A Deep Learning Framework for Noise Component Detection from Resting-State Functional MRI

Kam, Tae-Eui; Wen, Xuyun; Jin, Bing Jie; Jiao, Zhicheng; Hsu, Li-Ming; Zhou, Zhen; Liu, Yujie; Yamashita, Koji; Hung, Sheng-Che; Lin, Weili; Zhang, Han; Shen, Dinggang; Unc, for

doi:10.1007/978-3-030-32248-9_84

Cited by 16 publications

(23 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MRI data were pre-processed using an in-house infant-specific pipeline ( Wu et al, 2012 ; Wang et al, 2015 ; Jiang et al, 2019b ) which shares some common steps with the HCP pipeline ( https://github.com/Washington-University/Pipelines ), including head motion correction, alignment of rsfMRI images to T1 space, and band-pass filtering (0.01 Hz-0.08 Hz), but adds several unique steps tailored to infant functional connectivity MRI ( Kam et al, 2019 ). Brain tissue segmentation was first conducted to generate tissue labeling maps (gray matter, white matter, or cerebrospinal fluid) using a multi-site infant-dedicated computational toolbox, iBEAT v2.0 Cloud ( http://www.ibeat.cloud ) ( Wang et al, 2018 ).…”

Section: Methodsmentioning

confidence: 99%

“…The tissue labeling maps were used to register to the template (Colin 27 atlas) ( Holmes et al, 1998 ) in MNI space using advanced normalization tools (ANTs) ( Avants et al, 2011 ). Automatic noise-related component detection and regression were performed ( Kam et al, 2019 ). Specifically, a deep learning-based rsfMRI QC method, the long short-term memory (LSTM) neural network ( Yan et al, 2018 , 2019 ), was employed, capable of effectively extracting FC quality-related features from the raw data.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Effects of prenatal opioid exposure on functional networks in infancy

Merhar

Jiang

Parikh

et al. 2021

Developmental Cognitive Neuroscience

Self Cite

View full text Add to dashboard Cite

Prenatal opioid exposure has been linked to altered neurodevelopment and visual problems such as strabismus and nystagmus. The neural substrate underlying these alterations is unclear. Resting-state functional connectivity MRI (rsfMRI) is an advanced and well-established technique to evaluate brain networks. Few studies have examined the effects of prenatal opioid exposure on resting-state network connectivity in infancy. In this pilot study, we characterized network connectivity in opioid-exposed infants (n = 19) and controls (n = 20) between 4–8 weeks of age using both a whole-brain connectomic approach and a seed-based approach. Prenatal opioid exposure was associated with differences in distribution of betweenness centrality and connection length, with positive connections unique to each group significantly longer than common connections. The unique connections in the opioid-exposed group were more often inter-network connections while unique connections in controls and connections common to both groups were more often intra-network. The opioid-exposed group had smaller network volumes particularly in the primary visual network, but similar network strength as controls. Network topologies as determined by dice similarity index were different between groups, particularly in visual and executive control networks. These results may provide insight into the neural basis for the developmental and visual problems associated with prenatal opioid exposure.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Effects of prenatal opioid exposure on functional networks in infancy

Merhar

Jiang

Parikh

et al. 2021

Developmental Cognitive Neuroscience

Self Cite

View full text Add to dashboard Cite

show abstract

“…After conservative high-pass filtering with a sigma of 1,000 s to remove linear trends in the data, individual independent component analysis was conducted to decompose each of the preprocessed rs-fMRI data into 150 components using MELODIC in FSL. An automatic deep learning-based noise-related component identification algorithm was used to identify and remove non-signal components to clean the rs-fMRI data ( 48 ).…”

Section: Methodsmentioning

confidence: 99%

Generation of Infant-dedicated Fine-grained Functional Parcellation Maps of Cerebral Cortex

Wang

Zhang

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Infancy is a dynamic and immensely important period in human brain development. Studies of infant functional development using resting-state fMRI rely on precisely defined cortical parcellation maps. However, available adult-based functional parcellation maps are not applicable for infants due to their substantial differences in functional organizations. Fine-grained infant-dedicated cortical parcellation maps are highly desired but remain scarce, due to difficulties ranging from acquiring to processing of infant brain MRIs. In this study, leveraging 1,064 high-resolution longitudinal rs-fMRIs from 197 infants from birth to 24 months and advanced infant-dedicated processing tools, we create the first set of infant-specific, fine-grained cortical functional parcellation maps. Besides the conventional folding-based cortical registration, we specifically establish the functional correspondences across individuals using functional gradient densities and generate both age-specific and age-common fine-grained parcellation maps. The first set of comprehensive brain functional developmental maps are accordingly derived, and reveals a complex, hitherto unseen multi-peak fluctuation development pattern in temporal variations of gradient density, network sizes, and local efficiency, with more dynamic changes during the first 9 months than other ages. Our proposed method is applicable in generating fine-grained parcellations for the whole lifespan, and our parcellation maps will be available online to advance the neuroimaging field.

show abstract

“…However, the denoised neuroimaging datasets are generally not available for the supervised learning based denoising models because of the intimidating cost of the human labeling. The fixed featurebased denoising methods may help here when the cleaned images are not available [9], [10], but the tuning of the model for a slightly modified system are prone to be difficult and the result may diverge from what the neuroscientists expect [11]. Moreover, the optical imaging system is usually under a fast continuous development.…”

Section: Introductionmentioning

confidence: 99%

Denoising Across Data Acquisition Modalities for Mesoscopic Scale Optical Neuroimaging

Zhu

Han

et al. 2021

IEEE Access

View full text Add to dashboard Cite

The ever-evolving mesoscopic scale optical imaging systems facilitate the new discoveries in the neuroscience research. They excel at providing a complete view of the long projections of a single neuron across the whole brain. Although the preparation protocols and the optical imaging systems are rapidly evolving, there are still some noise and artifacts interfering downstream data processing and analysis due to the defects in the imaging system or during the sample preparation. A specific denoising procedure is usually developed for one optical imaging system as a post-processing measure. However, the development of the optical imaging systems usually follows in an incremental manner. It would be better to adapt the denoising model to the new optical imaging system than training a denoising model from scratch. In this paper, the proposed learning schema and practice learn a new denoising model for a new optical imaging system based on the existing denoise models and bootstrap a denoising model without the ground-truth denoising labels. We achieve this through a CycleGAN based model and the fact that the optical imaging systems usually produce images both from the signal area and the fixture area. The experiments show that our proposed procedure can provide a comparable denoising performance against other state-of-the-art denoising methods for several optical neuroimaging datasets.

show abstract

A Deep Learning Framework for Noise Component Detection from Resting-State Functional MRI

Cited by 16 publications

References 13 publications

Effects of prenatal opioid exposure on functional networks in infancy

Effects of prenatal opioid exposure on functional networks in infancy

Generation of Infant-dedicated Fine-grained Functional Parcellation Maps of Cerebral Cortex

Denoising Across Data Acquisition Modalities for Mesoscopic Scale Optical Neuroimaging

Contact Info

Product

Resources

About