Frequency Properties Of Spontaneous BOLD Signal And Resting-State Functional Networks In Moderate And Late Preterm Newborns

BackgroundA moderate and late preterm(MLPT)newborn is defined as an infant born after 32 but < 37 weeks of completed gestation [1].An estimated 14.9 million babies were born preterm around the world in 2010,and MLPT newborns account for more than 80 % of preterm newborns [1].Although MLPT newborns are born only a few weeks prior to term,they have been documented to have lower brain weights,less cortical folding and gyrification,and less myelinated white matter compared with term newborns [2-4].Even at termequivalent ages,MLPT infants are linked with altered brain white matter microstructures,brain sizes,and cortical gray matter volumes [5-7].These findings demonstrate that the final gestational weeks approaching term are critical for brain growth and development.Currently,numerous clinical studies elucidated that MLPT infants experience increased risk of neurodevelopmental delays,cognitive and behavioral problems compared with term newborns [8-11].Especially,these disorders are associated with abnormal resting-state networks [12].Recent advances in resting-state functional magnetic resonance imaging(rs-fMRI)based on blood oxygenation level–dependent(BOLD)may provide quantitative markers of normal neurodevelopment and brain injury[12-24].In this project we systemically explored the frequency of spontaneous BOLD signal,the effective connectivity of the common resting state networks,and small-world properties of the complex resting state networks in MLPT newborns.Materials and methods1.SubjectsWe recruited 129 newborns that included 71 MLPT newborns and 58 control term newborns(born from 2013 to 2014)from the Department of Pediatrics,Daping Hospital,Third Military Medical University(Chongqing,China).Based on the selection and exclusion criteria,26 MLPT and 35 term newborns were included in the final analyses.2.Data acquisition and rs-fMRI data preprocessingAll magnetic resonance imaging(MRI)scans were performed using a Magnetom Verio 3.0T scanner(Siemens,German).The rs-fMRI data were collected using a T2*-weighted echo-planar imaging(EPI)sequence(TR/TE: 2000 ms/30 ms,FOV = 220×220 mm2,matrix size = 64×64,33 slices,thickness: 3.0 mm,resolution: 3.4×3.4×3.0 mm3).Resting-state data were collected for 8 minutes,resulting in 240 volumes per subject.Rs-fMRI data preprocessing was carried out using both AFNI(http://afni.nimh.nih.gov/afni/)and SPM8(http://www.fil.ion.ucl.ac.uk/spm/software/spm8/)software packages.3.The study methods for the frequency of spontaneous BOLD signal in MLPT newbornsWe used rs-fMRI and the amplitude of low-frequency fluctuation(ALFF)method to investigate the frequency properties of spontaneous BOLD signals in 26 MLPT and 35 term newborns.Two frequency bands,slow-4(0.027–0.073 Hz)and slow-5(0.01–0.027 Hz),were analyzed.To determine the effects of group,frequency band and their interaction on ALFF,a two-way repeated-measures analysis of variance(ANOVA)was conducted using SPM8(http://www.fil.ion.ucl.ac.uk/spm/software/spm8/).The ALFF values of each subject were modeled using a flexible factorial design with group(MLPT and term newborns)as a between-subject factor and frequency band(slow-4 and slow-5)as a repeated measure.Post-hoc two-sample t tests were performed with those clusters showing significant main effects for group and frequency bands and for their interaction.The main and interaction effects were corrected by family-wise error(FWE,P < 0.05)and the topological false discovery rate(topo-FDR,P < 0.01),respectively.4.The study methods for the effective connectivity of resting state networks in MLPT NewbornsIn order to reveal the common resting state networks and to quantify the possible effective connectivity within them in MLPT newborns,we applied a multivariate Granger causality analysis(mGCA)[25] to the resting state networks retrieved by independent component analysis(ICA)from rs-fMRI data of 26 MLPT and 35 term newborns.The ICA procession were performed with GIFT(http://icatb.sourceforge.net/,version 1.3h)software packages [26].The measurement data were analyzed by two-sample t tests,and the enumeration data were tested by Chi-squared tests.P < 0.05 was considered statistically significant.All quantitative analyses were performed using IBM SPSS Statistics for Windows,version 21.0(IBM Corp.,Armonk,NY,USA).5.The study methods for the small-world properties of resting state networks in MLPT NewbornsIn this study,we described the important properties of functional connectivity between 90 cortical and sub-cortical regions in 26 MLPT newborns and 35 term newborns,and construct a set of undirected graphs.All network metrics were calculated using GRETNA(www.nitrc.org/frs/down load.php/5534/gretna.zip)software package.Statistical analysis of the data were performed by one-way ANOVA.Post-hoc t-tests were used for the two group(MLPT and Term newborns)comparison.P<0.01,Bonferroni corrected was accepted as statistically significant.Connectivity graphs were shown by Brainnet Viewer(www.nitrc.org/projects/bnv/)[27] software package.Results1.The results for the frequency of spontaneous BOLD signal in MLPT newborns showed widespread differences in ALFF between the two bands(slow-4 and slow-5);differences occurred mainly in the primary sensory and motor cortices and to a lesser extent in association cortices and subcortical areas.Compared with term newborns,MLPT newborns showed significantly altered neural activity predominantly in the primary sensory and motor cortices and in the posterior cingulate gyrus/precuneus.In addition,a significant interaction between frequency bands and groups was observed in the primary somatosensory cortex.Intriguingly,these primary sensory and motor regions have been proven to be the major cortical hubs during the neonatal period.2.The results for the effective connectivity of resting state networks in MLPT Newborns unveiled the detailed topology of seven resting state networks(visual network,VN;default mode network,DMN;sensorimotor network,SMN;auditory network,AN;salience network,SN;left frontoparietal network,FPN_L;right frontoparietal network,FPN_R)in MLPT and term newborns.Intrigingly,we found that MLPT newborns exhibited different effective connectivity patterns and decreased causal interactions among the RSNs relative to term newborns.There is one network circuit(DMN-SN-SMN)within RSNs in MLPT newborns,and two network circuits(DMN-SN-SMN and DMN-AN-SMN)in term newborns.In the further degree analysis of the two groups,there is no significant difference between two groups.In particular,only SMN was identified a hub in the brain functional architecture of MLPT newborns,and both SMN and DMN were hubs in term newborns.3.The results for the small-world properties of resting state networks in MLPT Newborns showed the important properties of functional connectivity between 90 cortical and sub-cortical regions in MLPT and term newborns,and constructed a set of undirected graphs.Both MLPT and term brains exhibited an efficient and economical small world topology: densely connected nearby regions,module formed,sparse,but well integrated,distant connections,and small world index greater than 1?Furthermore,brain networks showed the hubs identified using degree and betweenness centrality measures are more richly connected to others.Intriguingly,MLPT newborns showed lower small world index and lower global/local efficiency,and several different distributions of modules and hubs compared with term newborns.ConclusionsThis project revealed the frequency of spontaneous BOLD signal differences between MLPT and term newborns,which contribute to the understanding of regional development of spontaneous brain rhythms of MLPT newborns;unveiled the topology of common resting state networks and different effective connectivity patterns in MLPT and term newborns;uncovered an efficient and economical small world topology in MLPT and term newborns,which contribute to the understanding of underlying coordinated activity in these neonatal brains.This preliminary investigation may provide a new insight to the determination of the neurophysiological mechanisms of MLPT newborns.