A comprehensive macaque fMRI pipeline and hierarchical atlas

Jung, Benjamin; Taylor, Paul A.; Seidlitz, Jakob; Sponheim, Caleb; Perkins, Pierce; Ungerleider, Leslie G.; Glen, Daniel R.; Messinger, Adam

doi:10.1016/j.neuroimage.2021.117997

Cited by 107 publications

(106 citation statements)

References 67 publications

Supporting

Mentioning

106

Contrasting

Order By: Relevance

“…To compare our deep learning models with state-of-the-art methods for brain extraction, we employed five widely-used skull stripping pipelines implemented in commonly used MRI packages (AFNI, ANTs, FSL, and FreeSurfer) ( Avants et al, 2009 ; Cox, 1996 ; Fischl, 2012 ; Jenkinson et al, 2012 ). Specifically, we tested three intensity-based approaches (FSL BET, FreeSurfer HWA, and AFNI 3dSkullStrip) and two template-driven pipelines (Flirt+ANTS and AFNI @animal_warper) ( Jung et al, 2020 ; Seidlitz et al, 2018 ; Tustison et al, 2020 ). The command and parameters of intensity-based approaches were selected based on the experiments and suggestions from the prior studies as follows ( Xu et al, 2019 ; Zhao et al, 2018 ).…”

Section: Methodsmentioning

confidence: 99%

“…In recent years, registration-based label transferring (i.e. template-driven) approaches have been proposed as a potential solution for NHP brain extraction ( Jung et al, 2020 ; Lohmeier et al, 2019 ; Seidlitz et al, 2018 ; Tasserie et al, 2020 ). These approaches start by registering an individual’s anatomical image to the template in order to establish the deformation between the subject-specific head and template head.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

U-net model for brain extraction: Trained on humans for transfer to non-human primates

Wang

Cho

et al. 2021

NeuroImage

Self Cite

View full text Add to dashboard Cite

Brain extraction (a.k.a. skull stripping) is a fundamental step in the neuroimaging pipeline as it can affect the accuracy of downstream preprocess such as image registration, tissue classification, etc. Most brain extraction tools have been designed for and applied to human data and are often challenged by non-human primates (NHP) data. Amongst recent attempts to improve performance on NHP data, deep learning models appear to outperform the traditional tools. However, given the minimal sample size of most NHP studies and notable variations in data quality, the deep learning models are very rarely applied to multi-site samples in NHP imaging. To overcome this challenge, we used a transfer-learning framework that leverages a large human imaging dataset to pretrain a convolutional neural network (i.e. U-Net Model), and then transferred this to NHP data using a small NHP training sample. The resulting transfer-learning model converged faster and achieved more accurate performance than a similar U-Net Model trained exclusively on NHP samples. We improved the generalizability of the model by upgrading the transfer-learned model using additional training datasets from multiple research sites in the Primate Data-Exchange (PRIME-DE) consortium. Our final model outperformed brain extraction routines from popular MRI packages (AFNI, FSL, and FreeSurfer) across a heterogeneous sample from multiple sites in the PRIME-DE with less computational cost (20 s~10 min). We also demonstrated the transfer-learning process enables the macaque model to be updated for use with scans from chimpanzees, marmosets, and other mammals (e.g. pig). Our model, code, and the skull-stripped mask repository of 136 macaque monkeys are publicly available for unrestricted use by the neuroimaging community at https://github.com/HumanBrainED/NHP-BrainExtraction .

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

U-net model for brain extraction: Trained on humans for transfer to non-human primates

Wang

Cho

et al. 2021

NeuroImage

Self Cite

View full text Add to dashboard Cite

show abstract

“…One possible application of this method is the incorporation of QSM contrast into new macaque MRI brain atlases, such as the National Institute of Mental Health Macaque Template (NMT), which is a high-quality contrast made from T1w images of 31 macaque monkeys (Jung B et al, 2021;Seidlitz et al, 2018). Whereas the NMT shows clear structures in the basal ganglia, the T1w images, particularly in younger monkeys, do not provide sufficient contrast to segment the substantia nigra (SN).…”

Section: Discussionmentioning

confidence: 99%

“…σwm represents the noise measurement calculated as the standard deviation of Iwm. First of all, to create ROIs semiautomatically, each subject's T1w image was aligned to the standard NIMH macaque template (NMT v2.0) using analysis pipeline (NMT_subject_align) with software AFNI (Cox 1996;Jung et al, 2021;Seidlitz et al, 2018) images. Afterward, these masks were applied to the T2w and QSM images for calculating the CNRs after resampled the voxel size of the QSM from 0.4 mm to 0.5 mm as well as the T1w and T2w.…”

Section: Quantitative Image Evaluationmentioning

confidence: 99%

Visualization of subcortical structures in non-human primates in vivo by Quantitative Susceptibility Mapping at 3T MRI

Yoshida¹,

Ye²,

Yu³

et al. 2021

Preprint

View full text Add to dashboard Cite

Magnetic resonance imaging (MRI) is now an essential tool in the field of neuroscience involving non-human primates (NHP). Structural MRI scanning using T1-weighted (T1w) or T2-weighted (T2w) images provides anatomical information, particularly for experiments involving deep structures such as the basal ganglia and cerebellum. However, for certain subcortical structures, T1w and T2w images fail to reveal important anatomical details. To better visualize such structures in the macaque brain, we applied a relatively new method called quantitative susceptibility mapping (QSM), which enhances tissue contrast based on the local tissue magnetic susceptibility. To evaluate the visualization of important structures, we quantified the the contrast-to-noise ratios (CNRs) of the ventral pallidum (VP), globus pallidus external and internal segments (GPe and GPi), substantia nigra (SN), subthalamic nucleus (STN) in the basal ganglia and the dentate nucleus (DN) in the cerebellum. For these structures, the QSM method significantly increased the CNR, and thus the visibility, beyond that in either the T1w or T2w images. In addition, QSM values of some structures were correlated to the age of the macaque subjects. These results indicate that the QSM method can enable the clear identification of certain subcortical structures that are invisible in more traditional scanning sequences.

show abstract

“…Using the AFNI software (Cox, 1996), the segmentation of each brain of each session (anaesthetized and awake sessions) was performed on skull-stripped brains. To ensure optimized inter-session and inter-subject comparisons, both anatomical and functional images were then registered in a common atlas space, CHARM/SARM (Jung et al 2020, Reveley et al 2017, see https://afni.nimh.nih.gov/pub/dist/doc/htmldoc/nonhuman/macaque_tempatl/atlas_charm.html). Temporal filtering was then applied to extract the spontaneous slowly fluctuating brain activity (0.01–0.1Hz).…”

Section: Methodsmentioning

confidence: 99%

Frontal cortical functional connectivity is impacted by anaesthesia in macaques

Giacometti

Dureux

Autran-Clavagnier

et al. 2021

Preprint

View full text Add to dashboard Cite

A critical aspect of neuroscience is to establish whether and how brain networks evolved across primates. To date, most comparative studies have used resting-state functional Magnetic Resonance Imaging (rs-fMRI) in anaesthetized non-human primates and in awake humans. However, anaesthesia strongly affects rs-fMRI signals. The present study investigated the impact of the awareness state (anaesthesia vs. awake) within the same group of macaque monkeys on the rs-fMRI functional connectivity (FC) organization of a well characterized network in the human brain, the cingulo-frontal lateral network. Results in awake macaques revealed a similar FC pattern to that previously uncovered in the human brain. Rostral seeds in the cingulate sulcus exhibited stronger correlation strength with rostral compared to caudal lateral frontal cortical areas while caudal seeds in the cingulate sulcus displayed stronger correlation strength with caudal compared to anterior lateral frontal cortical areas. Critically, this inverse rostro-caudal functional gradient was abolished under anaesthesia. This study demonstrates that the FC pattern of cingulo-frontal cortical networks is preserved from macaque to human but some of its properties can only be observed in the awake state, warranting significant caution when comparing FC patterns across species under different states.

show abstract

A comprehensive macaque fMRI pipeline and hierarchical atlas

Cited by 107 publications

References 67 publications

U-net model for brain extraction: Trained on humans for transfer to non-human primates

U-net model for brain extraction: Trained on humans for transfer to non-human primates

Visualization of subcortical structures in non-human primates in vivo by Quantitative Susceptibility Mapping at 3T MRI

Frontal cortical functional connectivity is impacted by anaesthesia in macaques

Contact Info

Product

Resources

About