Neonates and children who have survived critical illness severe enough to require extracorporeal membrane oxygenation (ECMO) are at risk for neurologic insults, neurodevelopmental delays, worsening of underlying medical conditions, and development of new medical comorbidities. Structured neurodevelopmental follow-up is recommended for early identification and prompt interventions of any neurodevelopmental delays. Even children who initially survive this critical illness without new medical or neurologic deficits remain at risk of developing new morbidities/delays at least through adolescence, highlighting the importance of structured follow-up by personnel knowledgeable in the sequelae of critical illness and ECMO. Structured follow-up should be multifaceted, beginning predischarge and continuing as a coordinated effort after discharge through adolescence. Predischarge efforts should consist of medical and neurologic evaluations, family education, and co-ordination of long-term ECMO care. After discharge, programs should recommend a compilation of pediatric care, disease-specific care for underlying or acquired conditions, structured ECMO/neurodevelopmental care including school performance, parental education, and support. Institutionally, regionally, and internationally available resources will impact the design of individual center's follow-up program. Additionally, neurodevelopmental testing will need to be culturally and lingually appropriate for centers' populations. Thus, ECMO centers should adapt follow-up program to their specific populations and resources with the predischarge and postdischarge components described here.

No abstract

Transcranial magnetic stimulation (TMS) is a newer noninvasive language mapping tool that is safe and well-tolerated by children. We examined the accuracy of TMS-derived language maps in a clinical cohort by comparing it against functional magnetic resonance imaging (MRI)–derived language map. The number of TMS-induced speech disruptions and the volume of activation during functional MRI tasks were localized to Brodmann areas for each modality in 40 patients with epilepsy or brain tumor. We examined the concordance between TMS- and functional MRI–derived language maps by deriving statistical performance metrics for TMS including sensitivity, specificity, accuracy, and diagnostic odds ratio. Brodmann areas 6, 44, and 9 in the frontal lobe and 22 and 40 in the temporal lobe were the most commonly identified language areas by both modalities. Overall accuracy of TMS compared to functional MRI in localizing language cortex was 71%, with a diagnostic odds ratio of 1.27 and higher sensitivity when identifying left hemisphere regions. TMS was more accurate in determining the dominant hemisphere for language with a diagnostic odds ratio of 6. This study is the first to examine the accuracy of the whole brain language map derived by TMS in the largest cohort examined to date. While this comparison against functional MRI confirmed that TMS reliably localizes cortical areas that are not essential for speech function, it demonstrated only slight concordance between TMS- and functional MRI–derived language areas. That the localization of specific language cortices by TMS demonstrated low accuracy reveals a potential need to use concordant tasks between the modalities and other avenues for further optimization of TMS parameters.