Brain Connectivity Predicts Outcomes After Neonatal Injury, Study Finds
By Thomas R. Collins
May 5, 2022
Article In Brief
A study of children up to 30 months found that neonatal brain connectivity predicts outcomes after hypoxic-ischemic encephalopathy and congenital heart disease, revealing an important structure-function link in brain development. The findings could be a step toward a useful guide to early intervention in these common conditions.
Neonatal brain connectivity predicts outcomes after hypoxic-ischemic encephalopathy (HIE) and congenital heart disease (CHD), researchers reported in a new study in PLOS ONE.
The findings were seen in children at a young age—only as old as 30 months—but they could be a step toward a useful guide to early intervention in these common conditions, said Patrick McQuillen, MD, professor of pediatrics at the University of California, San Francisco, where he studies brain diseases in newborns.
While using brain connectivity as a predictive tool is the ultimate aim of their work, Dr. McQuillen said it was encouraging to see that, overall, the children had good outcomes given their conditions—a testament to the value of therapeutic hypothermia for HIE, or birth asphyxia.
“One way to put it would be, they did surprisingly well,” he said. For the congenital heart disease cohort, he said, “we could find some more areas that they performed worse in, but by and large they're doing OK. This is not a devastating brain injury that babies with congenital heart disease have. It's just that they have a lot of problems in a lot of different areas.”
Investigators performed diffusion MRI on 35 subjects with CHD and 62 with HIE and analyzed connectivity using graph theory metrics, which allows white matter tract connections to be described quantitatively. The patients were drawn from two single-center prospective cohort studies—Congenital Heart Disease MRI (CHDMRI) and Birth Asphyxia MRI (BAMRI)—but these studies didn't include a control cohort. The CHDMRI patients had brain MRIs performed pre- and post-operatively. The BAMRI patients had moderate to severe encephalopathy believed to be from hypoxia-ischemia and had MRIs performed in the first week after birth.
At 12 to 18 months—by which time neurodevelopmental testing was available for 42 HIE patients and 14 CHD patients—the HIE group scored within or above the normal range, researchers reported. The CHD group had lower language scores compared with the HIE group at this point. Still, about 93 percent of the CHD patients scored within or above the normal range in the cognitive domain, about 79 percent in the language domain and about 86 percent in the motor domain.
At 30 months—with testing available for 31 HIE patients and eight CHD patients—the HIE group scored within or above the normal range for all the domains. About 63 percent of the CHD patients scored in the normal range in the cognitive domain, 50 percent in the language domain, and 75 percent in the motor domain. The HIE group performed better than the CHD group in all domains.
To assess how well the networks of the brain are integrated, researchers used global efficiency, which essentially analyzes the balance of more local connections—like local roads within a city—and longer-term connections—similar to freeways. They found that this global efficiency was reduced in the CHD group compared with HIE (p = 0.03).
“We found distinct differences in how the brains are wired in children with congenital heart disease as compared with children with birth asphyxia,” Dr. McQuillen said.
Because the CHD group had the worse developmental scores in language, researchers examined pathways that have been associated with language in adults. They found connectivity in two areas—superior longitudinal fasciculus (SLF) tract 2 and SLF tract 3—to be linked with language scores. The greater the connectivity of SLF2, which connects the inferior parietal to the frontal lobe, the lower the expressive language scores were at 30 months. And the greater the connectivity of SLF3, which connects the supramarginal gyrus to the frontal lobe, the lower the overall language scores were at 12 to 18 months.
It might seem counterintuitive that greater connectivity might lead to worse outcomes, but it makes sense given the complex nature of brain connectivity and function, Dr. McQuillen said.
“In order to get normal language function, you have to have the orchestrated involvement of a number of different pathways,” he said. “So, any sort of imbalance can lead to dysfunction.”
Researchers are continuing to track the children as they age to see whether their connectivity patterns predict outcomes even later.
“Ultimately what we're trying to do is use the brain imaging, and the structural connectivity in particular, to really get—on a case-by-case basis—sort of personalized medicine,” that is, “a tool that we could use to predict later functional outcomes.”
There is particular interest in outcomes “at school age and beyond,” Dr. McQuillen said.
“So, we're going to continue to look at the neonatal imaging as we follow these kids and can test them doing much more sophisticated things, so we can continue to do the same sort of analysis,” he said. “We can use the neonatal imaging to identify patients for early intervention and maybe even for early specific interventions.”
Andras Jakab, MD, PhD, a research group leader at the Center for MR-Research at University Children's Hospital in Zürich, Switzerland, said the research shows the utility of the technology.
“This study provides convincing evidence for the value of advanced, noninvasive MRI techniques, such as diffusion MRI and connectomics, for the identification of infants at risk for neurodevelopmental impairments,” he said. “It stands in the line of previous works that found correlations between MRI markers and neurodevelopmental outcomes, revealing an important structure-function link in brain development.”
But he added, “It is important to understand that these correlations may only exist when larger groups of patients are compared. The question of how such MRI markers can be used to predict impaired neurodevelopment in the individual child remains open.”
The study also leaves questions unanswered, he said. “We know very little about why such wiring differences emerge between the two conditions. As the authors point out, one of the reasons for the differences in the brain wiring and outcomes could be the different treatment options—infants with birth asphyxia may benefit from the protective effects of therapeutic hypothermia,” Dr. Jakab said. “An interesting research question that still remains unexplored is if there are long-term differences in neurodevelopmental outcomes and how they are supported by the structural differences described by this article.”
Gabrielle deVeber, MD, a pediatric neurologist and professor of pediatrics at the Hospital for Sick Children in Toronto, noted how well the children fared. “Cardiac kids actually still do quite well even in this study,” she said. “The majority, 50 percent or above, scored normally in cognitive, language, or motor [domains] by 30 months. However, an important point is that outcomes may be even better than this report if the full spectrum of children with congenital heart disease are studied and not just the ones who are sick newborns, in-patients, and are imaged in the newborn period,” meaning they are more likely have more severe versions of CHD.
She said the study also shows that the presence of intact connectivity may not be sufficient to ensure normal network function.
Dr. deVeber said future work using this approach could shed light on whether the heart forming abnormally is due to abnormal brain formation, or vice versa. “This study employs recent advanced imaging techniques—structural connectivity—that in future analyses of these existing study data may also shed light on the debate on ‘nature versus nurture’ in the brains of infants and children with CHD.”
Dr. McQuillen had no disclosures.