If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To clarify the morphologic characteristics of the brain, which are the foundation of the emergence of general movements (GMs) in very-low-birth-weight infants.
Study design
Prospective cohort study. GMs were scored according to a semiquantitative scoring system: the GMs optimality score (GMOS) at preterm and term ages. Brain magnetic resonance imaging (MRI) at term-equivalent age was scored using a validated scoring system (MRI score). We examined the relationship between the two scores by multiple regression analysis with relevant clinical background.
Subjects
We included 50 very-low-birth-weight infants cared for at Oita University Hospital from August 2012 to August 2018 who underwent MRI and GMs assessment. Their median gestational age and birth weight were 29w2d and 1145 g, respectively.
Results
The MRI score and systemic steroid administration were related to preterm GMOS, and the MRI score was related to term GMOS. The component cerebellum score and cortical grey matter score of the MRI score were associated with preterm GMOS, and the cerebellum and the cerebral white matter scores were associated with term GMOS.
Conclusion
The quality of GMs was associated with brain morphological development. The co-evaluation of GMs and brain morphology leads to accurate developmental prediction.
The advancement of neonatal intensive care has improved the survival rate of very-low-birth-weight infants (VLBWI). In addition, there is an increasing number of VLBWI who develop cerebral palsy and intellectual and cognitive impairment. These are important issues for neonatal medicine. Early prediction is required for early intervention according to the character of impairment [
Brain magnetic resonance imaging (MRI) is also useful for predicting adverse neurological outcome. Cerebral MRI conducted at term-equivalent age (TEA) is a sensitive tool that can provide important information about the preterm infant’s brain injury and structure, which is associated with long-term neurodevelopment, including motor and cognitive outcomes [
]. Early detection of intellectual and cognitive impairment is also important. Understanding the relationship between both assessment methods leads to more accurate early prognostic prediction.
This study aims to clarify the morphologic characteristics of the immature brain, which is the foundation of GMs emergence by assessing the relationship between brain MRI and GMs in the neonatal period. Aiming for clinical application, both assessment methods can be practiced widely and generally. This enables early developmental prediction, leading to early intervention according to individual predicted developmental characteristics from the neonatal period.
2. Method
2.1 Participants
VLBWI were recruited from the neonatal intensive care unit (NICU) of Oita University Hospital, Japan, between August 2012 and August 2018. They were participants of the clinical research that investigated the association between GMs and long-term neurodevelopment. Brain MRI was routinely performed on all VLBWI for the clinical indication in our NICU. The subjects of this study were cases in whom MRI could be performed at the age of 37–43 postmenstrual weeks. Infants were enrolled after parental written consent for GMs assessment was obtained during their NICU stay. The ethics committee of Oita University approved this study. The authors (TM, MT, SK) obtained the informed consent for clinical research participation and collected the clinical data from medical records.
2.2 GMs assessment
GMs were recorded repeatedly in a standard way during the NICU stay [
]. We divided the recordings into two periods (preterm: 30–36 postmenstrual weeks; term: 37–43 postmenstrual weeks). When there were multiple records within the period, the GMs recording with the best quality was used for analysis, because the neonate’s behaviour should be assessed by his or her best performance [
Normal GMs are characterized by complexity, variability, and fluency. In the neonatal period, the GMs are referred to as writhing movements. Typically, they are ellipsoid in form and are of small to moderate amplitude and speed. Abnormal GMs in the writhing period are divided into three main types, namely, poor repertoire (PR) GMs, cramped synchronised (CS) GMs, and chaotic GMs. PR GMs appear monotonous in their sequence of successive movement components. CS GMs appear rigid because all limb and trunk muscles contract and relax almost simultaneously. Chaotic GMs occur chaotically and lack any fluency or smoothness [
]. The optimality score is based on a detailed analysis of the quality of the GMs. Eight different aspects of the GMs are distinguished. The first item is the quality of the GMs, and it can be scored as normal, variable, and complex (4 points), PR GMs (2 points), chaotic GMs (1 point), or CS GMs (1 point). The other seven items are sequence, amplitude, speed, space, rotatory components, onset-offset, and tremulous movement. All can be scored as optimal (2 points) or non-optimal (1 point). The highest and therefore the most optimal score is 18, whereas the lowest score is 8. GMs assessments were conducted before MRI scoring by an author (TM) who had been repeatedly certified on a GM Trust advanced course.
2.3 Scoring of brain MRI
Infants were scanned at a median postmenstrual age of 39 weeks 4 days (range, 37w1d–43w5d). All infants were scanned while sedated using thiamylal sodium and were monitored by pulse oximetry. Images were obtained with a 1.5- or 3.0-Tesla MR scanner (1.5-T MRI system, EXCELART Vantage XGV PPP, TOSHIBA, Japan; 3T MRI system, MAGNETOM Verio, Siemens, Japan). Because MRI was examined by clinical indication, the scanned age and setting of sequence parameters were not constant. Axial T2-weighted MRI brain scans were acquired using the following sequence parameters median (range): echo time, 93 (80–115); repetition time, 4000 (2810–6000); and slice thickness, 4 mm (3–5).
Brain MRIs were assessed using a standardised scoring system that has been validated with very preterm infants [
]. This scoring system for conventional MRIs provides a comprehensive and objective characterisation of regional and global brain injury and brain growth. This score consists of the sum of the scores at cerebral white matter (WM), cortical grey matter (CGM), deep grey matter (DGM), and cerebellum. The brain measurement values were one of the components of the score. The biparietal diameter (BPD) was a component of WM. The interhemispheric distance (IHD) was a component of CGM. The deep GM area (DGMA) was a component of DGM. The transcerebellar diameter (TCD) was a component of the cerebellum. A higher MRI score indicates that the brain injuries were strong and the brain growth was suppressed [
BPD and IHD were measured on an axial plane perpendicular to the cross-section through the bilateral cochlea. DGMA was measured on a single axial section at the level at which the caudate heads and the lentiform nuclei and the thalami are maximally visible. DGMA was manually segmented. TCD was measured using the maximum cerebellar diameters parallel to the line connecting the bilateral cochlea. Since there were differences depending on the scanned age, the brain measurement values were corrected according to the postmenstrual week at MRI scan. We applied a linear regression described in the article for the correction. A linear association was found between postmenstrual age at scan and BPD (R = 0.41; slope, 1.4 mm/week), DGMA (R = 0.36; slope, 0.26 cm2/week), and TCD (R = 0.35; slope, 0.83 mm/week) [
]. The corrected values are expressed as cBPD, cDGMA, and cTCD, respectively. Where a significant association for age was not identified for IHD, ventricular diameter, and callosal thickness, we used the measured values for analysis.
One author (IH), who was unaware of the infant’s clinical history, measured and scored the measurable values. Authors TM and KS scored the unmeasurable factors with reference to the radiologist’s findings.
2.4 Statistics
The data were analysed using SPSS Statistics version 24 (IBM, Tokyo, Japan). We examined the difference between the normal GMs group and the abnormal group using Mann-Whitney U analysis for continuous variables (e.g., birth weight, height, MRI score) and Fisher’s exact test for nominal variables. We used the Spearman rank correlation coefficient (r) to study the correlations among MRI score, the measured values of MRI score, and GMOS. Differences with p < 0.05 were considered statistically significant.
After confirming the correlation between GMOS and MRI score, we performed multiple regression analysis using the score of each part of the brain, constituting the MRI score and the GMs associating variables by univariate analysis. In addition, we examined associated brain measurement values and the optimal or nonoptimal elements of GMOS in Mann-Whitney U analysis.
3. Results
3.1 Participants
Within the study period, 65 VLBWI were admitted to our NICU. There were 3 deaths during the neonatal period, 1 case of chromosomal abnormality, and 1 case of congenital muscular dystrophy, which were excluded from the GMs clinical research assessment. Six cases did not give consent to participate in the clinical research. In addition, 4 cases were excluded from this study, as MRI could not be performed in TEA due to chronic respiratory failure. Eventually, 50 subjects were enrolled in this study. Their median gestational age was 29w2d, ranging from 24w0d to 37w4d. Their median birth weight, height, and head circumference were 1145 g, 37.1 cm, and 26.7 cm, respectively. Details of the body measurements at birth are listed in Table 1. Fetal growth restriction, defined as birth weight below the 10th percentile of the Japanese standard [
We were able to assess GMs in 48 cases at preterm age and in 49 cases at term age. Abnormal GMs were found in 10 cases and normal GMs in 38 cases at preterm age. All abnormal patterns were PR GMs. At term age, abnormal GMs were found in 25 cases and normal GMs in 24 cases. In one case, CS GMs were observed, but another sequential record presented relatively good-quality PR GMs. Therefore, the PR GMs recording was used for examination. The median GMOS was 16 (range, 11–18) at preterm age and 14 (range, 9–18) at term age. The global qualitative assessments of GMs and the GMOS were significantly correlated in each period. The median GMOS in preterm age of PR GMs cases was 12 (range, 11–14) and of normal GMs cases was 16 (range, 13–18). The median GMOS in term age of PR GMs cases was 13 (range, 9–14) and of normal GMs cases was 16 (range, 13–18).
The trajectories of GMs through the preterm-term period showed PR-PR in 7 cases, PR-normal in 3 cases, normal-PR in 19 cases, and normal-normal in 18 cases.
3.3 MRI scoring
MRI was performed in all cases. The median MRI score was 5.0 (range, 0–18). The brain part subscores were 4.0 (0–12) for WM, 0.0 (0–5) for CGM, 0.0 (0–2) for DGM, and 0.0 (0–2) for cerebellum. The mean (standard deviation) of measured brain values were as follows: cBPD, 71.8 (3.8) mm; IHD, 3.6 (1.1) mm; cDGMA, 12.0 (0.8) cm2; and cTCD, 50.1 (2.1) mm (Table 2).
Table 2The comparison of the MRI scores and measurement value for each brain parts between GMs normal group in preterm and term age.
The MRI scores were significantly higher in the abnormal GMs group than in the normal GMs group at both preterm and term age (p < 0.01; Fig. 1a, b). The GMOS at preterm and term ages were correlated with the MRI score (p < 0.01; Fig. 1c, d).
Fig. 1The correlation between GMs and MRI scores. (a, b) The comparison of MRI scores between GMs abnormal and normal group in preterm age (a) and term age (b). (p < 0.01) The MRI scores were significantly higher in the abnormal GMs group than in the normal GMs group at both ages. (p < 0.01). (c, d) The correlation between GMOS in preterm age (c) and term age (d) and MRI score. The GMOS at both ages were correlated with the MRI score (p < 0.01).
The GMOS also correlated with the measured values of each brain part that are elements of the MRI score. The preterm GMOS correlated with IHD, cTCD (p < 0.01), and cBPD (p < 0.05). The term GMOS correlated with cBPD (p < 0.01), IHD, cDGMA, and cTCD (p < 0.05) (Fig. 2).
Fig. 2The correlation between GMOS and measured values of MRI scores. The scatterplot on the left represents preterm GMOS and on the right represents term GMOS. The preterm GMOS were correlated with cBPD (p < 0.05), IHD, and cTCD (p < 0.01). The term GMOS were correlated with cBPD (p < 0.01), IHD, cDGMA, and cTCD (p < 0.05).
We examined the differences in clinical background between the GMs normal group and the GMs abnormal group (Table 1). Abnormal GMs in the preterm age group were associated with significantly earlier birth gestational age and a longer period of intubated ventilation (p < 0.05) and many cases of systemic steroid administration and indomethacin administration (p < 0.05) than in the normal GMs group. There was no significant related factor other than MRI score between the term GMs normal and abnormal groups.
There were significant differences in WM, CGM, and cerebellum scores between the normal and abnormal GMs groups of preterm age and in WM and CGM scores between the normal GMs and abnormal groups of term age. Comparison of the mean measurement values of each brain part presented significant differences in IHD and cTCD between the preterm normal and abnormal GMs groups (p < 0.01) and in cBPD between the term normal and abnormal GMs groups (p < 0.05). The abnormal GMs group of preterm age presented longer IHD and shorter cTCD than the preterm normal GMs group. The term abnormal GMs group presented shorter cBPD than the term normal GMs group (Table 2).
3.6 Multiple regression analysis
We performed multiple linear regression analysis with the GMOS as a dependent variable using the stepwise variable increase/decrease method. The MRI score and the related background factors were introduced as independent variables. In the preterm GMOS, MRI score and systemic steroid administration were recognised as related factors. The normality of the residual was confirmed: R2 was 0.40, and the Durbin-Watson ratio was 1.697. When analysing by the score of each part of the brain instead of using the MRI score, the cerebellum score and the CGM score remained relevant. R2 was 0.366, and the Durbin-Watson ratio was 1.942.
With regard to the term GMOS, because no relevant factors other than the MRI score were observed, the score of each part of the brain was examined as an independent variable. The WM score and cerebellum score remained relevant. The normality of the residual was confirmed. R2 was 0.339, and the Durbin-Watson ratio was 1.759 (Table 3).
Table 3Multivariate linear regression models. The related variables to GMOS.
Std. Error; standard error, CI; confidence intervals.
3.7 Relation between brain metrics and characteristics of GMs
Since multiple regression analysis showed the relation between the GMOS and the MRI score of each part of the brain, we compared the measured values of each brain part to items on the GMs optimality list to examine the characteristics of GMs by brain morphologic change. The relevance was recognised between preterm GMOS and CGM and cerebellum score and between term GMOS and WM and cerebellum score. We compared the difference in IHD and cTCD values between the preterm optimal GMs group and the nonoptimal group and the difference in cBPD and cTCD values between the term GMs optimal group and nonoptimal group. Because the “quality of the GMs” item on the GMOS is the same as the global qualitative assessment of GMs, we compared the other items of GMOS. In the preterm GMs groups, the space score of the nonoptimal group was associated with a significantly longer IHD (p < 0.05), and sequence, space (p < 0.01), and tremorous movements (p < 0.05) of the nonoptimal group were associated with a significantly longer cTCD. With regard to the GMs of the term group, there were no significant differences in any item, except for quality score (Table 4).
Table 4The comparison of measured value and the items of GMs optimality list.
Fig. 3 presents a boxplot diagram of the comparison of the cTCD values of each item on the preterm GMOS optimality list. The GMs of infants with small cTCD were characterised by poor sequence, lack of use of full space, and prominent tremulous movements.
Fig. 3The comparison of the cTCD values of each item on the preterm GMOS non-optimal and optimal groups. seq: sequence, amp; amplitude, rotat; rotatory components, on-offset; onset-offset of movements, tremor; tremulous movements *p < 0.01, **p < 0.05.
This study demonstrated the association between the quality of GMs and the MRI assessment scale, which reflects both injury and brain growth. The clinical background factors that may influence the quality of GMs can also be examined by multiple regression analysis. For preterm GMs, MRI score and systemic steroid administration were associated factors. For term GMs, the MRI score was the only associated factor. The association between the GMOS and the component score of each brain part were observed in the CGM and cerebellum for preterm GMs and in the WM and cerebellum for term GMs. With the exception of systemic steroid administration, clinical background was not recognised as a factor of GMs quality. There are reports that the quality of GMs in the preterm period deteriorates because of acute illness and systemic steroid administration [
]. Abnormal GMs are common in the first 14 days of life in extremely-low-birth-weight infants because of physiologic instability. Septicaemia and artificial ventilation are associated with deterioration of GMs [
]. In the present study, GMs recording was conducted when the patients were in a stable condition after acute illness. In addition, cases in which MRI could not be performed at TEA because of chronic respiratory failure were excluded from this study. Therefore, this study was conducted in subjects with a good general condition, which allows for less influence of clinical background factors. Our results suggested that hydrocortisone (HDC) was associated with a decrease in the GMOS in the preterm period. Hitzert reported that HDC treatment did not present a statistically significant difference of GMOS between the HDC treated group and the control group [
]. However, their subjects were few in number; there was a possibility of type 2 error. In their report the HDC group showed a tendency to lower GMOS on day 1 of HDC therapy. In addition, in all three cases of normal GMs before therapy, the GMs quality decreased on day 7 of therapy; two cases changed to PR, and one case changed to chaotic. We use HDC to prevent late-onset circulatory collapse after the acute illness phase. The GMs recording and steroid administration period overlapped. Therefore, the quality of the GMs might have been decreasing in association with systemic steroid administration.
Regarding the trajectory of GMs, preterm infants often show abnormal GMs before term, which may normalise at TEA [
]. Contrary to this, there were many subjects in the present study whose GMs in the preterm period were normal, but changed to PR at TEA. De Vries reported that abnormal GMs at TEA are common in extremely low-birth-weight infants. Their GMs often appear stiff and cramped with extended legs [
]. Our subjects often presented the same character of GMs at TEA, which might have caused the decreasing GMs quality. Another possibility was an influence of neonatal care. The infants were swaddled during the preterm period. We loosened the swaddling surrounding the baby in the incubator and recorded the GMs. Ferrari reported that swaddling promotes a flexed posture of the limbs with adduction of shoulders, facilitates elegant wrist movements and movements toward and across the midline, and reduces abrupt movements [
]. After infants are out of the incubator at TEA, we record the GMs of the infants wearing only a nappy without swaddling. Many VLBWI tend to present posture and state instability if they are not swaddled or wearing clothing. The infants in nappies only began to cry soon after awakening, accompanied by jerky, fast, and large GMs or reflexes such as startle.
There are several reports regarding the association between term GMs and brain MRI at TEA [
]. In the present study, we found a significant correlation between the GMOS for both age and the MRI score in TEA using semiquantitative detailed scoring. In consideration of the quality of GMs and related parts of the brain, it was reported that abnormal GMs at 1 and 3 months’ corrected age were associated with WM abnormalities on MRI at TEA [
Relationship between white matter pathology and performance on the general movement assessment and the test of infant motor performance in very preterm infants.
]. In the present study, we also found a significant association between the GMOS at term age and WM score. However, the association was not found between GMOS at preterm age and WM score by multiple regression analysis. This might be due to low incidence of cystic periventricular leukomalacia, which is major disorder leading to WM lesions and abnormal GMs in preterm period [
]. There was only one case of cystic periventricular leukomalacia in this study subject.
Regarding the CGM score, Olsen reported that abnormal GMs at both preterm and term age were associated with larger IHD and higher CGM abnormality on MRI at TEA [
]. These findings were consistent with our results by univariate comparisons. However, no association was found between term GMs and CGM score by multiple regression analysis. Olsen analysed the quality of GMs with binary value: categorised as normal and abnormal by generalised estimating equations; whereas, we applied GMOS as a continuous variable that made the multiple regression analysis possible with a small number of subjects. Since the degree of GMs normality or abnormality is variable, it would be more beneficial to assess the GMs quality as continuous variables in multivariate analysis.
From the results of the present study, IHD, which was the main component of the CGM score, correlates with cerebrospinal fluid volume, and a larger IHD is thought to reflect an impairment in brain growth [
]. The GMs in the term period might reflect WM injury rather than brain growth impairment.
The first year after conception is characterised by high developmental activity in the brain, especially in the cortical subplate and cerebellum. As the complexity and variation of GMs emergence and the synaptic activity in the cortical subplate emergence coincide, Hadders-Alga speculated that the cortical subplate is the neural substrate of GMs complexity and variation [
]. The subplate is a transient structure that gradually dissolves between 3 months before term and 3 months’ corrected age, when the cortical plate takes over. The subplate itself is not evaluated in MRI score; it would be related to CGM and WM scores.
Regarding the cerebellum score, in the present study, the GMs scores for both the preterm and term periods were associated with the cerebellum score. The relationship between infantile GMs and cerebellar size has been reported. Decreased cerebellar diameter at TEA was related to abnormal GMs at 3 months postterm age [
]. There has been no report showing the association between cerebellar size and the quality of GMs in the preterm period. It seems that this relevance was also found in preterm GMs and cerebellar size by applying the semiquantitative scoring of GMs quality. The normal value of the TCD at TEA in the preterm infant cohort was reported as mean 50.8 mm, standard deviation 3.3 mm [
]. In the present study, the mean cTCD of the preterm normal GMs group was 50.5 mm and that of the preterm abnormal GMs group was 48.8 mm. This showed that those with sufficient cerebellar growth had good-quality GMs.
The cerebellar volumetric growth impairment is prevalent in preterm infants even in the absence of destructive cerebellar injury [
]. The mechanism for impaired cerebellar development has been presumed to be the presence of severe supratentorial injury, that is, through the cerebro-cerebellar diaschisis [
]. However, in the present study, the cTCD was significantly small in the abnormal GMs group independently of the CGM or WM scores. It is unclear whether sufficient cerebellar growth created good-quality GMs or whether good-quality GMs promoted cerebellar growth. We speculated that motor and sensory feedback stimulus that was accompanied by GMs under in the extrauterine environment might be related to cerebellar growth. The cerebellar neurons receive excitatory inputs from the precerebellar nuclei outside the cerebellum. The Purkinje cells receive these inputs and send their axons to neurons in the deep cerebellar nuclei. The deep cerebellar nuclei neurons send their axons to other regions of the brain [
]. The anatomy and development of the cerebellum is conserved between mammals and bony fish. In zebrafish, it is reported that afferent sensory stimulation develops rapidly with the emergence of Purkinje cell excitability [
With respect to the relation between cerebellar size and development, Shah reported that cerebellar volume reduction was correlated with cognitive and motor development at 2 years of age [
]. In view of the role of the cerebellum in motor learning and cognitive functioning, impaired cerebellar development in preterm infants may contribute to the motor impairments and cognitive deficits commonly experienced by preterm infants [
]. In the present study, when comparing the score for each GMOS item and the cTCD, the cTCD was significantly larger in cases with optimal sequence, space, and tremulous movement. These elements of GMs might be related to the cerebellar development.
The MRI assessment scale, which reflects both injury and brain growth, can be applied easily to MRI scans used for clinical purposes. The measurable components of the score of IHD, BPW, and TCD can be repeatedly obtained by brain ultrasonography in the NICU. By evaluating those measurements together with the assessments of GMs over time, more detailed relationships between GMs and brain morphologic development will be found.
In the present study, the association between the quality of GMs and brain morphologic development was confirmed. Both GM quality and brain morphologic development have been reported to be associated with motor and cognitive prognosis [
]. It is important to evaluate them in combination to make an early accurate prediction of cognitive and intellectual impairments, which are currently difficult to predict in the neonatal period. This will lead to early intervention with family support, depending on the characteristics of the individual predicted impairment from the neonatal period.
We recognise several limitations of this study. First, the number of subjects was relatively small. There are possibilities of type 2 errors in the multiple regression analysis. GMs reflect the integrity and development of the central nervous system. Therefore, the development of any part of the brain may be somewhat related to GMs. We believe we have been able to clarify the predominant part of the brain that is strongly related to GMs quality. Second, the MRI scanned age and the setting of the scanning sequence parameters were not constant. On the other hand, this means that the scoring system can be applied to MRI scans that are done for clinical purposes. Therefore, general versatility of the MRI score is a high and promising indicator for the assessment of morphological character of the developing brain. Third, this study was an examination from a single facility. The GMs assessments and the MRI score are methods that can be generally performed in many facilities, and it is desirable that the results of the present study will be re-verified in multicentre study. Fourth, we have not presented the relationship in terms of prognosis of the subjects. Further prospective studies will be needed to confirm the association with long-term prognosis.
In conclusion, the developmental brain morphological characteristics that are the foundation of the emergence of GMs were clarified. Coevaluation of the GMs assessments and brain morphological assessments from the neonatal period leads to an accurate understanding of the developing brain and would be useful for the early prediction of various developmental impairments.
Acknowledgement
This work was supported by JSPS KAKENHI Grant Number 15K11705.
References
Jarjour I.T.
Neurodevelopmental outcome after extreme prematurity: a review of the literature.
Relationship between white matter pathology and performance on the general movement assessment and the test of infant motor performance in very preterm infants.
30536-9&id=gr1.jpg){kind=link}
30536-9&id=gr2.jpg){kind=link}
30536-9&id=gr3.jpg){kind=link}