BrainNetCNN: Convolutional neural networks for brain networks; towards predicting neurodevelopment

Kawahara, Jeremy; Brown, Colin J.; Miller, Steven P.; Booth, Brian G.; Chau, Vann; Grunau, Ruth E.; Zwicker, Jill G.; Hamarneh, Ghassan

doi:10.1016/j.neuroimage.2016.09.046

Cited by 472 publications

(369 citation statements)

References 39 publications

Supporting

Mentioning

356

Contrasting

Unclassified

Order By: Relevance

“…They also used a machine learning approach to find the regions that contribute more to classify between those two groups. Kawahara et al (2017) specifically adapted a convolutional neural network framework to structural brain networks and demonstrated the capacity to predict neurodevelopmental scores at 2 years of age based in structural brain networks at birth. Although showing a relatively modest predictive power (correlations up to .3), the approach of predicting continuous variables instead of dichotomised indices shows great potential.…”

Section: Machine Learning As a Solution?mentioning

confidence: 99%

Annual Research Review: Not just a small adult brain: understanding later neurodevelopment through imaging the neonatal brain

Batalle

Edwards

O’Muircheartaigh

2017

Child Psychology Psychiatry

View full text Add to dashboard Cite

BackgroundThere has been a recent proliferation in neuroimaging research focusing on brain development in the prenatal, neonatal and very early childhood brain. Early brain injury and preterm birth are associated with increased risk of neurodevelopmental disorders, indicating the importance of this early period for later outcome.Scope and methodologyAlthough using a wide range of different methodologies and investigating diverse samples, the common aim of many of these studies has been to both track normative development and investigate deviations in this development to predict behavioural, cognitive and neurological function in childhood. Here we review structural and functional neuroimaging studies investigating the developing brain. We focus on practical and technical complexities of studying this early age range and discuss how neuroimaging techniques have been successfully applied to investigate later neurodevelopmental outcome.ConclusionsNeuroimaging markers of later outcome still have surprisingly low predictive power and their specificity to individual neurodevelopmental disorders is still under question. However, the field is still young, and substantial challenges to both acquiring and modeling neonatal data are being met.

show abstract

Section: Machine Learning As a Solution?mentioning

confidence: 99%

Annual Research Review: Not just a small adult brain: understanding later neurodevelopment through imaging the neonatal brain

Batalle

Edwards

O’Muircheartaigh

2017

Child Psychology Psychiatry

View full text Add to dashboard Cite

show abstract

“…Besides its role as a potential method for providing supportive correlative measurable values, DTI analysis already functions as a base for the connectome analysis of the brain . Several methodological motor structural connectome studies in the population of preterm born children with a short follow‐up period have already been published . This analysis may offer additional information about the brain plasticity in preterm infants and provide a possible diagnostic tool in the future.…”

Section: Discussionmentioning

confidence: 99%

Diffusion tensor imaging is associated with motor outcomes of very preterm born children at 11 years of age

et al. 2019

View full text Add to dashboard Cite

Aim Very preterm children born <32 weeks of gestation are at risk for motor difficulties such as cerebral palsy and developmental coordination disorder. This study explores the association between diffusion tensor imaging metrics at term and motor outcomes at 11 years of age. Methods A cohort of 37 very preterm infants (mean gestational age 29 4/7, SD 2 0/7) born in 2004‐2006 in Turku University Hospital underwent diffusion tensor imaging at term. A region of interest analysis of fractional anisotropy and mean diffusivity was performed. Motor outcomes at 11 years of age were measured with the Movement Assessment Battery for Children – Second Edition. Results The diffusion metrics of the corpus callosum (genu P = .005, splenium P = .049), the left corona radiata (P = .035) and the right optic radiation (P = .017) were related to later motor performance. Mean diffusivity decreased and fractional anisotropy increased in proportion to the improving performance. Conclusion The diffusion metrics of the genu and splenium of the corpus callosum, the left corona radiata and the right optic radiation at term were associated with motor skills at 11 years of age. Diffusion tensor imaging should be further studied as a potential tool in recognising children at risk for motor impairment.

show abstract

“…[79][80][81][82][83][84][85][86] Although there are many hundreds of descriptive features designed by prior knowledge, the current feature sets still may not be optimal for a given task. [104][105][106][107] DL involves abstraction by building networks with >2 processing layers. [87][88] Only recently applied to radiomics, DL has proven to be valuable in both differential diagnosis [89][90][91][92][93][94][95][96][97][98][99][100][101][102][103] and prognosis.…”

Section: Deep Learningmentioning

confidence: 99%

“…The first CNN was proposed by LeCun et al in 1998, 109 but its success was limited until the advent of graphic processing units and the development of learning algorithms. [104][105][106][107] Cancer December 15, 2018 Two major factors influencing CNN applications in diagnostic imaging are computational power and the availability of training data. With medical imaging, the input layer during training includes images or subregions of labeled images, which then are convolved in sublayers along with their known classifiers (for example, benign or malignant) to identify those image features that are most related to the classification.…”

Section: Deep Learningmentioning

confidence: 99%

Quantitative imaging of cancer in the postgenomic era: Radio(geno)mics, deep learning, and habitats

Napel

Jardim‐Perassi

et al. 2018

Cancer

129

View full text Add to dashboard Cite

Although cancer often is referred to as “a disease of the genes,” it is indisputable that the (epi)genetic properties of individual cancer cells are highly variable, even within the same tumor. Hence, preexisting resistant clones will emerge and proliferate after therapeutic selection that targets sensitive clones. Herein, the authors propose that quantitative image analytics, known as “radiomics,” can be used to quantify and characterize this heterogeneity. Virtually every patient with cancer is imaged radiologically. Radiomics is predicated on the beliefs that these images reflect underlying pathophysiologies, and that they can be converted into mineable data for improved diagnosis, prognosis, prediction, and therapy monitoring. In the last decade, the radiomics of cancer has grown from a few laboratories to a worldwide enterprise. During this growth, radiomics has established a convention, wherein a large set of annotated image features (1‐2000 features) are extracted from segmented regions of interest and used to build classifier models to separate individual patients into their appropriate class (eg, indolent vs aggressive disease). An extension of this conventional radiomics is the application of “deep learning,” wherein convolutional neural networks can be used to detect the most informative regions and features without human intervention. A further extension of radiomics involves automatically segmenting informative subregions (“habitats”) within tumors, which can be linked to underlying tumor pathophysiology. The goal of the radiomics enterprise is to provide informed decision support for the practice of precision oncology.

show abstract

BrainNetCNN: Convolutional neural networks for brain networks; towards predicting neurodevelopment

Cited by 472 publications

References 39 publications

Annual Research Review: Not just a small adult brain: understanding later neurodevelopment through imaging the neonatal brain

Annual Research Review: Not just a small adult brain: understanding later neurodevelopment through imaging the neonatal brain

Diffusion tensor imaging is associated with motor outcomes of very preterm born children at 11 years of age

Quantitative imaging of cancer in the postgenomic era: Radio(geno)mics, deep learning, and habitats

Contact Info

Product

Resources

About