Impaired Trophic Interactions Between the Cerebellum and the Cerebrum Among Preterm Infants
Background. Advanced neuroimaging techniques have brought increasing recognition of cerebellar injury among premature infants. The developmental relationship between early brain injury and effects on the cerebrum and cerebellum remains unclear.
Objectives. To examine whether cerebral parenchymal brain lesions among preterm infants are associated with subsequent decreases in cerebellar volume and, conversely, whether primary cerebellar injury is associated with decreased cerebral brain volumes, with advanced, 3-dimensional, volumetric MRI at term gestational age equivalent.
Methods. Total cerebellar volumes and cerebellar gray and myelinated white matter volumes were determined through manual outlining for 74 preterm infants with unilateral periventricular hemorrhagic infarction (14 infants), bilateral diffuse periventricular leukomalacia (20 infants), cerebellar hemorrhage (10 infants), or normal term gestational age equivalent MRI findings (30 infants). Total brain and right/left cerebral and cerebellar hemispheric volumes were calculated.
Results. Unilateral cerebral brain injury was associated with significantly decreased volume of the contralateral cerebellar hemisphere. Conversely, unilateral primary cerebellar injury was associated with a contralateral decrease in supratentorial brain volume. Cerebellar gray matter and myelinated white matter volumes were reduced significantly not only among preterm infants with primary cerebellar hemorrhage but also among infants with cerebral parenchymal brain injury.
Conclusions. These data suggest strongly that both reduction in contralateral cerebellar volume with unilateral cerebral parenchymal injury and reduction in total cerebellar volume with bilateral cerebral lesions are related to trophic transsynaptic effects. Early-life cerebellar injury may contribute importantly to the high rates of cognitive, behavioral, and motor deficits reported for premature infants.
Cerebellar injury is increasingly recognized through advanced neonatal brain imaging as a complication of premature birth.1–3 Survivors of preterm birth demonstrate a constellation of long-term neurodevelopmental deficits, many of which are potentially referable to cerebellar injury, including impaired motor functions such as fine motor incoordination, ataxia, and impaired motor sequencing.4,5 Cerebellar injury has been implicated in cognitive, social, and behavioral dysfunction among older patients6–8 and may contribute to the 25% to 50% incidence of long-term cognitive, language, and behavioral dysfunction among formerly preterm infants.9–12 Currently, limited information is available regarding the nature and consequences of cerebellar injury among preterm infants.
Cerebellar atrophy may occur as a result of a primary acquired injury, such as hemorrhage or infarction.1 In the absence of direct cerebellar injury, impaired cerebellar growth may also ensue as a secondary effect related to damage in remote but connected areas of the brain (ie, diaschisis).13 Crossed cerebellar diaschisis is defined as a reduction in metabolism and blood flow in the cerebellar hemisphere contralateral to a cerebral cortical insult.14,15 The possibility of an effect of cerebral injury on the developing cerebellum among preterm infants has not been well studied or appreciated. Moreover, no study to date has elucidated the effects of primary cerebellar injury on subsequent cerebral cortical brain development. We hypothesized that cerebral parenchymal brain lesions among preterm infants would be associated with subsequent decreases in cerebellar volume and, conversely, primary cerebellar injury would be associated with decreased cerebral brain volume. To address these hypotheses, we used quantitative, 3-dimensional, volumetric MRI to study formerly preterm infants at term gestational age equivalent (GAE) (ie, gestational age plus postnatal age). In particular, for preterm infants with unilateral cerebral or cerebellar injuries, we sought to determine whether the subsequent cerebellar or cerebral volume, respectively, was altered. As a secondary objective, we examined whether cerebral and cerebellar lesions among preterm infants were associated with decreased volumes of cerebellar gray matter and myelinated cerebellar white matter at term GAE.
MRI studies for preterm infants (gestational age: <32 weeks) with a term GAE MRI study that demonstrated unilateral periventricular hemorrhagic infarction (PVHI), periventricular leukomalacia (PVL), or unilateral or bilateral cerebellar hemorrhage (CBH) and preterm infants with normal MRI studies (with not only the absence of focal lesions but also the absence of abnormal, diffuse, high-intensity T2 signal in the white matter) were selected from existing MRI data sets acquired during previous16 and ongoing unpublished, prospective, research studies involving preterm infants. All studies were performed with informed parental consent and were in accordance with the ethical standards of the institutional review board of the Brigham and Women’s Hospital. We excluded MRI scans for infants with known or suspected brain malformations, dysmorphic features, or congenital anomalies suggesting a genetic syndrome, metabolic disorder, or central nervous system infection. In addition, we excluded infants with MRI studies that showed combined cerebral (eg, PVL or PVHI) and cerebellar (eg, hemorrhagic intraparenchymal or extraparenchymal lesions) parenchymal lesions. Perinatal data, including birth weight, gestational age at birth, Apgar scores at 1 and 5 minutes, and gender, were collected from the infants’ medical records.
MRI Image Acquisition
MRI was performed with a 1.5-T General Electric Signa system (GE Medical Systems, Milwaukee, WI). The MRI data were acquired with 2 different imaging modes, ie, a coronal, 3-dimensional, Fourier transform, spoiled gradient recalled (SPGR) sequence (slice thickness: 1.5 mm; flip angle: 45°; repetition time: 35 milliseconds; echo time: 5 milliseconds; field of view: 18 cm; matrix: 256 × 256) and an axial or coronal, double-echo (proton-density and T2-weighted), spin-echo sequence (slice thickness: 3 mm; repetition time: 3000 milliseconds; echo times: 84 and 168 milliseconds; field of view: 18 cm; matrix: 256 × 256, interleaved acquisition). The voxel dimensions for the SPGR acquisition were 0.7 × 0.7 × 1.5 mm3.
MRI Image Processing
Images were analyzed on workstations (Sun Microsystems, Mountain View, CA) with an established, postacquisition, image-processing protocol that was described previously and validated by our group for cerebral tissue segmentation.16,17 The protocol is based on mathematical algorithms designed to reduce imaging noise, to align gradient-echo images and T2-weighted and proton-density, 3-dimensional data sets, to classify cerebral tissue according to signal intensity and anatomic localization, and to estimate the volume of each tissue class [ie, to compute absolute volumes of cerebral cortical gray matter, deep nuclear gray matter, unmyelinated and myelinated white matter, and cerebrospinal fluid (CSF)]. The total cerebral brain volume was calculated from the sum of the voxels representing all gray and white matter of the brain, excluding CSF and the cerebellum.
Total cerebellar volume, cerebellar gray matter and myelinated white matter volumes, and fourth ventricle CSF volume were measured through manual outlining on the original, coronal, SPGR images with Slicer software (www.slicer.org) (Fig 1). The vermis was included as part of our total cerebellar volume measurements. The small amount of unmyelinated white matter present in the hemispheres of the immature cerebellum could not be quantified in the term GAE MRI studies because of the inherent lack of signal contrast in the tissues at that age. Therefore, we elected to include the small amount of unmyelinated cerebellar white matter that might be present in the cerebellar hemispheres as part of our gray matter volume measurements. All manual outlining of the total cerebellar volume and cerebellar segmentations were performed by a single investigator (C.L.). The intra-rater reliability coefficients (10 MRI scans) for total cerebellar volume, cerebellar gray matter, and myelinated white matter were α = .96, α = .95, and α = .93, respectively. The investigators who performed cerebral volume measurements (J.S.S) and cerebellar measurements (C.L.) were blinded to each other’s measurements.
To divide the cerebellum and cerebrum into right and left hemispheres, one investigator (C.L.) defined an “ideal” midsagittal plane as a virtual geometric plane about which the brain demonstrates maximal bilateral symmetry (Fig 2). We specifically selected the ideal midsagittal plane to ensure that the vermis was divided with a consistent approach for all MRI studies. We used an interactive tool for automatic identification of the midsagittal plane.18 This tool allows users to determine anatomic landmarks of the midsagittal plane by choosing points on any of the coronal, axial, or sagittal sections of the MRI data; a least-squares algorithm is used to fit the set of chosen points to the equation of the desired plane.
Cerebral parenchymal lesions were diagnosed by an experienced neuroradiologist (R.L.R.) in conventional, T1/T2-weighted, MRI scans. We categorized white matter lesions as cystic PVL or diffuse PVL (dPVL). Cystic PVL consisted of bilateral cystic lesions in the periventricular regions, whereas dPVL was defined as diffuse, excessive, high signal intensity in the periventricular white matter on T2-weighted scans, as described previously.19–22 PVHI consisted of a unilateral or asymmetric lesion of increased T2 signal in the periventricular white matter associated with ipsilateral germinal matrix-intraventricular hemorrhage. Infratentorial parenchymal lesions included unilateral or bilateral intraparenchymal CBH. The neuroradiologist (R.L.R.) was blinded to the clinical data for the cohort.
Continuous perinatal factors were summarized with means and SDs; categorical factors were summarized with proportions. One-way analysis of variance was used to compare birth weights, gestational ages, Apgar scores at 1 and 5 minutes, and GAEs at the time of term MRI studies between the 4 groups, and Fisher’s exact test was used to compare gender differences. Differences in cerebellar and cerebral brain volumes for preterm infants with normal MRI findings and for preterm infants with PVL, PVHI, or CBH were evaluated with 1-way analysis of variance, and the Bonferroni method was used to determine group differences. Paired t tests were used to compare ipsilateral and contralateral cerebellar and cerebral volumes in each subgroup (ie, preterm infants with PVL, PVHI, CBH, or normal MRI findings).
We reviewed 85 MRI studies for preterm infants. Of these, 11 were excluded because of direct parenchymal injury to both the cerebrum and cerebellum (5 cases of PVHI, 3 cases of dPVL, 2 cases of cystic PVL, and 1 case of posthemorrhagic hydrocephalus). No infants were excluded because of metabolic or genetic disorders. We analyzed the MRI data for the remaining 74 preterm infants. For the 74 preterm infants studied, term MRI studies were performed at a mean GAE of 40.1 ± 1.5 weeks. Cerebellar volumes and cerebral volumes were calculated for all MRI studies. Table 1 summarizes the baseline characteristics of our preterm cohort. There were no statistically significant differences in gestational age, birth weight, gender, Apgar scores, or GAE at the time of term MRI studies among the preterm infants studied.
Conventional MRI Findings
Of the 74 MRI studies performed for preterm infants at term GAE, 30 were normal, whereas 44 were abnormal. The abnormalities in the latter 44 studies consisted of 20 cases of dPVL, 14 cases of unilateral PVHI, and 10 cases of CBH. Table 2 summarizes the MRI abnormalities for our preterm infant cohort.
Cerebral and Cerebellar Hemispheric Volumes Among Preterm Infants With Normal and Abnormal MRI Studies
Cerebral brain volumes were significantly reduced for all preterm infants with abnormal term MRI studies, compared with preterm infants with normal term MRI scans (361.0 ± 22.2 mL vs 310.2 ± 19.3 mL, P < .001). The greatest reduction in cerebral brain volume was documented for preterm infants with PVHI, followed by infants with dPVL or CBH. When the relationship between cerebral and cerebellar hemispheric volumes was examined, preterm infants with unilateral PVHI showed significantly reduced volumes of the contralateral cerebellar hemisphere at term GAE (ipsilateral cerebellar mean: 10.3 ± 2.4 mL; contralateral cerebellar mean: 8.1 ± 2.0 mL; P < .001). Similarly, infants with unilateral CBH had significantly smaller contralateral cerebral hemispheric volumes (ipsilateral cerebral mean: 166.5 ± 21.9 mL; contralateral cerebral mean: 155.8 ± 19.7 mL; P < .01). However, there was no significant difference in the 2 cerebellar hemispheric volume measurements for infants with a diagnosis of bilateral dPVL or in cerebral hemispheric volume measurements for infants with bilateral CBH. Similarly, preterm infants with normal term GAE MRI studies did not demonstrate any significant differences when ipsilateral and contralateral cerebellar and cerebral hemispheric volumes were compared (Table 3).
Cerebellar Tissue Class Volumes Among Preterm Infants With Normal and Abnormal MRI Studies
Total cerebellar volume was decreased significantly in MRI studies performed at term GAE for all preterm infants with abnormal MRI scans, compared with preterm infants with normal MRI studies (17.9 ± 2.2 mL vs 24.2 ± 2.8 mL, P < .001) (Table 4). Specifically, cerebellar gray matter volumes in term MRI scans were decreased significantly for infants with CBH. Notably, preterm infants with PVHI and PVL had marked reductions in cerebellar cortical gray matter at term in the absence of direct cerebellar injury, compared with gray matter volume for premature infants with normal term MRI studies. Reductions in the volume of cerebellar myelinated white matter were also noted in all 3 diagnostic groups, compared with preterm infants with normal MRI scans (P < .05). Preterm infants with CBH showed the greatest reductions in total cerebellar volume and cerebellar gray and white matter volumes, followed by infants with PVHI or dPVL. There were no differences in cerebellar gray and myelinated white matter volumes between preterm infants with PVHI and those with dPVL. CSF volumes were significantly greater for preterm infants with abnormal MRI findings (P < .01) (Table 4). Interestingly, only preterm infants with unilateral PVHI showed significantly smaller cerebellar gray and white matter volumes in the contralateral cerebellar hemisphere, compared with the ipsilateral cerebellar hemisphere (cerebellar gray matter: 4.3 ± 1.4 mL vs 7.0 ± 1.4 mL; cerebellar white matter: 1.9 ± 1.2 mL vs 3.4 ± 1.3 mL; P < .01).
This study used 3-dimensional quantitative MRI to demonstrate for the first time in vivo evidence for a significant crossed trophic effect between the developing cerebral and cerebellar structures among preterm infants. Specifically, we showed that unilateral injury confined to the preterm cerebral hemisphere was associated with a significantly decreased volume of the contralateral cerebellar hemisphere and vice versa and that these effects were evident as early as term GAE.
To date, in vivo studies of the effects of prematurity and injury on brain development among preterm infants have focused principally on cerebral structures.23–26 For example, quantitative MRI studies have shown a reduction in total brain tissue volume, as well as cerebral cortical gray matter and myelinated white matter volumes, among preterm infants with periventricular white matter injury. As a secondary objective, we sought to determine whether the decrease in total cerebellar volume was attributable primarily to a decrease in cerebellar gray or myelinated white matter, or both. We demonstrated that both cerebellar gray and myelinated white matter volumes were decreased at term GAE among premature infants with injuries confined to the cerebral hemispheres. This apparent reduction in cerebellar gray matter volume suggests that injury to the supratentorial periventricular white matter impairs not only cerebral cortical development27 but also development of the remote developing cerebellar neurons. Moreover, the cerebellar myelinated white matter volume was also reduced among premature infants with dPVL or PVHI, even in the absence of direct CBH. At present, we cannot determine whether this decrease in myelinated white matter is attributable to a reduction in afferent fiber tracts to the cerebellum or is secondary to failed cerebellar cortical growth, with subsequent effects on efferent fiber tracts from the cerebellum.
Although the cerebellum has long been known to play a central role in the coordination of movement, more recent studies among older children and adults support its importance for cognition, language, attention, and social function.6,8,28 Survivors of prematurity manifest a range of neuromotor impairments and disabilities.29–31 By school age, an astonishing 25% to 50% of these children also demonstrate cognitive impairments, learning disabilities, and academic failure. The decreases in cerebellar gray and myelinated white matter volumes documented in this study suggest a potential role for cerebellar dysfunction in the high prevalence of motor and cognitive deficits reported for survivors of prematurity. Our data raise the intriguing question of whether the development of such nonmotor deficits among these children may be explained in part by a secondary remote effect on cerebellar development.
Preterm infants in our study with bilateral dPVL demonstrated symmetrical decreases in cerebellar volumes in the absence of direct cerebellar injury. Similarly, preterm infants with bilateral CBH showed bilateral decreases in cerebral hemispheric volumes. In contrast, preterm infants with unilateral PVHI, but not bilateral dPVL, demonstrated decreased volume of the contralateral cerebellar hemisphere, which suggests a mechanism mediated by disruption of crossing pathways between the cerebral and cerebellar hemispheres. The cerebellum receives afferent input from large parts of the cerebrum. One major source of excitatory activation in the cerebellum is from the contralateral frontoparietal cortex via the corticopontocerebellar tracts. Functional disconnection of these transneural pathways likely underlies the phenomenon of crossed cerebellar diaschisis, in which remote cerebral injury results in hypometabolism, cell loss, and atrophy of the contralateral cerebellar hemisphere.32 We reported previously that the third trimester is a period of particularly rapid growth for the cerebellum.33 In the current study, our data suggest that an important stimulus for this rapid growth is likely to be transsynaptic excitatory input from the contralateral cerebral hemisphere.34
Controversy exists regarding the relative contribution of early-life unilateral cortical lesions to cerebellar growth among young children.14,35–38 The presence of crossed cerebellar diaschisis has been demonstrated after postnatal injury. However, crossed cerebellar diaschisis typically is not present in the first 4 weeks after perinatal, unilateral, ischemic events.14 In our study of preterm infants, the period between birth and the term GAE scans averaged 10 weeks. Furthermore, in our previous work, we showed that the growth of the immature cerebellum is accelerated during late gestation.33 These 2 features may explain the apparent trophic effects noted in our study.
In contrast to crossed cerebellar diaschisis, the reverse situation (ie, remote effects in the cerebral cortex contralateral to a cerebellar lesion) has been described only rarely and among older subjects, such as adults after cerebellar stroke or resection of posterior fossa tumors, studied with single-photon emission computed tomography.15,37,39–41 The underlying mechanism is postulated to be interruption of the cerebellothalamocortical pathway, which projects from the deep cerebellar nuclei through the superior cerebellar peduncle to the contralateral cerebral cortex. The cerebellum, like the basal ganglia and thalamus, receives information from the cerebral cortex for processing and projects back to the sensorimotor and associative areas of the cerebral cortex, in a reverberative manner.42–44 Our findings indicated that unilateral CBH resulted in decreased contralateral cerebral brain volume, whereas bilateral CBH was associated with bilateral reductions in cerebral brain volumes. These data suggest that early-life primary cerebellar injury among extremely preterm infants can result in disruption of supratentorial neural systems. We postulate that crossed cerebellocerebral diaschisis may also play an important role in the long-term behavioral and cognitive impairments documented for formerly preterm infants.
This study has several potential limitations. Our cerebellar gray matter volumes might be slightly overestimated because we likely included a small amount of unmyelinated white matter tissue, given the inherent difficulty of differentiating the cerebellar white matter within the cerebellar gray matter at this age because of the lack of signal contrast in these tissues. However, currently there is no other available method to quantify cerebellar gray and white matter volumes. Moreover, a single investigator performed all manual segmentations of the cerebellum; therefore, all measurements were conducted with a consistent approach for all subgroups of preterm infants. In addition, intra-rater reliability was determined to be high for cerebellar gray and white matter volume measurements.
Unilateral cerebral parenchymal brain injury among premature infants was shown for the first time to result in subsequent impairment in the contralateral cerebellar volume; conversely, primary unilateral cerebellar injury was associated with impaired contralateral cerebral brain volume at term GAE. Cerebral parenchymal brain injury (bilateral or unilateral) was also associated with impaired cerebellar gray and white matter development, in the absence of direct cerebellar injury. These findings provide important insights into the highly integrated anatomic and functional interactions between the cerebrum and the cerebellum during development. It is reasonable to suggest that early-life cerebellar impairment, related to either direct cerebellar injury or cerebellar underdevelopment secondary to cerebral injury, plays a previously under-recognized role in the long-term cognitive, behavioral, and motor deficits associated with brain injury among premature infants. The full extent of the role of cerebellar injury in the genesis of intellectual, motor, and behavioral deficits among premature infants remains to be determined.
C.L. is supported by a postdoctoral fellowship from the Canadian Institutes of Health Research. We acknowledge support from the Hearst Fund, LifeBridge Fund, the United Cerebral Palsy Foundation, the Whitaker Foundation, National Institutes of Health grants P41 RR13218 and R21 MH67054, and the Children’s Hospital Boston Clinical Research Program.
We thank Shaye Moore for help with manuscript preparation.
- Accepted January 13, 2005.
- Address correspondence to Adré J. du Plessis, MBChB, MPH, Department of Neurology, Fegan 11, Children’s Hospital, 300 Longwood Ave, Boston, MA 02115. E-mail:
No conflict of interest declared.
- Merrill JD, Piecuch RE, Fell SC, Barkovich AJ, Goldstein RB. A new pattern of cerebellar hemorrhages in preterm infants. Pediatrics.1998;102(6) . Available at: www.pediatrics.org/cgi/content/full/102/6/e62
- ↵Johnsen SD, Tarby TJ, Lewis KS, Bird R, Prenger E. Cerebellar infarction: an unrecognized complication of very low birthweight. J Child Neurol.2002;17 :320– 324
- ↵Powls A, Botting N, Cooke RW, Marlow N. Motor impairment in children 12 to 13 years old with a birthweight of less than 1250 g. Arch Dis Child Fetal Neonatal Ed.1995;73 :F62–F66
- ↵Schmahmann JD, Sherman JC. The cerebellar cognitive affective syndrome. Brain.1998;121 :561– 579
- Berquin PC, Giedd JN, Jacobsen LK, et al. Cerebellum in attention-deficit hyperactivity disorder: a morphometric MRI study. Neurology.1998;50 :1087– 1093
- ↵Levisohn L, Cronin-Golomb A, Schmahmann JD. Neuropsychological consequences of cerebellar tumour resection in children: cerebellar cognitive affective syndrome in a paediatric population. Brain.2000;123 :1041– 1050
- Allin M, Matsumoto H, Santhouse AM, et al. Cognitive and motor function and the size of the cerebellum in adolescents born very pre-term. Brain.2001;124 :60– 66
- ↵Rollins NK, Wen TS, Dominguez R. Crossed cerebellar atrophy in children: a neurologic sequela of extreme prematurity. Pediatr Radiol.1995;25(suppl 1) :S20– S25
- ↵Shamoto H, Chugani HT. Glucose metabolism in the human cerebellum: an analysis of crossed cerebellar diaschisis in children with unilateral cerebral injury. J Child Neurol.1997;12 :407– 414
- ↵Haidar H, Warfield SK, Soul JS. A new accurate approach for “Talairach”-based parcellation for neonatal MRI. J Neuroimaging.2005; In press
- Volpe JJ. Cerebral white matter injury of the premature infant: more common than you think. Pediatrics.2003;112 :176– 180
- ↵Maalouf EF, Duggan PJ, Counsell SJ, et al. Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics.2001;107 :719– 727
- McKinstry RC, Miller JH, Snyder AZ, et al. A prospective, longitudinal diffusion tensor imaging study of brain injury in newborns. Neurology.2002;59 :824– 833
- Miller SP, Cozzio CC, Goldstein RB, et al. Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings. AJNR Am J Neuroradiol.2003;24 :1661– 1669
- ↵Murphy BP, Inder TE, Huppi PS, et al. Impaired cerebral cortical gray matter growth after treatment with dexamethasone for neonatal chronic lung disease. Pediatrics.2001;107 :217– 221
- Kilbride HW, Thorstad K, Daily DK. Preschool outcome of less than 801-gram preterm infants compared with full-term siblings. Pediatrics.2004;113 :742– 747
- ↵Limperopoulos C, Soul JS, Gauvreau K, et al. Late gestation cerebellar growth is rapid and impeded by premature birth. Pediatrics.2005;115 :688– 695
- Le Strange E, Saeed N, Cowan FM, Edwards AD, Rutherford MA. MR imaging quantification of cerebellar growth following hypoxic-ischemic injury to the neonatal brain. AJNR Am J Neuroradiol.2004;25 :463– 468
- ↵Sargent MA, Poskitt KJ, Roland EH, Hill A, Hendson G. Cerebellar vermian atrophy after neonatal hypoxic-ischemic encephalopathy. AJNR Am J Neuroradiol.2004;25 :1008– 1015
- ↵Rousseaux M, Steinling M. Crossed hemispheric diaschisis in unilateral cerebellar lesions. Stroke.1992;23 :511– 514
- Copyright © 2005 by the American Academy of Pediatrics