statistically associated with poor clinical outcome as tested neurodevelopmentally.

After the optimal NIRS instrument is selected and shown to be reliable, blinded observational studies should be done in which infants are monitored with NIRS at least 80% of the time, and other relevant clinical factors such as 02 requirements, apneic spells, seizures, and infections are recorded. A clinical validation study will need an adequate number of easily identified babies, so will need to focus on a prevalent problem seen in a NICU before severe brain injury or death occurs. Blinding will be necessary to prevent intervention on the basis of NIRS readings which might alter the outcome. Examples are babies who are near-term in hypoxic pulmonary failure, premature infants 32 weeks or more on high frequency ventilation, and infants with severe congenital heart disease. Short-term monitoring of the preterm infant may be difficult to relate to outcomes because of the large number of contributing variables. Monitoring will need to be continuous and last at least 2 weeks to cover the most vulnerable postpartum period.

Intermediate outcomes which should be correlated with NIRS include seizures, IVH, cystic lesions on ultrasound, MRI findings, and survival and time to discharge. Long term outcome includes motor and developmental level. These studies are needed to determine the predictive value of NIRS: can it identify infants at risk for poor outcome, at a time when there is the potential for intervention, by quantitatively monitoring cerebral oxygenation, CBF and CBV, and by measuring scattering changes and functional activation responses.? The range and duration of changes in these parameters which lead to reversible or irreversible damage must be determined. If this can be done, NIRS will be of major clinical benefit in the management and development of treatment strategies for prevention of long term neurological disability in high-risk infants.


Dr. Deborah Hirtz
Program Director -NINDS, NIH

Dr. Peter A. Bandettini
Laboratory of Brain and Cognition
National Institute of Mental Health
National Institutes of Health

Dr. Robert Baughman
Director, NINDS Division of Fundamental
Neuroscience & Developmental Disorders
NIH Neuroscience Center

Dr. Ben Barbieri
ISS Inc.

Dr. Jack Belliveau
Massachusetts General Hospital, NMR Center

Dr. David Benaron

Dr. David Boas
Electro-Optics Technology Center
Tufts University

Dr. Britton Chance
Professor Emeritus
Department of Biochemistry and Biophysics
University of Pennsylvania

Dr. Mark Cope
University College London

Dr. Adre du Plessis
Neurology Department
Children's Hospital

Dr. Donna Ferriero
Department of Pediatric Neurology
University of California San Francisco

Dr. Gerald Fischbach
Director -NINDS, NIH

Dr. Amir H. Gandjbakhche
Laboratory of Integrative and Medical Biophysics

Dr. Gorm Greisen
Dept. Neonatology

Dr. Susan Hintz
Division of Neonatology
Stanford University

Dr. Michael F. Huerta
Associate Director
Division of Neuroscience and Basic Behavioral Science

Ms. Margaret Jacobs
Program Director , NINDS,NIH

Dr. Ken Kaufmann
Hamamatsu Corporation

Dr. Stephen Koslow
Associate Director, NIMH,NIH

Dr. C. Dean Kurth
Children's Hospital of Philadephia
Department of Anesthesiology

Mr. Richard Moberg
Moberg Research, Inc.

Dr. Karin Nelson
Acting Chief Neuroepidemiology Branch, NINDS

Dr. Shoko Nioka
Department of Biochemistry and Biophysics
University of Pennsylvania

Dr. Edward Novotny
Yale University School of Medicine
Dept. of Pediatrics

Dr. Eugene Oliver
Program Director NINDS,NIH

Dr. LuAnn Papile
Professor of Pediatrics & Obstetrics and Gynecology
Department of Pediatrics
Division of Neonatology
University of New Mexico School of Medicine

Dr. Jeffrey Perlman
UT SW Medical Center Dallas
Department of Pediatrics

Dr. Kevin J. Quinn
Chief, Cognitive Neuroscience Program

Dr. Chandra Ramamoorlty
Department of Anesthesiology
University of Washington

Dr. Nimmi Ramanujan
Department of Biochemistry and Biophysics
University of Pennsylvania

Dr. Mark Scher
Department of Pediatric Neurology
University Hospital of Cleveland

Dr. Robert Sclabassi
Department of Neurological Surgery
Presbyterian University Hospital

Dr. Philip Sheridan

Dr. Janet Soul
Instructor in Neurology
Dept. of Neurology
Children's Hospital

Dr. Giovanna Spinella
Program Director -NINDS, NIH

Dr. Miljan Stankovic
Winthrop University Hospital

Dr. Joseph Volpe
Bronson Crothers Professor of Neurology
Department of Neurology
Children's Hospital

Mr. Don Wallace
ISS, Inc.

Dr. Edward B. (Ted) Weiler
Director of Special Projects
Olympic Medical

Dr. David A. Wilson
Department of Anesthesiology
Johns Hopkins University

Dr. Linda Wright
Center for Research for Mothers and Children

Dr. Thomas A. Zeffiro
Research Bldg.
Georgetown University

Last updated April 18, 2011