When should a newborn's brain MRI be done after a difficult birth?

The report of the American College of Obstetricians and Gynecologists' Task Force on Neonatal Encephalopathy, Neonatal Encephalopathy and Neurologic Outcome, Second Edition, published in Pediatrics in 2014, describes two scans rather than one. The first, between 24 and 96 hours of life, emphasizes diffusion and spectroscopic findings and is more useful for evaluating the timing of a cerebral injury, because diffusion abnormalities are most prominent in that window. The second, at day 10 of life or later, best shows the full extent of injury, because the complete picture often is not visible during the first week. Many hospitals perform only one scan, usually after rewarming from cooling therapy, which is why the date and time of the MRI is one of the first things a reviewer checks.

Can a brain MRI look normal even though the baby was injured?

Yes, and the timing of the scan is a common reason. Diffusion abnormalities do not stay abnormal. They fall, bottom out, and then drift back toward normal values in a process called pseudonormalization. In a 2012 study in Neurology, Dr. Nathalie Bednarek and colleagues found that in cooled newborns, diffusion measurements returned to apparently normal values after the tenth day of life, compared with 6 to 8 days in a comparison group that had not been cooled. A diffusion-weighted scan performed inside that recovery window can read as unremarkable in a brain that is genuinely injured. Conventional T1 and T2 images taken at the same late point usually still show the injury, which is one reason the ACOG Task Force recommends a later scan in addition to an early one rather than instead of it.

How well does a newborn brain MRI predict a child's future?

It predicts adverse outcomes well and reassuring outcomes less well. In a meta-analysis of 32 studies covering 860 infants with neonatal encephalopathy, published in Pediatrics in 2010 by Dr. Sudhin Thayyil and colleagues, conventional MRI performed during the neonatal period had a pooled sensitivity of 91% but a specificity of only 51% for predicting adverse neurodevelopmental outcome. In practical terms, most children who go on to significant impairment had an abnormal scan, but among children who went on to a normal outcome roughly half still had abnormal neonatal imaging, so an abnormal scan is a weak basis for ruling a good outcome out. The most accurate single quantitative measure in that analysis was the deep gray matter lactate to N-acetylaspartate ratio on MR spectroscopy, with pooled sensitivity of 82% and specificity of 95%. No newborn scan settles a child's future, which is why formal developmental follow-up matters.

Your Baby’s Brain MRI After a Difficult Birth — How to Read the Report and Why the Timing Matters

Q: What does a basal ganglia and thalamus pattern on a newborn MRI mean?

The basal ganglia-thalamus pattern, sometimes called the central or deep gray matter pattern, involves injury to structures deep in the brain along with the perirolandic cortex. In the 2008 Radiographics review by Dr. Benjamin Huang and Dr. Mauricio Castillo, severe hypoxia-ischemia in term neonates preferentially damages the deep gray matter, with perirolandic involvement, while less profound insults instead produce parasagittal watershed territory infarcts. In the French national LyTONEPAL cohort of 520 encephalopathic newborns, reported by Dr. Julie Beck and colleagues in Children in 2022, 52.4% had a brain injury; the basal ganglia and thalamus were involved in 33.8%, white matter in 33.5%, and cortex in 25.6%. The pattern is one piece of evidence about the nature of an event, not a timestamp, and it is read alongside the fetal monitoring strip, cord gases, and the delivery timeline.

Q: Does an abnormal newborn brain MRI mean there was medical negligence?

No. An MRI describes an injury; it does not establish who caused it or whether anyone was careless. Some newborn brain injuries occur before labor begins, arise from infection or a stroke, or reflect a genetic or metabolic condition, and no obstetric decision would have changed them. A birth injury review uses the imaging as one input alongside the fetal heart rate tracing, the cord blood gas values, the Apgar scores, the nursing notes, and the minute-by-minute delivery timeline. The questions that decide these cases are whether warning signs in the record were recognized, whether the response was timely, whether cooling therapy was offered within the six-hour window, and whether the recommended imaging was actually performed and correctly read.

Somewhere in the folder you brought home from the NICU there is a page of radiology prose that almost nobody explains to parents. It uses words like restricted diffusion, lentiform nuclei, loss of the normal T1 hyperintensity of the posterior limb of the internal capsule. A neonatologist may have summarized it in one sentence at the bedside. That page is often the single most consequential document in the chart — and the one parents are least equipped to question, including on the point that matters most, which is not what it says but what day the scan was taken.

This is a plain-English guide: the windows the national guidance specifies, why a scan done a little too late can read as falsely reassuring, what the injury patterns do and do not prove, and how the imaging is read alongside the delivery record.

Why Is Brain MRI the Test That Matters After a Difficult Birth?

Because it is the only test that shows the injury itself, rather than its effects. Apgar scores, cord blood gases, and a neurologic exam describe a baby in trouble; they cannot say which structures were hurt, how widely, or in what pattern. A 2025 consensus position in Pediatric Research by Dr. Abbot Laptook, Dr. Kasper Kyng, and colleagues for the Brain, Development and Imaging section of the European Society for Paediatric Research states it plainly: MRI is the gold standard for characterizing brain injury in neonatal encephalopathy, with diffusion-weighted imaging enabling early detection and MR spectroscopy providing robust prognostic indicators of two-year outcomes.

What MRI is not is interchangeable with the head ultrasound many babies get first. In a 2014 meta-analysis by Dr. Bin Lin and colleagues at Fudan University, cranial ultrasound had a pooled sensitivity of 68.3% and specificity of 45.9%, against 91.3% sensitivity for T1- and T2-weighted MRI. A normal head ultrasound in the first days of life rules very little out.

When Should the MRI Have Been Done?

National guidance describes two scans, not one, and the two answer different questions. The report of the ACOG Task Force on Neonatal Encephalopathy — Neonatal Encephalopathy and Neurologic Outcome, Second Edition, developed with the American Academy of Pediatrics and published in Pediatrics in 2014 — recommends an MRI or MR spectroscopy study between 24 and 96 hours of life, emphasizing diffusion and spectroscopic findings, to assist with management and with evaluating the timing of a cerebral injury; and a second study at day 10 of life or later to delineate the full nature and extent of injury.

The split is physiologic. Diffusion abnormalities are most prominent between 24 and 96 hours of life, so the early scan is where evidence about when an insult occurred is most visible. Conventional T1- and T2-weighted abnormalities become most evident only after about a week, so the full footprint of the damage is a later picture. One study cannot be optimal for both.

In practice, most babies get one scan, not two — usually shortly after rewarming from cooling therapy. That is a resource reality, not a scandal. But it means the single scan your baby received was well placed for one question and may be poorly placed for the other. The date stamped on the MRI is one of the first things a reviewer checks, before reading a word of the findings.

Why Can a Late MRI Look Falsely Reassuring?

Because diffusion abnormalities do not stay abnormal — they fall, bottom out, and then drift back toward normal values. Radiologists call that return to apparently normal readings pseudonormalization, and it is one of the most important and least explained facts in this area of medicine. A diffusion-weighted scan taken inside the recovery window can read as unremarkable in a brain that is genuinely and permanently injured.

Cooling therapy moves the goalposts. In a 2012 study in Neurology, Dr. Nathalie Bednarek and colleagues reviewed the MRIs of newborns treated with whole-body therapeutic hypothermia and found diffusion measurements reduced in every injured infant during the first 7 days of life, but returned to normal values after the tenth day — compared with 6 to 8 days in a previously published group that had not been cooled. Cooling slows the evolution of the imaging findings along with everything else.

So a report saying “no restricted diffusion” means something quite different on day 3 than on day 9. The conventional T1 and T2 sequences from a late scan usually still show the injury — which is why the guidance asks for a late study in addition to an early one, and why a lone late scan reported as normal deserves a closer look rather than relief.

What Do the Injury Patterns Mean?

Pattern is the part of the report that speaks to the nature of what happened. In the widely cited 2008 Radiographics review by Dr. Benjamin Huang and Dr. Mauricio Castillo, severe hypoxia-ischemia in term newborns preferentially damages the deep gray matter — the basal ganglia and thalamus — with involvement of the perirolandic cortex. Less profound insults instead produce parasagittal watershed territory infarcts, in the border zones between major arterial territories; in preterm newborns, that same category tends to produce intraventricular hemorrhage and periventricular white matter injury.

Both patterns are common. In the French national LyTONEPAL cohort of 520 encephalopathic newborns of at least 36 weeks, described by Dr. Julie Beck and colleagues in Children in 2022, 52.4% had an identifiable brain injury; the basal ganglia and thalamus were involved in 33.8%, white matter in 33.5%, and cortex in 25.6%.

What a pattern is not is a clock. It is evidence about the character of an oxygen-deprivation event — abrupt and severe versus prolonged and partial — read together with the fetal heart rate tracing, the cord blood gas values, the Apgar scores, the placental pathology, and the delivery timeline. An abrupt, severe pattern in a baby whose monitoring was reassuring until a sudden catastrophic change tells a coherent story. Prolonged partial injury in a baby with hours of documented late decelerations tells a different one.

What Do the Words in the Report Actually Mean?

A neonatal MRI report is written for other clinicians. A short glossary of the terms parents encounter most often:

Restricted diffusion / reduced ADC. Water molecules are moving less freely than they should, typically because injured cells have swollen. ADC is the apparent diffusion coefficient, the number behind the picture. Low ADC in the first days signals acute injury.
T1 and T2 signal change. The conventional sequences. Abnormal brightness or darkness in a structure that should look uniform indicates tissue injury, clearest after roughly the first week.
PLIC. The posterior limb of the internal capsule, a tract normally bright on T1 in a term newborn. Loss of that brightness is a long-recognized abnormal sign, though the Thayyil meta-analysis below found its standalone predictive power poor.
Lentiform nuclei, putamen, globus pallidus, thalami. The deep gray matter structures — named involvement here is the basal ganglia–thalamus pattern.
Parasagittal / watershed. The cortex and white matter beside the midline, at the boundary between major arterial supplies — the region injured by prolonged partial oxygen deprivation.
Punctate white matter injury. Small scattered spots, common and variable in significance.
Lac/NAA on MR spectroscopy. The ratio of lactate to N-acetylaspartate in deep gray matter. Rising lactate with falling NAA indicates injured tissue — the strongest single quantitative predictor in the evidence.

How Well Does the MRI Predict Your Child’s Future?

It predicts bad outcomes far better than it predicts good ones. In a meta-analysis of 32 studies covering 860 infants with neonatal encephalopathy, published in Pediatrics in 2010 by Dr. Sudhin Thayyil and colleagues at University College London, conventional MRI in the neonatal period had a pooled sensitivity of 91% (95% CI 87–94%) but a specificity of only 51% (95% CI 45–58%) for adverse neurodevelopmental outcome at a year or more. Deep gray matter Lac/NAA on spectroscopy did markedly better on specificity: sensitivity 82%, specificity 95%.

Read that specificity figure carefully, because it cuts in the direction parents rarely hear. A specificity of 51% means that among the children who went on to a normal outcome, roughly half still had abnormal findings on their neonatal MRI. An abnormal scan, in other words, is a poor tool for ruling a good outcome out — which is why structured developmental follow-up across the first two years matters more than any single image, and why a life care plan is built from a child's actual functional trajectory rather than a radiology impression.

How these scans get scored also still varies by hospital. A 2025 consensus statement in Pediatric Neurology from the Newborn Brain Health Working Group of the Canadian Neonatal Network, led by Dr. Khorshid Mohammad, was written precisely because considerable variability exists among institutions in MRI timing, protocols, injury classification, and scoring systems — two radiologists at two hospitals can describe the same brain in materially different language.

When Do MRI Findings Point to Possible Negligence?

On their own, they usually don't. An MRI describes an injury; it does not identify a cause or assign responsibility. Newborn brain injuries can arise before labor begins, from infection, from a perinatal stroke, or from a genetic or metabolic condition, and in many cases no obstetric decision would have altered the outcome. A responsible review treats the imaging as one input among several, and asks narrower questions:

Was an MRI obtained at all, and on what day of life — a date that can answer the question being asked of it?
Where the only scan was late, was pseudonormalized diffusion considered, or was “no restricted diffusion” allowed to stand as reassurance?
Does the imaging pattern fit the clinical story in the monitoring record, the cord gases, and the resuscitation notes?
Was the baby evaluated for cooling therapy within the six-hour window when criteria were met?
Were the findings acted on — neurology referral, EEG monitoring for seizures, developmental follow-up — or filed without a plan?

What a Trained Reader Looks For

Imaging cases turn on the fit between two records created days apart by people who never spoke to each other. When Herb Borroto, M.D., J.D., the firm's Medical-Legal Expert, reviews a file like this, the reading is deliberately paired:

The MRI report with its timestamp attached. Day of life, whether cooling had ended, which sequences ran, whether spectroscopy was done at all.
The images, not only the impression. A late scan called unremarkable on diffusion may carry conventional-sequence findings the impression line underweighted.
The delivery record beside the pattern. The tracing, the timing of any sentinel event such as a uterine rupture or cord prolapse, and the decision-to-delivery interval.
The placental pathology. Often the clearest evidence of whether a process began before labor or during it.

Alex Alvarez, the firm's Managing Partner and a Board Certified Civil Trial Lawyer (The Florida Bar), presents these cases the way the medicine reads them. The disagreement is rarely about whether a brain is injured — the pictures settle that. It is about whether the pattern and the paper trail together describe an unforeseeable catastrophe or a deterioration that was documented, watched, and not acted on.

If Your Baby Had an Abnormal Brain MRI After Birth

If your child was diagnosed with HIE or birth asphyxia, treated with cooling, or later diagnosed with cerebral palsy, the imaging is only part of what needs reading — the parts you have probably never seen are the fetal monitoring strips, the nursing notes, and the placental pathology. A free, confidential case review can help you understand what the records show and whether your state's filing deadlines remain open. See also our guides to cooling therapy, cord blood gas results, and the birth injury statute of limitations.

Free case review. No Fees Unless We Recover Money for You.

Sources

American College of Obstetricians and Gynecologists’ Task Force on Neonatal Encephalopathy, with the American Academy of Pediatrics — “Neonatal Encephalopathy and Neurologic Outcome, Second Edition,” Pediatrics 133(5):e1482 (2014), on the recommendation of an MRI or MR spectroscopy study between 24 and 96 hours of life for evaluation of diffusion and spectroscopic abnormalities and the timing of cerebral injury, and a second study at day 10 of life or later to delineate the full nature and extent of injury. publications.aap.org
Bednarek N, Mathur A, Inder T, Wilkinson J, Neil J, Shimony J — “Impact of therapeutic hypothermia on MRI diffusion changes in neonatal encephalopathy,” Neurology 78(18):1420–7 (2012), on reduced mean diffusivity in all injured infants within the first 7 days of life and pseudonormalization after the tenth day with cooling versus 6 to 8 days in the uncooled control group. pubmed.ncbi.nlm.nih.gov
Huang BY, Castillo M — “Hypoxic-ischemic brain injury: imaging findings from birth to adulthood,” Radiographics 28(2):417–39 (2008), on severe insults preferentially damaging deep gray matter with perirolandic involvement in term neonates, and less profound insults producing parasagittal watershed territory infarcts in term neonates and periventricular white matter injury in preterm neonates. pubmed.ncbi.nlm.nih.gov
Thayyil S, Chandrasekaran M, Taylor A, et al. — “Cerebral magnetic resonance biomarkers in neonatal encephalopathy: a meta-analysis,” Pediatrics 125(2):e382–95 (2010), on 32 studies and 860 infants, conventional neonatal MRI pooled sensitivity of 91% and specificity of 51%, deep gray matter Lac/NAA sensitivity of 82% and specificity of 95%, and poor discriminatory power of the PLIC sign and brain-water ADC. pubmed.ncbi.nlm.nih.gov
Beck J, Loron G, Ancel PY, et al. — “An Updated Overview of MRI Injuries in Neonatal Encephalopathy: LyTONEPAL Cohort,” Children 9(4):561 (2022), on 520 included newborns, brain injury in 52.4%, and involvement of the basal ganglia and thalamus in 33.8%, white matter in 33.5%, and cortex in 25.6%. pubmed.ncbi.nlm.nih.gov
Laptook A, Garvey AA, Adams C, et al., for the Brain, Development and Imaging section of the European Society for Paediatric Research — “Magnetic resonance imaging and spectroscopy in neonatal encephalopathy: current consensus position and future opportunities,” Pediatric Research (2025), on MRI as the gold standard for characterizing brain injury in neonatal encephalopathy, diffusion-weighted imaging for early detection, and basal ganglia and thalamus Lac/NAA as a robust prognostic indicator of two-year neurodevelopmental outcomes. pubmed.ncbi.nlm.nih.gov
Mohammad K, Reddy Gurram Venkata SK, Wintermark P, et al., Newborn Brain Health Working Group of the Canadian Neonatal Network — “Consensus Approach for Standardization of the Timing of Brain Magnetic Resonance Imaging and Classification of Brain Injury in Neonates With Neonatal Encephalopathy/Hypoxic-Ischemic Encephalopathy: A Canadian Perspective,” Pediatric Neurology 166:16–31 (2025), on considerable variability among institutions in MRI timing, protocols, injury classification, and scoring systems. pubmed.ncbi.nlm.nih.gov
Lin B, Zhang P, Cheng G, Zhou W, Wang L — “Meta-analysis of prognostic tests in neonates over 35-week gestational age with hypoxic-ischemic encephalopathy,” Zhonghua Yi Xue Za Zhi 94(2):115–21 (2014), on pooled cranial ultrasound sensitivity of 68.3% and specificity of 45.9% compared with T1/T2-weighted MRI sensitivity of 91.3%. pubmed.ncbi.nlm.nih.gov