Early prediction of cognitive deficits in very preterm infants using functional connectome data in an artificial neural network framework

Supervised contrastive learning enhances graph convolutional networks for predicting neurodevelopmental deficits in very preterm infants using brain structural connectome

Very preterm (VPT) infants (born at less than 32 weeks gestational age) are at high risk for various adverse neurodevelopmental deficits. Unfortunately, most of these deficits cannot be accurately diagnosed until the age of 2-5 years old. Given the benefits of early interventions, accurate diagnosis and prediction soon after birth are urgently needed for VPT infants. Previous studies have applied deep learning models to learn the brain structural connectome (SC) to predict neurodevelopmental deficits in the preterm population. However, none of these models are specifically designed for graph-structured data, and thus may potentially miss certain topological information conveyed in the brain SC. In this study, we aim to develop deep learning models to learn the SC acquired at term-equivalent age for early prediction of neurodevelopmental deficits at 2 years corrected age in VPT infants. We directly treated the brain SC as a graph, and applied graph convolutional network (GCN) models to capture complex topological information of the SC. In addition, we applied the supervised contrastive learning (SCL) technique to mitigate the effects of the data scarcity problem, and enable robust training of GCN models. We hypothesize that SCL will enhance GCN models for early prediction of neurodevelopmental deficits in VPT infants using the SC. We used a regional prospective cohort of ∼280 VPT infants who underwent MRI examinations at term-equivalent age from the Cincinnati Infant Neurodevelopment Early Prediction Study (CINEPS). These VPT infants completed neurodevelopmental assessment at 2 years corrected age to evaluate cognition, language, and motor skills. Using the SCL technique, the GCN model achieved mean areas under the receiver operating characteristic curve (AUCs) in the range of 0.72∼0.75 for predicting three neurodevelopmental deficits, outperforming several competing models. Our results support our hypothesis that the SCL technique is able to enhance the GCN model in our prediction tasks.

Resting-State Functional Magnetic Resonance Imaging Network Association With Mortality, Epilepsy, Cognition, and Motor Two-Year Outcomes in Suspected Severe Neonatal Acute Brain Injury

BACKGROUND AND OBJECTIVES

In acute brain injury of neonates, resting-state functional magnetic resonance imaging (MRI) (RS) showed incremental association with consciousness, mortality, cognitive and motor development, and epilepsy, with correction for multiple comparisons, at six months postgestation in neonates with suspected acute brain injury (ABI). However, there are relatively few developmental milestones at six months to benchmark against, thus, we extended this cohort study to evaluate two-year outcomes.

METHODS

In 40 consecutive neonates with ABI and RS, ordinal scores of resting-state networks; MRI, magnetic resonance spectroscopy, and electroencephalography; and up to 42-month outcomes of mortality, general and motor development, Pediatric Cerebral Performance Category Scale (PCPC), and epilepsy informed associations between tests and outcomes.

RESULTS

Mean gestational age was 37.8 weeks, 68% were male, and 60% had hypoxic-ischemic encephalopathy. Three died in-hospital, four at six to 42 months, and five were lost to follow-up. Associations included basal ganglia network with PCPC (P = 0.0003), all-mortality (P = 0.005), and motor (P = 0.0004); language/frontoparietal network with developmental delay (P = 0.009), PCPC (P = 0.006), and all-mortality (P = 0.01); default mode network with developmental delay (P = 0.003), PCPC (P = 0.004), neonatal intensive care unit mortality (P = 0.01), and motor (P = 0.009); RS seizure onset zone with epilepsy (P = 0.01); and anatomic MRI with epilepsy (P = 0.01).

CONCLUSION

For the first time, at any age, resting state functional MRI in ABI is associated with long-term epilepsy and RSNs predicted mortality in neonates. Severity of RSN abnormality was associated with incrementally worsened neurodevelopment including cognition, language, and motor function over two years.

High-accuracy machine learning techniques for functional connectome fingerprinting and cognitive state decoding

The human brain is a complex network comprised of functionally and anatomically interconnected brain regions. A growing number of studies have suggested that empirical estimates of brain networks may be useful for discovery of biomarkers of disease and cognitive state. A prerequisite for realizing this aim, however, is that brain networks also serve as reliable markers of an individual. Here, using Human Connectome Project data, we build upon recent studies examining brain-based fingerprints of individual subjects and cognitive states based on cognitively demanding tasks that assess, for example, working memory, theory of mind, and motor function. Our approach achieves accuracy of up to 99% for both identification of the subject of an fMRI scan, and for classification of the cognitive state of a previously unseen subject in a scan. More broadly, we explore the accuracy and reliability of five different machine learning techniques on subject fingerprinting and cognitive state decoding objectives, using functional connectivity data from fMRI scans of a high number of subjects (865) across a number of cognitive states (8). These results represent an advance on existing techniques for functional connectivity-based brain fingerprinting and state decoding. Additionally, 16 different functional connectome (FC) matrix construction pipelines are compared in order to characterize the effects of different aspects of the production of FCs on the accuracy of subject and task classification, and to identify possible confounds.

Deep Hashing Mutual Learning for Brain Network Classification

Recently, clinical phenotypic semantic information has begun to play an important role in some brain network classification methods based on deep learning. However, most current methods only consider the phenotypic semantic information of individual brain networks but ignore the potential phenotypic characteristics among group brain networks. To address this problem, we present a deep hashing mutual learning (DHML)-based brain network classification method. Specifically, we first design a separable CNN-based deep hashing learning to extract individual topological features of brain networks and map them into hash codes. Secondly, we construct a group brain network relationship graph based on the similarity of phenotypic semantic information, in which each node is a brain network, and the properties of the nodes are the individual features extracted in the previous step. Then, we adopt a GCN-based deep hashing learning to extract the group topological features of the brain network and map them to hash codes. Finally, the two deep hashing learning models perform mutual learning by measuring the distribution differences between the hash codes to achieve the interaction of individual and group features. The experimental results on the three commonly used brain atlases (AAL Atlas, Dosenbach160 Atlas, and CC200 Atlas) of the ABIDE I dataset show that our proposed DHML method achieves optimal classification performance compared with some state-of-the-art methods.

Machine Learning and Prediction in Fetal, Infant, and Toddler Neuroimaging: A Review and Primer

Predictive models in neuroimaging are increasingly designed with the intent to improve risk stratification and support interventional efforts in psychiatry. Many of these models have been developed in samples of children school-aged or older. Nevertheless, despite growing evidence that altered brain maturation during the fetal, infant, and toddler (FIT) period modulates risk for poor mental health outcomes in childhood, these models are rarely implemented in FIT samples. Applications of predictive modeling in children of these ages provide an opportunity to develop powerful tools for improved characterization of the neural mechanisms underlying development. To facilitate the broader use of predictive models in FIT neuroimaging, we present a brief primer and systematic review on the methods used in current predictive modeling FIT studies. Reflecting on current practices in more than 100 studies conducted over the past decade, we provide an overview of topics, modalities, and methods commonly used in the field and under-researched areas. We then outline ethical and future considerations for neuroimaging researchers interested in predicting health outcomes in early life, including researchers who may be relatively new to either advanced machine learning methods or using FIT data. Altogether, the last decade of FIT research in machine learning has provided a foundation for accelerating the prediction of early-life trajectories across the full spectrum of illness and health.

Graph Neural Networks in Network Neuroscience

Noninvasive medical neuroimaging has yielded many discoveries about the brain connectivity. Several substantial techniques mapping morphological, structural and functional brain connectivities were developed to create a comprehensive road map of neuronal activities in the human brain -namely brain graph. Relying on its non-euclidean data type, graph neural network (GNN) provides a clever way of learning the deep graph structure and it is rapidly becoming the state-of-the-art leading to enhanced performance in various network neuroscience tasks. Here we review current GNN-based methods, highlighting the ways that they have been used in several applications related to brain graphs such as missing brain graph synthesis and disease classification. We conclude by charting a path toward a better application of GNN models in network neuroscience field for neurological disorder diagnosis and population graph integration. The list of papers cited in our work is available at https://github.com/basiralab/GNNs-in-Network-Neuroscience.

A Semi-Supervised Graph Convolutional Network for Early Prediction of Motor Abnormalities in Very Preterm Infants

Approximately 32-42% of very preterm infants develop minor motor abnormalities. Earlier diagnosis soon after birth is urgently needed because the first two years of life represent a critical window of opportunity for early neuroplasticity in infants. In this study, we developed a semi-supervised graph convolutional network (GCN) model that is able to simultaneously learn the neuroimaging features of subjects and consider the pairwise similarity between them. The semi-supervised GCN model also allows us to combine labeled data with additional unlabeled data to facilitate model training. We conducted our experiments on a multisite regional cohort of 224 preterm infants (119 labeled subjects and 105 unlabeled subjects) who were born at 32 weeks or earlier from the Cincinnati Infant Neurodevelopment Early Prediction Study. A weighted loss function was applied to mitigate the impact of an imbalanced positive:negative (~1:2) subject ratio in our cohort. With only labeled data, our GCN model achieved an accuracy of 66.4% and an AUC of 0.67 in the early prediction of motor abnormalities, outperforming prior supervised learning models. By taking advantage of additional unlabeled data, the GCN model had significantly better accuracy (68.0%, p = 0.016) and a higher AUC (0.69, p = 0.029). This pilot work suggests that the semi-supervised GCN model can be utilized to aid early prediction of neurodevelopmental deficits in preterm infants.

Genetic and environmental factors influencing neonatal resting-state functional connectivity

Functional magnetic resonance imaging has been used to identify complex brain networks by examining the correlation of blood-oxygen-level-dependent signals between brain regions during the resting state. Many of the brain networks identified in adults are detectable at birth, but genetic and environmental influences governing connectivity within and between these networks in early infancy have yet to be explored. We investigated genetic influences on neonatal resting-state connectivity phenotypes by generating intraclass correlations and performing mixed effects modeling to estimate narrow-sense heritability on measures of within network and between-network connectivity in a large cohort of neonate twins. We also used backwards elimination regression and mixed linear modeling to identify specific demographic and medical history variables influencing within and between network connectivity in a large cohort of typically developing twins and singletons. Of the 36 connectivity phenotypes examined, only 6 showed narrow-sense heritability estimates greater than 0.10, with none being statistically significant. Demographic and obstetric history variables contributed to between- and within-network connectivity. Our results suggest that in early infancy, genetic factors minimally influence brain connectivity. However, specific demographic and medical history variables, such as gestational age at birth and maternal psychiatric history, may influence resting-state connectivity measures.

Correlation of abnormal brain changes with perinatal factors in very preterm infants based on diffusion tensor imaging

Background

It remains unclear whether very preterm (VP) infants have the same level of brain structure and function as full-term (FT) infants. In addition, the relationship between potential differences in brain white matter microstructure and network connectivity and specific perinatal factors has not been well characterized.

Objective

This study aimed to investigate the existence of potential differences in brain white matter microstructure and network connectivity between VP and FT infants at term-equivalent age (TEA) and examine the potential association of these differences with perinatal factors.

Methods

A total of 83 infants were prospectively selected for this study: 43 VP infants (gestational age, or GA: 27–32 weeks) and 40 FT infants (GA: 37–44 weeks). All infants at TEA underwent both conventional magnetic resonance imaging (MRI) and diffusion tensor imaging (DTI). Significant differences in white matter fractional anisotropy (FA) and mean diffusivity (MD) images between the VP and FT groups were observed using tract-based spatial statistics (TBSS). The fibers were tracked between each pair of regions in the individual space, using the automated anatomical labeling (AAL) atlas. Then, a structural brain network was constructed, where the connection between each pair of nodes was defined by the number of fibers. Network-based statistics (NBS) were used to examine differences in brain network connectivity between the VP and FT groups. Additionally, multivariate linear regression was conducted to investigate potential correlations between fiber bundle numbers and network metrics (global efficiency, local efficiency, and small-worldness) and perinatal factors.

Results

Significant differences in FA were observed between the VP and FT groups in several regions. These differences were found to be significantly associated with perinatal factors such as bronchopulmonary dysplasia (BPD), activity, pulse, grimace, appearance, respiratory (APGAR) score, gestational hypertension, and infection. Significant differences in network connectivity were observed between the VP and FT groups. Linear regression results showed significant correlations between maternal years of education, weight, the APGAR score, GA at birth, and network metrics in the VP group.

Conclusions

The findings of this study shed light on the influence of perinatal factors on brain development in VP infants. These results may serve as a basis for clinical intervention and treatment to improve the outcome of preterm infants.

Development and validation of a prediction model for intrapartum cesarean delivery based on the artificial neural networks approach: a protocol for a prospective nested case-control study

INTRODUCTION

Although intrapartum caesarean delivery can resolve dystocia, it would still lead to several adverse outcomes for mothers and children. The obstetric care professionals need effective tools that can help them to identify the possibility and risk factors of intrapartum caesarean delivery, and further implement interventions to avoid unnecessary caesarean birth. This study aims to develop a prediction model for intrapartum caesarean delivery with real-life data based on the artificial neural networks approach.

METHODS AND ANALYSIS

This study is a prospective nested case-control design. Pregnant women who plan to deliver vaginally will be recruited in a tertiary hospital in Southwest China from March 2022 to March 2024. The clinical data of prelabour, intrapartum period and psychosocial information will be collected. The case group will be the women who finally have a baby with intrapartum caesarean deliveries, and the control group will be those who deliver a baby vaginally. An artificial neural networks approach with the backpropagation algorithm multilayer perceptron topology will be performed to construct the prediction model.

ETHICS AND DISSEMINATION

Ethical approval for data collection was granted by the Ethics Committee of West China Second University Hospital, Sichuan University, and the ethical number is 2021 (204). Written informed consent will be obtained from all participants and they can withdraw from the study at any time. The results of this study will be published in peer-review journal.

Differentiate preterm and term infant brains and characterize the corresponding biomarkers via DICCCOL-based multi-modality graph neural networks

Preterm birth is a worldwide problem that affects infants throughout their lives significantly. Therefore, differentiating brain disorders, and further identifying and characterizing the corresponding biomarkers are key issues to investigate the effects of preterm birth, which facilitates the interventions for neuroprotection and improves outcomes of prematurity. Until now, many efforts have been made to study the effects of preterm birth; however, most of the studies merely focus on either functional or structural perspective. In addition, an effective framework not only jointly studies the brain function and structure at a group-level, but also retains the individual differences among the subjects. In this study, a novel dense individualized and common connectivity-based cortical landmarks (DICCCOL)-based multi-modality graph neural networks (DM-GNN) framework is proposed to differentiate preterm and term infant brains and characterize the corresponding biomarkers. This framework adopts the DICCCOL system as the initialized graph node of GNN for each subject, utilizing both functional and structural profiles and effectively retaining the individual differences. To be specific, functional magnetic resonance imaging (fMRI) of the brain provides the features for the graph nodes, and brain fiber connectivity is utilized as the structural representation of the graph edges. Self-attention graph pooling (SAGPOOL)-based GNN is then applied to jointly study the function and structure of the brain and identify the biomarkers. Our results successfully demonstrate that the proposed framework can effectively differentiate the preterm and term infant brains. Furthermore, the self-attention-based mechanism can accurately calculate the attention score and recognize the most significant biomarkers. In this study, not only 87.6% classification accuracy is observed for the developing Human Connectome Project (dHCP) dataset, but also distinguishing features are explored and extracted. Our study provides a novel and uniform framework to differentiate brain disorders and characterize the corresponding biomarkers.

Brain Connectivity Based Graph Convolutional Networks and Its Application to Infant Age Prediction

Infancy is a critical period for the human brain development, and brain age is one of the indices for the brain development status associated with neuroimaging data. The difference between the predicted age based on neuroimaging and the chronological age can provide an important early indicator of deviation from the normal developmental trajectory. In this study, we utilize the Graph Convolutional Network (GCN) to predict the infant brain age based on resting-state fMRI data. The brain connectivity obtained from rs-fMRI can be represented as a graph with brain regions as nodes and functional connections as edges. However, since the brain connectivity is a fully connected graph with features on edges, current GCN cannot be directly used for it is a node-based method for sparse graphs. Hence, we propose an edge-based Graph Path Convolution (GPC) method, which aggregates the information from different paths and can be naturally applied on dense graphs. We refer the whole model as Brain Connectivity Graph Convolutional Networks (BC-GCN). Further, two upgraded network structures are proposed by including the residual and attention modules, referred as BC-GCN-Res and BC-GCN-SE to emphasize the information of the original data and enhance influential channels. Moreover, we design a two-stage coarse-to-fine framework, which determines the age group first and then predicts the age using group-specific BC-GCN-SE models. To avoid accumulated errors from the first stage, a cross-group training strategy is adopted for the second stage regression models. We conduct experiments on infant fMRI scans from 6 to 811 days of age. The coarse-to-fine framework shows significant improvements when being applied to several models (reducing error over 10 days). Comparing with state-of-the-art methods, our proposed model BC-GCN-SE with coarse-to-fine framework reduces the mean absolute error of the prediction from >70 days to 49.9 days. The code is now available at https://github.com/SCUT-Xinlab/BC-GCN.

Acute histologic chorioamnionitis independently and directly increases the risk for brain abnormalities seen on magnetic resonance imaging in very preterm infants

BACKGROUND

The independent risk for neurodevelopmental impairments attributed to chorioamnionitis in premature infants remains controversial. Delayed brain maturation or injury identified on magnetic resonance imaging at term-equivalent age can be used as a surrogate measure of neurodevelopmental impairments that is less confounded by postdelivery neonatal intensive care unit environmental factors to investigate this relationship more clearly.

OBJECTIVE

This study aimed to determine whether preterm infants born with moderate to severe acute histologic chorioamnionitis would have a higher magnetic resonance imaging-determined global brain abnormality score, independent of early premature birth, when compared with preterm infants with no or mild chorioamnionitis.

STUDY DESIGN

This was a prospective, multicenter cohort study involving infants born very prematurely ≤32 weeks' gestational age with acute moderate to severe histologic chorioamnionitis, graded using standard histologic criteria. Brain abnormalities were diagnosed and scored using a well-characterized, standardized scoring system captured using a high-resolution 3 Tesla magnetic resonance imaging research magnet. In secondary analyses, total brain volume and 4 magnetic resonance imaging metrics of cortical maturation (cortical surface area, sulcal depth, gyral index, and inner cortical curvature) were calculated using an automated algorithm and correlated with chorioamnionitis. The association of funisitis (any grade) with brain abnormalities was also explored. We investigated if premature birth mediated the relationship between histologic chorioamnionitis and brain abnormality score using mediation analysis.

RESULTS

Of 353 very preterm infants, 297 infants had mild or no chorioamnionitis (controls), and 56 were diagnosed with moderate to severe acute histologic chorioamnionitis. The primary outcome brain abnormality score was significantly higher in histologic chorioamnionitis-exposed infants than in the controls (median, 4 vs 7; P<.001). Infants with acute histologic chorioamnionitis had significantly lower brain tissue volume (P=.03) and sulcal depth (P=.04), whereas other morphometric indices did not differ statistically. In the multiple regression analysis, we observed persistent significant relationships between moderate to severe acute histologic chorioamnionitis and brain abnormality scores (β=2.84; 1.51-4.16; P<.001), total brain volume (P=.03), and sulcal depth (P=.02). Funisitis was also significantly associated with brain abnormality score after adjustment for clinical confounders (P=.005). Mediation analyses demonstrated that 50% of brain abnormalities was an indirect consequence of premature birth, and the remaining 50% was a direct effect of moderate to severe acute histologic chorioamnionitis when compared with preterm infants with no or mild chorioamnionitis exposure. Examining gestational age as a mediator, funisitis did not exert a significant direct effect on brain abnormalities after the significant indirect effects of preterm birth were accounted for.

CONCLUSION

Acute histologic chorioamnionitis increases the risk for brain injury and delayed maturation, both directly and indirectly, by inducing premature birth.

A self-training deep neural network for early prediction of cognitive deficits in very preterm infants using brain functional connectome data

BACKGROUND

Deep learning has been employed using brain functional connectome data for evaluating the risk of cognitive deficits in very preterm infants. Although promising, training these deep learning models typically requires a large amount of labeled data, and labeled medical data are often very difficult and expensive to obtain.

OBJECTIVE

This study aimed to develop a self-training deep neural network (DNN) model for early prediction of cognitive deficits at 2 years of corrected age in very preterm infants (gestational age ≤32 weeks) using both labeled and unlabeled brain functional connectome data.

MATERIALS AND METHODS

We collected brain functional connectome data from 343 very preterm infants at a mean (standard deviation) postmenstrual age of 42.7 (2.5) weeks, among whom 103 children had a cognitive assessment at 2 years (i.e. labeled data), and the remaining 240 children had not received 2-year assessments at the time this study was conducted (i.e. unlabeled data). To develop a self-training DNN model, we built an initial student model using labeled brain functional connectome data. Then, we applied the trained model as a teacher model to generate pseudo-labels for unlabeled brain functional connectome data. Next, we combined labeled and pseudo-labeled data to train a new student model. We iterated this procedure to obtain the best student model for the early prediction task in very preterm infants.

RESULTS

In our cross-validation experiments, the proposed self-training DNN model achieved an accuracy of 71.0%, a specificity of 71.5%, a sensitivity of 70.4% and an area under the curve of 0.75, significantly outperforming transfer learning models through pre-training approaches.

CONCLUSION

We report the first self-training prognostic study in very preterm infants, efficiently utilizing a small amount of labeled data with a larger share of unlabeled data to aid the model training. The proposed technique is expected to facilitate deep learning with insufficient training data.

Using deep clustering to improve fMRI dynamic functional connectivity analysis

Dynamic functional connectivity (dFC) analysis of resting-state fMRI data is commonly performed by calculating sliding-window correlations (SWC), followed by k-means clustering in order to assign each window to a given state. Studies using synthetic data have shown that k-means performance is highly dependent on sliding window parameters and signal-to-noise ratio. Additionally, sources of heterogeneity between subjects may affect the accuracy of group-level clustering, thus affecting measurements of dFC state temporal properties such as dwell time and fractional occupancy. This may result in spurious conclusions regarding differences between groups (e.g. when comparing a clinical population to healthy controls). Therefore, is it important to quantify the ability of k-means to estimate dFC state temporal properties when applied to cohorts of multiple subjects, and to explore ways in which clustering performance can be maximised. Here, we explore the use of dimensionality reduction methods prior to clustering in order to map high-dimensional data to a lower dimensional space, providing salient features to the subsequent clustering step. We assess the use of deep autoencoders for dimensionality reduction prior to applying k-means clustering to the encoded data. We compare this deep clustering method to dimensionality reduction using principle component analysis (PCA), uniform manifold approximation and projection (UMAP), as well as applying k-means to the original feature space using either L1 or L2 distance. We provide extensive quantitative evaluation of clustering performance using synthetic datasets, representing data from multiple heterogeneous subjects. In synthetic data we find that deep clustering gives the best performance, while other approaches are often insufficient to capture temporal properties of dFC states. We then demonstrate the application of each method to real-world data from human subjects and show that the choice of dimensionality reduction method has a significant effect on group-level measurements of state temporal properties.

Systematic Bibliometric and Visualized Analysis of Research Hotspots and Trends on Autism Spectrum Disorder Neuroimaging

Background

Autism spectrum disorder (ASD) is a chronic developmental disability caused by differences in the brain. The gold standard for the diagnosis of this condition is based on behavioral science, but research on the application of neurological detection to diagnose the atypical nervous system of ASD is ongoing. ASD neuroimaging research involves the examination of the brain's structure, functional connections, and neurometabolic. However, limited medical resource and the unique heterogeneity of ASD have resulted in many challenges when neuroimaging is utilized.

Objective

This bibliometric study is aimed at summarizing themes and trends in research on autism spectrum disorder neuroimaging and at proposing potential directions for future inquiry.

Methods

Citations were downloaded from the Web of Science Core Collection database on neuroimaging published from January 1, 2012, to December 31, 2021. The retrieved information was analyzed using Bibliometric.com, CiteSpace.5.8. R3, and VOS viewer.

Results

A total of 1,363 papers were published across 58 regions. The United States was the leading source of publications. The League of European Research Universities published the largest number of articles (171). Burst keywords from 2018 to 2021 include identification and network. The clusters of references that continued into 2020 included graph theory, functional connectivity, and classification, which represent key research topics.

Conclusions

Imaging data is being used to identify neuro-network models with higher accuracy for ASD discrimination. Functional near-infrared imaging is advantageous compared to other neuroimaging. In the future, research on systematic and accurate computer-aided diagnosis technology should be encouraged. Moreover, the study of neuroimaging of ASD in different psychological and behavioral states can inspire new ideas about the diagnosis and intervention training of ASD and should be explored.

SSPNet: An interpretable 3D-CNN for classification of schizophrenia using phase maps of resting-state complex-valued fMRI data

Convolutional neural networks (CNNs) have shown promising results in classifying individuals with mental disorders such as schizophrenia using resting-state fMRI data. However, complex-valued fMRI data is rarely used since additional phase data introduces high-level noise though it is potentially useful information for the context of classification. As such, we propose to use spatial source phase (SSP) maps derived from complex-valued fMRI data as the CNN input. The SSP maps are not only less noisy, but also more sensitive to spatial activation changes caused by mental disorders than magnitude maps. We build a 3D-CNN framework with two convolutional layers (named SSPNet) to fully explore the 3D structure and voxel-level relationships from the SSP maps. Two interpretability modules, consisting of saliency map generation and gradient-weighted class activation mapping (Grad-CAM), are incorporated into the well-trained SSPNet to provide additional information helpful for understanding the output. Experimental results from classifying schizophrenia patients (SZs) and healthy controls (HCs) show that the proposed SSPNet significantly improved accuracy and AUC compared to CNN using magnitude maps extracted from either magnitude-only (by 23.4 and 23.6% for DMN) or complex-valued fMRI data (by 10.6 and 5.8% for DMN). SSPNet captured more prominent HC-SZ differences in saliency maps, and Grad-CAM localized all contributing brain regions with opposite strengths for HCs and SZs within SSP maps. These results indicate the potential of SSPNet as a sensitive tool that may be useful for the development of brain-based biomarkers of mental disorders.

Machine Learning Prediction Models for Neurodevelopmental Outcome After Preterm Birth: A Scoping Review and New Machine Learning Evaluation Framework

BACKGROUND AND OBJECTIVES

Outcome prediction of preterm birth is important for neonatal care, yet prediction performance using conventional statistical models remains insufficient. Machine learning has a high potential for complex outcome prediction. In this scoping review, we provide an overview of the current applications of machine learning models in the prediction of neurodevelopmental outcomes in preterm infants, assess the quality of the developed models, and provide guidance for future application of machine learning models to predict neurodevelopmental outcomes of preterm infants.

METHODS

A systematic search was performed using PubMed. Studies were included if they reported on neurodevelopmental outcome prediction in preterm infants using predictors from the neonatal period and applying machine learning techniques. Data extraction and quality assessment were independently performed by 2 reviewers.

RESULTS

Fourteen studies were included, focusing mainly on very or extreme preterm infants, predicting neurodevelopmental outcome before age 3 years, and mostly assessing outcomes using the Bayley Scales of Infant Development. Predictors were most often based on MRI. The most prevalent machine learning techniques included linear regression and neural networks. None of the studies met all newly developed quality assessment criteria. Studies least prone to inflated performance showed promising results, with areas under the curve up to 0.86 for classification and R2 values up to 91% in continuous prediction. A limitation was that only 1 data source was used for the literature search.

CONCLUSIONS

Studies least prone to inflated prediction results are the most promising. The provided evaluation framework may contribute to improved quality of future machine learning models.

Artificial Intelligence in NICU and PICU: A Need for Ecological Validity, Accountability, and Human Factors

Pediatric patients, particularly in neonatal and pediatric intensive care units (NICUs and PICUs), are typically at an increased risk of fatal decompensation. That being said, any delay in treatment or minor errors in medication dosage can overcomplicate patient health. Under such an environment, clinicians are expected to quickly and effectively comprehend large volumes of medical information to diagnose and develop a treatment plan for any baby. The integration of Artificial Intelligence (AI) into the clinical workflow can be a potential solution to safeguard pediatric patients and augment the quality of care. However, before making AI an integral part of pediatric care, it is essential to evaluate the technology from a human factors perspective, ensuring its readiness (technology readiness level) and ecological validity. Addressing AI accountability is also critical to safeguarding clinicians and improving AI acceptance in the clinical workflow. This article summarizes the application of AI in NICU/PICU and consecutively identifies the existing flaws in AI (from clinicians’ standpoint), and proposes related recommendations, which, if addressed, can improve AIs’ readiness for a real clinical environment.

ConCeptCNN: A novel multi-filter convolutional neural network for the prediction of neurodevelopmental disorders using brain connectome

BACKGROUND

Deep convolutional neural network (CNN) and its derivatives have recently shown great promise in the prediction of brain disorders using brain connectome data. Existing deep CNN methods using single global row and column convolutional filters have limited ability to extract discriminative information from brain connectome for prediction tasks.

PURPOSE

This paper presents a novel deep Connectome-Inception CNN (ConCeptCNN) model, which is developed based on multiple convolutional filters. The proposed model is used to extract topological features from brain connectome data for neurological disorders classification and analysis.

METHODS

The ConCeptCNN uses multiple vector-shaped filters extract topological information from the brain connectome at different levels for complementary feature embeddings of brain connectome. The proposed model is validated using two datasets: the Neuro Bureau ADHD-200 dataset and the Cincinnati Early Prediction Study (CINEPS) dataset.

RESULTS

In a cross-validation experiment, the ConCeptCNN achieved a prediction accuracy of 78.7% for the detection of ADHD in adolescents and an accuracy of 81.6% for the prediction of cognitive deficits at 2-year corrected age in very preterm infants. In addition to the classification tasks, the ConCeptCNN identified several brain regions that are discriminative to neurodevelopmental disorders.

CONCLUSIONS

We compared the ConCeptCNN with several peer CNN methods. The results demonstrated that proposed model improves overall classification performance of neurodevelopmental disorders prediction tasks. This article is protected by copyright. All rights reserved.

Association of network connectivity via resting state functional MRI with consciousness, mortality, and outcomes in neonatal acute brain injury

•

Basal ganglia and seizure onset zone networks were associated with motor outcomes.

•

Broad language/cognitive region networks were associated with developmental delay.

•

Discharge with mortality was linked to default mode and language/cognitive networks.

•

Exams were not linked to networks after multiple testing corrections.

•

Lack of detection of all studied networks only occurred in those who did not survive.

Background

An accurate and comprehensive test of integrated brain network function is needed for neonates during the acute brain injury period to inform on morbidity. This retrospective cohort study assessed whether integrated brain network function acquired by resting state functional MRI during the acute period in neonates with brain injury, is associated with acute exam, neonatal mortality, and 6-month outcomes.

Methods

Study subjects included 40 consecutive neonates with resting state functional MRI acquired within 31 days after suspected brain insult from March 2018 to July 2019 at Phoenix Children’s Hospital. Acute-period exam and test results were assigned ordinal scores based on severity as documented by respective treating specialists. Analyses (Fisher exact, Wilcoxon-rank sum test, ordinal/multinomial logistic regression) examined association of resting state networks with demographics, presentation, neurological exam, electroencephalogram, anatomical MRI, magnetic resonance spectroscopy, passive task functional MRI, and outcomes of discharge condition, outpatient development, motor tone, seizure, and mortality.

Results

Subjects had a mean (standard deviation) gestational age of 37.8 (2.6) weeks, a majority were male (63%), with a diagnosis of hypoxic ischemic encephalopathy (68%). Findings at birth included mild distress (48%), moderately abnormal neurological exam (33%), and consciousness characterized as awake but irritable (40%). Significant associations after multiple testing corrections were detected for resting state networks: basal ganglia with outpatient developmental delay (odds ratio [OR], 14.5; 99.4% confidence interval [CI], 2.00–105; P < .001) and motor tone/weakness (OR, 9.98; 99.4% CI, 1.72–57.9; P < .001); language/frontoparietal network with discharge condition (OR, 5.13; 99.4% CI, 1.22–21.5; P = .002) and outpatient developmental delay (OR, 4.77; 99.4% CI, 1.21–18.7; P=.002); default mode network with discharge condition (OR, 3.72; 99.4% CI, 1.01–13.78; P=.006) and neurological exam (P = .002 (FE); OR, 11.8; 99.4% CI, 0.73–191; P = .01 (OLR)); and seizure onset zone with motor tone/weakness (OR, 3.31; 99.4% CI, 1.08–10.1; P=.003). Resting state networks were not detected in three neonates, who died prior to discharge.

Conclusions

This study provides level 3 evidence (OCEBM Levels of Evidence Working Group) demonstrating that in neonatal acute brain injury, the degree of abnormality of resting state networks is associated with acute exam and outcomes. Total lack of brain network detection was only found in patients who did not survive.

Cognitive Outcome Prediction in Infants With Neonatal Hypoxic-Ischemic Encephalopathy Based on Functional Connectivity and Complexity of the Electroencephalography Signal

Impaired neurodevelopmental outcome, in particular cognitive impairment, after neonatal hypoxic-ischemic encephalopathy is a major concern for parents, clinicians, and society. This study aims to investigate the potential benefits of using advanced quantitative electroencephalography analysis (qEEG) for early prediction of cognitive outcomes, assessed here at 2 years of age. EEG data were recorded within the first week after birth from a cohort of twenty infants with neonatal hypoxic-ischemic encephalopathy (HIE). A proposed regression framework was based on two different sets of features, namely graph-theoretical features derived from the weighted phase-lag index (WPLI) and entropies metrics represented by sample entropy (SampEn), permutation entropy (PEn), and spectral entropy (SpEn). Both sets of features were calculated within the noise-assisted multivariate empirical mode decomposition (NA-MEMD) domain. Correlation analysis showed a significant association in the delta band between the proposed features, graph attributes (radius, transitivity, global efficiency, and characteristic path length) and entropy features (Pen and SpEn) from the neonatal EEG data and the cognitive development at age two years. These features were used to train and test the tree ensemble (boosted and bagged) regression models. The highest prediction performance was reached to 14.27 root mean square error (RMSE), 12.07 mean absolute error (MAE), and 0.45 R-squared using the entropy features with a boosted tree regression model. Thus, the results demonstrate that the proposed qEEG features show the state of brain function at an early stage; hence, they could serve as predictive biomarkers of later cognitive impairment, which could facilitate identifying those who might benefit from early targeted intervention.

Artificial Intelligence to Improve Health Outcomes in the NICU and PICU: A Systematic Review

CONTEXT

Artificial intelligence (AI) technologies are increasingly used in pediatrics and have the potential to help inpatient physicians provide high-quality care for critically ill children.

OBJECTIVE

We aimed to describe the use of AI to improve any health outcome(s) in neonatal and pediatric intensive care.

DATA SOURCE

PubMed, IEEE Xplore, Cochrane, and Web of Science databases.

STUDY SELECTION

We used peer-reviewed studies published between June 1, 2010, and May 31, 2020, in which researchers described (1) AI, (2) pediatrics, and (3) intensive care. Studies were included if researchers assessed AI use to improve at least 1 health outcome (eg, mortality).

DATA EXTRACTION

Data extraction was conducted independently by 2 researchers. Articles were categorized by direct or indirect impact of AI, defined by the European Institute of Innovation and Technology Health joint report.

RESULTS

Of the 287 publications screened, 32 met inclusion criteria. Approximately 22% (n = 7) of studies revealed a direct impact and improvement in health outcomes after AI implementation. Majority were in prototype testing, and few were deployed into an ICU setting. Among the remaining 78% (n = 25) AI models outperformed standard clinical modalities and may have indirectly influenced patient outcomes. Quantitative assessment of health outcomes using statistical measures, such as area under the receiver operating curve (56%; n = 18) and specificity (38%; n = 12), revealed marked heterogeneity in metrics and standardization.

CONCLUSIONS

Few studies have revealed that AI has directly improved health outcomes for pediatric critical care patients. Further prospective, experimental studies are needed to assess AI's impact by using established implementation frameworks, standardized metrics, and validated outcome measures.

Deep Multimodal Learning From MRI and Clinical Data for Early Prediction of Neurodevelopmental Deficits in Very Preterm Infants

The prevalence of disabled survivors of prematurity has increased dramatically in the past 3 decades. These survivors, especially, very preterm infants (VPIs), born ≤ 32 weeks gestational age, are at high risk for neurodevelopmental impairments. Early and clinically effective personalized prediction of outcomes, which forms the basis for early treatment decisions, is urgently needed during the peak neuroplasticity window—the first couple of years after birth—for at-risk infants, when intervention is likely to be most effective. Advances in MRI enable the noninvasive visualization of infants' brains through acquired multimodal images, which are more informative than unimodal MRI data by providing complementary/supplementary depicting of brain tissue characteristics and pathology. Thus, analyzing quantitative multimodal MRI features affords unique opportunities to study early postnatal brain development and neurodevelopmental outcome prediction in VPIs. In this study, we investigated the predictive power of multimodal MRI data, including T2-weighted anatomical MRI, diffusion tensor imaging, resting-state functional MRI, and clinical data for the prediction of neurodevelopmental deficits. We hypothesize that integrating multimodal MRI and clinical data improves the prediction over using each individual data modality. Employing the aforementioned multimodal data, we proposed novel end-to-end deep multimodal models to predict neurodevelopmental (i.e., cognitive, language, and motor) deficits independently at 2 years corrected age. We found that the proposed models can predict cognitive, language, and motor deficits at 2 years corrected age with an accuracy of 88.4, 87.2, and 86.7%, respectively, significantly better than using individual data modalities. This current study can be considered as proof-of-concept. A larger study with external validation is important to validate our approach to further assess its clinical utility and overall generalizability.

Convolutional Neural Network with Sparse Strategies to Classify Dynamic Functional Connectivity

Classification of dynamic functional connectivity (DFC) is becoming a promising approach for diagnosing various neurodegenerative diseases. However, the existing methods generally face the problem of overfitting. To solve it, this paper proposes a convolutional neural network with three sparse strategies named SCNN to classify DFC. Firstly, an element-wise filter is designed to impose sparse constraints on the DFC matrix by replacing the redundant elements with zeroes, where the DFC matrix is specially constructed to quantify the spatial and temporal variation of DFC. Secondly, a 11 convolutional filter is adopted to reduce the dimensionality of the sparse DFC matrix, and remove meaningless features resulted from zero elements in the subsequent convolution process. Finally, an extra sparse optimization classifier is employed to optimize the parameters of the above two filters, which can effectively improve the ability of SCNN to extract discriminative features. Experimental results on multiple resting-state fMRI datasets demonstrate that the proposed model provides a better classification performance of DFC compared with several state-of-the-art methods, and can identify the abnormal brain functional connectivity.

Neonatal spectral EEG is prognostic of cognitive abilities at school age in premature infants without overt brain damage

Prematurity is a prototype of biological risk that could affect the late neurocognitive outcome; however, the condition itself remains a non-specific marker. This longitudinal 6-year study aimed to evaluate the prognostic role of neonatal spectral EEG in premature infants without neurological complications. The study cohort was 26 children born 23-34 gestational ages; all neonates underwent multichannel EEG recordings at 35 weeks post-conception. EEG data were transformed into the frequency domain and divided into delta (0.5-4 Hz), theta (5-7 Hz), alpha (8-13 Hz), and beta (14-20 Hz) frequency bands. At 6 years, a neuropsychological and behavioral evaluation was performed. Correlations between spectral bands and neuropsychological assessments were performed with a conservative and robust Bayesian correlation model using weakly informative priors. The correlation of neuropsychological tasks to spectral frequency bands highlighted a significant association with visual and auditory attention tests. The performance on the same tests appears to be mainly impaired.Conclusions: We found that spectral EEG frequencies are independent predictors of performance in attention tasks. We hypothesized that spectral EEG might reflect early circuitries' imbalance in the reticular ascending system and cumulative effect on ongoing development, pointing to the importance of early prognostic instruments. What is Known: • Prematurity is a non-specific marker of late neurocognitive risk. • Precise prognostic instruments are lacking, mostly in patients with low-grade conditions. What is New: • Longitudinal long-term studies are scarce but crucial for the inferential attributive process. • Spectral EEG frequencies are independent predictors of performance in attention tasks.

Bedside functional monitoring of the dynamic brain connectivity in human neonates

Clinicians have long been interested in functional brain monitoring, as reversible functional losses often precedes observable irreversible structural insults. By characterizing neonatal functional cerebral networks, resting-state functional connectivity is envisioned to provide early markers of cognitive impairments. Here we present a pioneering bedside deep brain resting-state functional connectivity imaging at 250-μm resolution on human neonates using functional ultrasound. Signal correlations between cerebral regions unveil interhemispheric connectivity in very preterm newborns. Furthermore, fine-grain correlations between homologous pixels are consistent with white/grey matter organization. Finally, dynamic resting-state connectivity reveals a significant occurrence decrease of thalamo-cortical networks for very preterm neonates as compared to control term newborns. The same method also shows abnormal patterns in a congenital seizure disorder case compared with the control group. These results pave the way to infants' brain continuous monitoring and may enable the identification of abnormal brain development at the bedside.

The cingulum in very preterm infants relates to language and social-emotional impairment at 2 years of term-equivalent age

•

Maturation of specific WM tracts in preterm individuals differs from those of term controls.

•

The elastic net logistic regression model was used to identify altered white matter tracts in the preterm brain.

•

The alteration of the cingulum in the preterm at near-term correlate with neurodevelopmental scores at 18–22 months of age.

Background

Relative to full-term infants, very preterm infants exhibit disrupted white matter (WM) maturation and problems related to development, including motor, cognitive, social-emotional, and receptive and expressive language processing.

Objective

The present study aimed to determine whether regional abnormalities in the WM microstructure of very preterm infants, as defined relative to those of full-term infants at a near-term age, are associated with neurodevelopmental outcomes at the age of 18–22 months.

Methods

We prospectively enrolled 89 very preterm infants (birth weight < 1500 g) and 43 normal full-term control infants born between 2016 and 2018. All infants underwent a structural brain magnetic resonance imaging scan at near-term age. The diffusion tensor imaging (DTI) metrics of the whole-brain WM tracts were extracted based on the neonatal probabilistic WM pathway. The elastic net logistic regression model was used to identify altered WM tracts in the preterm brain. We evaluated the associations between the altered WM microstructure at near-term age and motor, cognitive, social-emotional, and receptive and expressive language developments at 18–22 months of age, as measured using the Bayley Scales of Infant Development, Third Edition.

Results

We found that the elastic net logistic regression model could classify preterm and full-term neonates with an accuracy of 87.9% (corrected p < 0.008) using the DTI metrics in the pathway of interest with a 10% threshold level. The fractional anisotropy (FA) values of the body and splenium of the corpus callosum, middle cerebellar peduncle, left and right uncinate fasciculi, and right portion of the pathway between the premotor and primary motor cortices (premotor-PMC), as well as the mean axial diffusivity (AD) values of the left cingulum, were identified as contributive features for classification. Increased adjusted AD values in the left cingulum pathway were significantly correlated with language scores after false discovery rate (FDR) correction (r = 0.217, p = 0.043). The expressive language and social-emotional composite scores showed a significant positive correlation with the AD values in the left cingulum pathway (r = 0.226 [p = 0.036] and r = 0.31 [p = 0.003], respectively) after FDR correction.

Conclusion

Our approach suggests that the cingulum pathways of very preterm infants differ from those of full-term infants and significantly contribute to the prediction of the subsequent development of the language and social-emotional domains. This finding could improve our understanding of how specific neural substrates influence neurodevelopment at later ages, and individual risk prediction, thus helping to inform early intervention strategies that address developmental delay.

Neonatal brain connectivity outliers identify over forty percent of IQ outliers at 4 years of age

BACKGROUND

Defining reliable brain markers for the prediction of abnormal behavioral outcomes remains an urgent but extremely challenging task in neuroscience research. This is particularly important for infant studies given the most dramatic brain and behavioral growth during infancy.

METHODS

In this study, we proposed a novel prediction scheme through abstracting individual newborn's whole-brain functional connectivity pattern to three outlier measures (Triple O) and tested the hypothesis that neonates identified as "brain outliers" based on Triple O were more likely to develop as IQ outliers at 4 years of age. Without need for training with behavioral data, Triple O represents a novel proof-of-concept approach to predict later IQ outcomes based on neonatal brain data.

RESULTS

Triple O correctly identified 42.1% true IQ outliers among a mixed cohort of 175 newborns with different term, twin, and maternal disorder statuses. Triple O also reached a high level of specificity (96.2%) and overall accuracy (90.3%). Further incorporating a demographic information indicator, the enhanced Triple O+ could further differentiate between high and low 4YR IQ outliers. Validation tests against seven independent reference samples revealed highly consistent results and a minimum sample size of ~50 for robust performance.

CONCLUSIONS

Considering that postnatal brain growth and various environmental factors likely also contribute to 4YR IQ, the fact that Triple O, based purely on neonatal functional connectivity data, could identify >40% of 4YR IQ outliers is striking. Together with the very high level of specificity, each outlier predicted by Triple O represents a meaningful risk but future efforts are needed to explore ways to identify the rest of outliers. Overall, with no need for training, a high level of robustness, and a minimal requirement on sample size, the proposed Triple O approach demonstrates great potential to predict later outlying IQ performances using neonatal functional connectivity data.

Early Prediction of Cognitive Deficit in Very Preterm Infants Using Brain Structural Connectome With Transfer Learning Enhanced Deep Convolutional Neural Networks

Up to 40% of very preterm infants (≤32 weeks’ gestational age) were identified with a cognitive deficit at 2 years of age. Yet, accurate clinical diagnosis of cognitive deficit cannot be made until early childhood around 3–5 years of age. Recently, brain structural connectome that was constructed by advanced diffusion tensor imaging (DTI) technique has been playing an important role in understanding human cognitive functions. However, available annotated neuroimaging datasets with clinical and outcome information are usually limited and expensive to enlarge in the very preterm infants’ studies. These challenges hinder the development of neonatal prognostic tools for early prediction of cognitive deficit in very preterm infants. In this study, we considered the brain structural connectome as a 2D image and applied established deep convolutional neural networks to learn the spatial and topological information of the brain connectome. Furthermore, the transfer learning technique was utilized to mitigate the issue of insufficient training data. As such, we developed a transfer learning enhanced convolutional neural network (TL-CNN) model for early prediction of cognitive assessment at 2 years of age in very preterm infants using brain structural connectome. A total of 110 very preterm infants were enrolled in this work. Brain structural connectome was constructed using DTI images scanned at term-equivalent age. Bayley III cognitive assessments were conducted at 2 years of corrected age. We applied the proposed model to both cognitive deficit classification and continuous cognitive score prediction tasks. The results demonstrated that TL-CNN achieved improved performance compared to multiple peer models. Finally, we identified the brain regions most discriminative to the cognitive deficit. The results suggest that deep learning models may facilitate early prediction of later neurodevelopmental outcomes in very preterm infants at term-equivalent age.

A multi-task, multi-stage deep transfer learning model for early prediction of neurodevelopment in very preterm infants

Survivors following very premature birth (i.e., ≤ 32 weeks gestational age) remain at high risk for neurodevelopmental impairments. Recent advances in deep learning techniques have made it possible to aid the early diagnosis and prognosis of neurodevelopmental deficits. Deep learning models typically require training on large datasets, and unfortunately, large neuroimaging datasets with clinical outcome annotations are typically limited, especially in neonates. Transfer learning represents an important step to solve the fundamental problem of insufficient training data in deep learning. In this work, we developed a multi-task, multi-stage deep transfer learning framework using the fusion of brain connectome and clinical data for early joint prediction of multiple abnormal neurodevelopmental (cognitive, language and motor) outcomes at 2 years corrected age in very preterm infants. The proposed framework maximizes the value of both available annotated and non-annotated data in model training by performing both supervised and unsupervised learning. We first pre-trained a deep neural network prototype in a supervised fashion using 884 older children and adult subjects, and then re-trained this prototype using 291 neonatal subjects without supervision. Finally, we fine-tuned and validated the pre-trained model using 33 preterm infants. Our proposed model identified very preterm infants at high-risk for cognitive, language, and motor deficits at 2 years corrected age with an area under the receiver operating characteristic curve of 0.86, 0.66 and 0.84, respectively. Employing such a deep learning model, once externally validated, may facilitate risk stratification at term-equivalent age for early identification of long-term neurodevelopmental deficits and targeted early interventions to improve clinical outcomes in very preterm infants.

Machine Learning in Neuroimaging: A New Approach to Understand Acupuncture for Neuroplasticity

The effects of acupuncture facilitating neural plasticity for treating diseases have been identified by clinical and experimental studies. In the last two decades, the application of neuroimaging techniques in acupuncture research provided visualized evidence for acupuncture promoting neuroplasticity. Recently, the integration of machine learning (ML) and neuroimaging techniques becomes a focus in neuroscience and brings a new and promising approach to understand the facilitation of acupuncture on neuroplasticity at the individual level. This review is aimed at providing an overview of this rapidly growing field by introducing the commonly used ML algorithms in neuroimaging studies briefly and analyzing the characteristics of the acupuncture studies based on ML and neuroimaging, so as to provide references for future research.

Deep Neural Network for Automatic Classification of Pathological Voice Signals

OBJECTIVES

Computer-aided pathological voice detection is efficient for initial screening of pathological voice, and has received high academic and clinical attention. This paper proposes an automatic diagnosis method of pathological voice based on deep neural network (DNN). Other two classification models (support vector machines and random forests) were used to verify the effectiveness of DNN.

METHODS

In this paper, we extracted 12 Mel frequency cepstral coefficients of each voice sample as row features. The constructed DNN consists a two-layer stacked sparse autoencoders network and a softmax layer. The stacked sparse autoencoders layer can learn high-level features from raw Mel frequency cepstral coefficients features. Then, the softmax layer can diagnose pathological voice according to high-level features. The DNN and the other two comparison models used the same train set and test set for the experiment.

RESULTS

Experimental results reveal that the value of sensitivity, specificity, precision, accuracy, and F1 score of the DNN can reach 97.8%, 99.4%, 99.4%, 98.6%, and 98.4%, respectively. The five indexes of DNN classification results are at least 6.2%, 5%, 5.6%, 5.7%, and 6.2% higher than the comparison models (support vector machine and random forest).

CONCLUSIONS

The proposed DNN can learn advanced features from raw acoustic features, and distinguish pathological voice from healthy voice. To the extent of this preliminary study, future studies can further explore the application of DNN in other experiments and clinical practice.

Early brain abnormalities in infants born very preterm predict under-reactive temperament

BACKGROUND

Early temperament may mediate the association between brain abnormalities following preterm birth and neurodevelopmental outcomes.

AIMS

This exploratory study investigated whether brain abnormalities in infants born very preterm predicted temperament.

STUDY DESIGN

Infants born prematurely underwent structural MRI at term. Mother self-reported depression symptoms at the scanning visit, and the Infant Behavior Questionnaire-Revised-Short (IBQ-R-S) about their infant at 3-months corrected age.

SUBJECTS

Infants (n = 214) born at ≤32 weeks gestation (M = 29.29, SD = 2.60). Average post-menstrual age at the MRI scan was 42.72 weeks (SD = 1.30). The majority of the infants were male (n = 115), and Caucasian (n = 145) or African American (n = 58). The average birthweight in grams was 1289.75 (SD = 448.5).

OUTCOME MEASURES

Infant Behavior Questionnaire-Revised-Short (IBQ-R-S) subscales.

RESULTS

Multivariate regression showed white matter abnormalities predicted lower ratings on High Intensity Pleasure and Vocal Reactivity, grey matter abnormalities predicted lower ratings on High Intensity Pleasure and Cuddliness, and cerebellar abnormalities predicted lower ratings on Fear and Sadness IBQ-R-S subscales adjusting for gestational age and sex. The pattern of results was essentially unchanged when maternal depression and socioeconomic status were included in the model.

CONCLUSIONS

Early MRI-diagnosed brain abnormalities in infants born very preterm were associated less vocalization and engagement during cuddling, decreased ability to take pleasure in stimulating activities, and lower emotionality in fear and sadness domains. Although replication is warranted, an under-reactive temperament in infants born preterm may have a neurobiological basis.

Predicting motor outcome in preterm infants from very early brain diffusion MRI using a deep learning convolutional neural network (CNN) model

BACKGROUND AND AIMS

Preterm birth imposes a high risk for developing neuromotor delay. Earlier prediction of adverse outcome in preterm infants is crucial for referral to earlier intervention. This study aimed to predict abnormal motor outcome at 2 years from early brain diffusion magnetic resonance imaging (MRI) acquired between 29 and 35 weeks postmenstrual age (PMA) using a deep learning convolutional neural network (CNN) model.

METHODS

Seventy-seven very preterm infants (born <31 weeks gestational age (GA)) in a prospective longitudinal cohort underwent diffusion MR imaging (3T Siemens Trio; 64 directions, b ​= ​2000 ​s/mm²). Motor outcome at 2 years corrected age (CA) was measured by Neuro-Sensory Motor Developmental Assessment (NSMDA). Scores were dichotomised into normal (functional score: 0, normal; n ​= ​48) and abnormal scores (functional score: 1-5, mild-profound; n ​= ​29). MRIs were pre-processed to reduce artefacts, upsampled to 1.25 ​mm isotropic resolution and maps of fractional anisotropy (FA) were estimated. Patches extracted from each image were used as inputs to train a CNN, wherein each image patch predicted either normal or abnormal outcome. In a postprocessing step, an image was classified as predicting abnormal outcome if at least 27% (determined by a grid search to maximise the model performance) of its patches predicted abnormal outcome. Otherwise, it was considered as normal. Ten-fold cross-validation was used to estimate performance. Finally, heatmaps of model predictions for patches in abnormal scans were generated to explore the locations associated with abnormal outcome.

RESULTS

For the identification of infants with abnormal motor outcome based on the FA data from early MRI, we achieved mean sensitivity 70% (standard deviation SD 19%), mean specificity 74% (SD 39%), mean AUC (area under the receiver operating characteristic curve) 72% (SD 14%), mean F1 score of 68% (SD 13%) and mean accuracy 73% (SD 19%) on an unseen test data set. Patch-based prediction heatmaps showed that the patches around the motor cortex and somatosensory regions were most frequently identified by the model with high precision (74%) as a location associated with abnormal outcome. Part of the cerebellum, and occipital and frontal lobes were also highly associated with abnormal NSMDA/motor outcome.

DISCUSSION/CONCLUSION

This study established the potential of an early brain MRI-based deep learning CNN model to identify preterm infants at risk of a later motor impairment and to identify brain regions predictive of adverse outcome. Results suggest that predictions can be made from FA maps of diffusion MRIs well before term equivalent age (TEA) without any prior knowledge of which MRI features to extract and associated feature extraction steps. This method, therefore, is suitable for any case of brain condition/abnormality. Future studies should be conducted on a larger cohort to re-validate the robustness and effectiveness of these models.

Deep Learning Techniques for Automatic Detection of Embryonic Neurodevelopmental Disorders

The increasing rates of neurodevelopmental disorders (NDs) are threatening pregnant women, parents, and clinicians caring for healthy infants and children. NDs can initially start through embryonic development due to several reasons. Up to three in 1000 pregnant women have embryos with brain defects; hence, the primitive detection of embryonic neurodevelopmental disorders (ENDs) is necessary. Related work done for embryonic ND classification is very limited and is based on conventional machine learning (ML) methods for feature extraction and classification processes. Feature extraction of these methods is handcrafted and has several drawbacks. Deep learning methods have the ability to deduce an optimum demonstration from the raw images without image enhancement, segmentation, and feature extraction processes, leading to an effective classification process. This article proposes a new framework based on deep learning methods for the detection of END. To the best of our knowledge, this is the first study that uses deep learning techniques for detecting END. The framework consists of four stages which are transfer learning, deep feature extraction, feature reduction, and classification. The framework depends on feature fusion. The results showed that the proposed framework was capable of identifying END from embryonic MRI images of various gestational ages. To verify the efficiency of the proposed framework, the results were compared with related work that used embryonic images. The performance of the proposed framework was competitive. This means that the proposed framework can be successively used for detecting END.

A Multichannel Deep Neural Network Model Analyzing Multiscale Functional Brain Connectome Data for Attention Deficit Hyperactivity Disorder Detection

Purpose

To develop a multichannel deep neural network (mcDNN) classification model based on multiscale brain functional connectome data and demonstrate the value of this model by using attention deficit hyperactivity disorder (ADHD) detection as an example.

Materials and Methods

In this retrospective case-control study, existing data from the Neuro Bureau ADHD-200 dataset consisting of 973 participants were used. Multiscale functional brain connectomes based on both anatomic and functional criteria were constructed. The mcDNN model used the multiscale brain connectome data and personal characteristic data (PCD) as joint features to detect ADHD and identify the most predictive brain connectome features for ADHD diagnosis. The mcDNN model was compared with single-channel deep neural network (scDNN) models and the classification performance was evaluated through cross-validation and hold-out validation with the metrics of accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC).

Results

In the cross-validation, the mcDNN model using combined features (fusion of the multiscale brain connectome data and PCD) achieved the best performance in ADHD detection with an AUC of 0.82 (95% confidence interval [CI]: 0.80, 0.83) compared with scDNN models using the features of the brain connectome at each individual scale and PCD, independently. In the hold-out validation, the mcDNN model achieved an AUC of 0.74 (95% CI: 0.73, 0.76).

Conclusion

An mcDNN model was developed for multiscale brain functional connectome data, and its utility for ADHD detection was demonstrated. By fusing the multiscale brain connectome data, the mcDNN model improved ADHD detection performance considerably over the use of a single scale.© RSNA, 2019.

Fetal Brain Abnormality Classification from MRI Images of Different Gestational Age

Magnetic resonance imaging (MRI) is a common imaging technique used extensively to study human brain activities. Recently, it has been used for scanning the fetal brain. Amongst 1000 pregnant women, 3 of them have fetuses with brain abnormality. Hence, the primary detection and classification are important. Machine learning techniques have a large potential in aiding the early detection of these abnormalities, which correspondingly could enhance the diagnosis process and follow up plans. Most research focused on the classification of abnormal brains in a primary age has been for newborns and premature infants, with fewer studies focusing on images for fetuses. These studies associated fetal scans to scans after birth for the detection and classification of brain defects early in the neonatal age. This type of brain abnormality is named small for gestational age (SGA). This article proposes a novel framework for the classification of fetal brains at an early age (before the fetus is born). As far as we could know, this is the first study to classify brain abnormalities of fetuses of widespread gestational ages (GAs). The study incorporates several machine learning classifiers, such as diagonal quadratic discriminates analysis (DQDA), K-nearest neighbour (K-NN), random forest, naïve Bayes, and radial basis function (RBF) neural network classifiers. Moreover, several bagging and Adaboosting ensembles models have been constructed using random forest, naïve Bayes, and RBF network classifiers. The performances of these ensembles have been compared with their individual models. Our results show that our novel approach can successfully identify and classify numerous types of defects within MRI images of the fetal brain of various GAs. Using the KNN classifier, we were able to achieve the highest classification accuracy and area under receiving operating characteristics of 95.6% and 99% respectively. In addition, ensemble classifiers improved the results of their respective individual models.

Assessment of lipid peroxidation and artificial neural network models in early Alzheimer Disease diagnosis

OBJECTIVE

Lipid peroxidation constitutes a molecular mechanism involved in early Alzheimer Disease (AD) stages, and artificial neural network (ANN) analysis is a promising non-linear regression model, characterized by its high flexibility and utility in clinical diagnosis. ANN simulates neuron learning procedures and it could provide good diagnostic performances in this complex and heterogeneous disease compared with linear regression analysis.

DESIGN AND METHODS

In our study, a new set of lipid peroxidation compounds were determined in urine and plasma samples from patients diagnosed with early Alzheimer Disease (n = 70) and healthy controls (n = 26) by means of ultra-performance liquid chromatography coupled with tandem mass-spectrometry. Then, a model based on ANN was developed to classify groups of participants.

RESULTS

The diagnostic performances obtained using an ANN model for each biological matrix were compared with the corresponding linear regression model based on partial least squares (PLS), and with the non-linear (radial and polynomial) support vector machine (SVM) models. Better accuracy, in terms of receiver operating characteristic-area under curve (ROC-AUC), was obtained for the ANN models (ROC-AUC 0.882 in plasma and 0.839 in urine) than for PLS and SVM models.

CONCLUSION

Lipid peroxidation and ANN constitute a useful approach to establish a reliable diagnosis when the prognosis is complex, multidimensional and non-linear.

Objective and Automated Detection of Diffuse White Matter Abnormality in Preterm Infants Using Deep Convolutional Neural Networks

Diffuse white matter abnormality (DWMA), or diffuse excessive high signal intensity is observed in 50-80% of very preterm infants at term-equivalent age. It is subjectively defined as higher than normal signal intensity in periventricular and subcortical white matter in comparison to normal unmyelinated white matter on T2-weighted MRI images. Despite the well-documented presence of DWMA, it remains debatable whether DWMA represents pathological tissue injury or a transient developmental phenomenon. Manual tracing of DWMA exhibits poor reliability and reproducibility and unduly increases image processing time. Thus, objective and ideally automatic assessment is critical to accurately elucidate the biologic nature of DWMA. We propose a deep learning approach to automatically identify DWMA regions on T2-weighted MRI images. Specifically, we formulated DWMA detection as an image voxel classification task; that is, the voxels on T2-weighted images are treated as samples and exclusively assigned as DWMA or normal white matter voxel classes. To utilize the spatial information of individual voxels, small image patches centered on the given voxels are retrieved. A deep convolutional neural networks (CNN) model was developed to differentiate DWMA and normal voxels. We tested our deep CNN in multiple validation experiments. First, we examined DWMA detection accuracy of our CNN model using computer simulations. This was followed by in vivo assessments in a cohort of very preterm infants (N = 95) using cross-validation and holdout validation. Finally, we tested our approach on an independent preterm cohort (N = 28) to externally validate our model. Our deep CNN model achieved Dice similarity index values ranging from 0.85 to 0.99 for DWMA detection in the aforementioned validation experiments. Our proposed deep CNN model exhibited significantly better performance than other popular machine learning models. We present an objective and automated approach for accurately identifying DWMA that may facilitate the clinical diagnosis of DWMA in very preterm infants.

Enhancing Diagnosis of Autism With Optimized Machine Learning Models and Personal Characteristic Data

Autism spectrum disorder (ASD) is a developmental disorder, affecting about 1% of the global population. Currently, the only clinical method for diagnosing ASD are standardized ASD tests which require prolonged diagnostic time and increased medical costs. Our objective was to explore the predictive power of personal characteristic data (PCD) from a large well-characterized dataset to improve upon prior diagnostic models of ASD. We extracted six personal characteristics (age, sex, handedness, and three individual measures of IQ) from 851 subjects in the Autism Brain Imaging Data Exchange (ABIDE) database. ABIDE is an international collaborative project that collected data from a large number of ASD patients and typical non-ASD controls from 17 research and clinical institutes. We employed this publicly available database to test nine supervised machine learning models. We implemented a cross-validation strategy to train and test those machine learning models for classification between typical non-ASD controls and ASD patients. We assessed classification performance using accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). Of the nine models we tested using six personal characteristics, the neural network model performed the best with a mean AUC (SD) of 0.646 (0.005), followed by k-nearest neighbor with a mean AUC (SD) of 0.641 (0.004). This study established an optimal ASD classification performance with PCD as features. With additional discriminative features (e.g., neuroimaging), machine learning models may ultimately enable automated clinical diagnosis of autism.

A Novel Transfer Learning Approach to Enhance Deep Neural Network Classification of Brain Functional Connectomes

Early diagnosis remains a significant challenge for many neurological disorders, especially for rare disorders where studying large cohorts is not possible. A novel solution that investigators have undertaken is combining advanced machine learning algorithms with resting-state functional Magnetic Resonance Imaging to unveil hidden pathological brain connectome patterns to uncover diagnostic and prognostic biomarkers. Recently, state-of-the-art deep learning techniques are outperforming traditional machine learning methods and are hailed as a milestone for artificial intelligence. However, whole brain classification that combines brain connectome with deep learning has been hindered by insufficient training samples. Inspired by the transfer learning strategy employed in computer vision, we exploited previously collected resting-state functional MRI data for healthy subjects from existing databases and transferred this knowledge for new disease classification tasks. We developed a deep transfer learning neural network (DTL-NN) framework for enhancing the classification of whole brain functional connectivity patterns. Briefly, we trained a stacked sparse autoencoder (SSAE) prototype to learn healthy functional connectivity patterns in an offline learning environment. Then, the SSAE prototype was transferred to a DTL-NN model for a new classification task. To test the validity of our framework, we collected resting-state functional MRI data from the Autism Brain Imaging Data Exchange (ABIDE) repository. Using autism spectrum disorder (ASD) classification as a target task, we compared the performance of our DTL-NN approach with a traditional deep neural network and support vector machine models across four ABIDE data sites that enrolled at least 60 subjects. As compared to traditional models, our DTL-NN approach achieved an improved performance in accuracy, sensitivity, specificity and area under receiver operating characteristic curve. These findings suggest that DTL-NN approaches could enhance disease classification for neurological conditions, where accumulating large neuroimaging datasets has been challenging.