News

Below is the news since the year 2018. For previous publications and other achievements, please see my CV.

2/28/2020: A nice software development work done by Zhen Zhou (our visiting student, now research assistant) et al. was recently accepted by Human Brain Mapping.  BrainNetClass (v1.1) is now available at https://github.com/zzstefan/BrainNetClass, together with software manual and example datasets. When installed in Matlab, BrainNetClass allows users to generate brain functional connectivity networks with advanced methods and conduct disease classification in a one-click-run-all fashion. It is highly recommended that neuroscientists and clinicians should use this software to conduct rigorous machine learning-based disease diagnosis.  Paper can be openly accessed here.

  • Zhou, Z., Chen, X., Zhang, Y., Hu, D., Qiao, L., Yu, R., Yap, P.-T., Pan, G.*, Zhang, H.*, Shen, D.*, 2020, A Toolbox for Brain Network Construction and Classification (BrainNetClass). Human Brain Mapping.

1/2/2020: Works of Dr. Zhenyu Tang (our former lab member), Mr. Xi-Ze Jia (my former colleague), and Mr. Bing Cao (visiting PhD student) have been accepted by IEEE Transactions on Medical Imaging (TMI), AAAI 2020, and other journals. In our TMI work, we use brain tumor patients’ multimodal MRI data to predict genotype for their overall survival time prediction, where imaging phenotype was used to both predict genotypic and prognostic data at the same time to let them support each other. AAAI is a top conference in CV&PR field. Mr. Bing Cao also received AAAI’s travel award (for only a few highly acclaimed papers).

  • Tang, Z., Xu, Y., Jin, L., Aibaidula, A., Lu, J., Jiao, Z., Wu, J.*, Zhang, H.*, Shen, D.*, 2020, Deep Learning of Imaging Phenotype and Genotype for Predicting Overall Survival Time of Glioblastoma Patients. IEEE Transactions on Medical Imaging.
  • Jia, X.-Z., Ji, G.-J., Liao, W., Lv, Y.-T., Wang, J., Wang, Z., Zhang, H., Liu, D.-Q., Zang, Y.-F., 2020. Percent amplitude of fluctuation: a simple measure for resting-state fMRI signal at single voxel level. PLoS ONE.
  • Cao, B., Zhang, H.*, Wang, N.*, Gao, X., Shen, D.*, Auto-GAN: Self-Supervised Collaborative Learning for Medical Image Synthesis, AAAI 2020, New York, USA, Feb 7-12, 2020.
  • Jia, X.-Z., Wang, J., Sun, H.-Y., Zhang, H., Liao, W., Wang, Z., Yan, C.-G., Song, X.-W., Zang, Y.-F., 2019. RESTplus: an improved toolkit for resting-state functional magnetic resonance imaging data processing, Science Bulletin, 64:953-954.

12/3/2019: Dr. Guoshi Li, My team member, has his paper “Multiscale Modeling of Intra-Regional and Inter-Regional Connectivities and Their Alterations in Major Depressive Disorder” featured in RSNA 2019! He made a very interesting oral presentation at the conference and then accepted a media interview in Chicago. Later, news in MedicalNewsToday titled “New MRI scans reveal brain features of depression” reported his work to the public. In this paper, we built a novel model that can disentangle the detailed inhibitory/excitatory neural mass model (see the upper figure below) inside each brain region based on common resting-state BOLD fMRI. We, for the first time, demonstrate that by modeling the intra-regional neuronal circuit in addition to the traditional inter-regional large-scale functional connectivity modeling, one can better capture the neural mechanism of depression. The imbalanced DLPFC (dorsal lateral prefrontal cortex) and hyperactivated thalamus could be the cause of depressive symptoms. The work is also included in UNC’s Vital Signs“, a week’s collection of news and events from the School of Medicine. In addition, I also present our recent work (Dr. Pu Huang’s) on deep learning (GAN)-based brain tumor image synthesis, tumor delineation, and tumor grading, all in one model (see the lower figure below)!

  1. Li, G., Liu, Y., Zheng, Y., Wu, Y., Yap, P.-T., Qiu, S.*, Zhang, H.*, Shen, D.*, Multiscale Modeling of Intra-Regional and Inter-Regional Connectivities and Their Alterations in Major Depressive Disorder, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019. (Oral Presentation).
  2. Huang, P., Zhang, H.*, Jiao, Z., Wei, D., Shi, F., Li, D.*, Shen, D.*, Common-space-learning from Multi-modality for Missing MRI Synthesis and Glioma Grading, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019.

 

10/22/2019: Congratulations to Guoshi Li and his newly accepted Human Brain Mapping paper focusing on large-scale Dynamic Causal Modeling of resting-state fMRI for subjects with Major Depressive Disorder (MDD). We found that MDD has reduced excitatory connectivity within the default mode network, and between the default mode and salience networks. MDD also manifests significantly elevated network-averaged inhibitory effective connectivity within the default mode network. This imbalance of excitation and inhibition may underlie disrupted self-recognition and emotional control in MDD. See paper below:

  • Li, G., Liu, Y., Zheng, Y., Li, D., Liang, X., Chen, Y., Cui, Y., Yap, P.-T., Qiu, S.*, Zhang, H.*, Shen, D.*, Large-scale Dynamic Causal Modeling of Major Depressive Disorder based on Resting-state fMRI. Human Brain Mapping, 2019.

8/9/2019: Congratulations to the following authors for their abstracts accepted by RSNA 2019. We have 4 oral presentations and 2 poster presentations! See below:

  1. Jiang, W., Zhang, H.*, Shen, D.*, Development of Graph Theoretical Biomarkers in Early Infancy, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019.
  2. Sun, K., Jiao, Z., Yan, X., Zhang, H., Cheng, J.-Z., Yan, F., Shen, D., Comparison of Four Radiomics-based Classification Methods in Diagnosis of Breast Lesions with Multi-b Diffusion-Weighted MR imaging, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019. (Oral Presentation).
  3. Sun, K., Jiao, Z., Yan, X., Zhang, H., Cheng, J.-Z., Shen, D., Yan, F., Learning Effective Radiomic Features for Characterization of Breast Lesions with Multi-b Diffusion-Weighted MR Imaging, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019. (Oral Presentation).
  4. Li, G., Liu, Y., Zheng, Y., Wu, Y., Yap, P.-T., Qiu, S.*, Zhang, H.*, Shen, D.*, Multiscale Modeling of Intra-Regional and Inter-Regional Connectivities and Their Alterations in Major Depressive Disorder, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019. (Oral Presentation).
  5. Huang, P., Zhang, H.*, Jiao, Z., Wei, D., Shi, F., Li, D.*, Shen, D.*, Common-space-learning from Multi-modality for Missing MRI Synthesis and Glioma Grading, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019.
  6. Yamashita, K., Zhang, H., Li, T., Wen, X., Jing, B., Kam, T.-E., Hsu, L.-M., Yap, P.-T., Wang, L., Li, G., Baluyot, K.R., Howell, B.R., Styner, M.A., Yacoub, E., Chen, G., Potts, T., Gilmore, J.H., Piven, J., Smith, J.K., Ugurbil, K., Zhu, H., Elison, J.T., Hazlett, H., Zhu, H., Shen, D., Lin, W., Symmetrical Functional Connectivity Strength Between Bilateral Anterior Heschl’s Gyri are Negatively Associated with Receptive Function During Infancy, 105th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Dec. 1-6, 2019. (Oral Presentation).

7/8/2019: Congratulations to Tae-Eui Kam and his paper published on TMI! We propose a new deep learning method that is feasible to learn deep discriminative features from resting-state fMRI, essentially, from the resting-state brain networks (voxel-wise, both static and dynamic networks):

  • Kam, T.-E., Zhang, H.*, Jiao, Z., Shen, D., Deep Learning of Static and Dynamic Brain Functional Networks for Early MCI Detection, IEEE Transactions on Medical Imaging, 2019.

7/2/2019: I gave a talk at MICS (Medical Image Computing Seminar (MICS) Webinar at 8:00 AM EST on “Resting-state brain network construction and network-based disease classification”, with more than 200 audiences. The PPT slides can be downloaded from here.

7/1/2019: The following papers have been accepted by the 22nd International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2019), Shenzhen, China, Oct 13-17, 2019. Congratulations to all the authors!

  1. Early Development of Infant Brain Complex Network”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Weixiong Jiang, Han Zhang*, Ye Wu, Liming Hsu, Dan Hu, Dinggang Shen*]
  2. “Multi-layer temporal network analysis reveals increasing temporal reachability and spreadability in the first two years of life”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Zhen Zhou, Han Zhang*, Li-Ming Hsu, Weili Lin, Gang Pan*, Dinggang Shen*, and for the UNC/UMN Baby Connectome Project Consortium]
  3. “Deep Granular Feature-Label Distribution Learning for Neuroimaging-based Infant Age Prediction”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Dan Hu, Han Zhang, Zhengwang Wu, Weili Lin, Gang Li, Dinggang Shen, and for the UNC/UMN Baby Connectome Project Consortium]
  4. “Automated Parcellation of the Cortex Using Structural Connectome Harmonics”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [H. Patrick Taylor IV, Zhengwang Wu, Ye Wu, Dinggang Shen, Han Zhang, Pew-Thian Yap]
  5. “Dynamic Routing Capsule Networks for Mild Cognitive Impairment Diagnosis”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Zhicheng Jiao, Pu Huang, Tae-Eui Kam, Li-Ming Hsu, Ye Wu, Han Zhang*, and Dinggang Shen*]
  6. “CoCa-GAN: Common-feature-learning-based Context-aware Generative Adversarial Network for Glioma Grading”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Pu Huang, Dengwang Li, Zhicheng Jiao, Dongming Wei, Guoshi Li, Han Zhang*, and Dinggang Shen*]
  7. “Identification of Abnormal Circuit Dynamics in Major Depressive Disorder via Multiscale Neural Modeling of Resting-state fMRI”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Guoshi Li, Yujie Liu, Yanting Zheng, Ye Wu, Pew-Thian Yap, Shijun Qiu, Han Zhang*, and Dinggang Shen*]
  8. “Pre-operative Overall Survival Time Prediction for Glioblastoma Patients Using Deep Learning on Both Imaging Phenotype and Genotype”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Zhenyu Tang, Yuyun Xu, Zhicheng Jiao, Junfeng Lu, Lei Jin, Abudumijiti Aibaidula, Jinsong Wu, Qian Wang, Han Zhang*, and Dinggang Shen*]
  9. “A Deep Learning Framework for Noise Component Detection from Resting-state Functional MRI”, MICCAI 2019, Shenzhen, China, Oct 13-17, 2019. [Tae-Eui Kam, Xuyun Wen, Bing Jin, Zhicheng Jiao, Li-Ming Hsu, Zhen Zhou, Yujie Liu, Koji Yamashita, Sheng-Che Hung, Weili Lin, Han Zhang*, and Dinggang Shen*, for UNC/UMN Baby Connectome Project Consortium]

6/25/2019: Our toolbox BrainNetClass v1.0 for Brain Network Construction and Network-based Classification has been released! See the software paper for details. Download link (toolbox, data, manual). Cite:

  • Zhou, Z., Chen, X., Zhang, Y., Qiao, L., Yu, R., Pan, G., Zhang, H.*, Shen, D.*, 2019. Brain network construction and classification toolbox (BrainNetClass). arXiv:1906.09908

6/20/2019: PNAS paper got accepted, which is focusing on early brain structural changes using data from Baby Connectome Project (BCP):

  • Wang, F., Lian, C., Wu, Z., Zhang, H., Li, T., Meng, Y., Wang, L., Lin, W., Shen, D., Li, G., Developmental Topography of Cortical Thickness during Infancy”, PNAS, Accepted.

6/17/2019: Our toolbox for Brain Network Construction and Network-based Classification (BrainNetClass v1.0) to be released! It is a GUI-based, fully automated, easy-to-use connectomics-based individualized classification tool designed for doctors and neuroscientists.

6/11/2019: The collaborative paper on relationship between olfactory function and Alzheimer’s disease has been published. It for the first time shows correlated BOLD fMRI signals between the default mode network and the primary olfactory cortex and the altered synchronization among these two functional networks in AD. It also explained why AD patients tend to have olfactory (smelling) deficit at the early stage. Congratulations!

  • Lu, J., Yang, Q.X., Zhang, H., Eslinger, P.J., Zhang, X., Wu, S., Zhang, B., Zhu, B., Karunanayaka, P.R., Disruptions of the olfactory and default mode networks in Alzheimer’s disease, Brain and Behavior, Accepted.

5/28/2019: Our collaborative paper has been accepted by Journal of Affective Disorders. Congratulations!

  • Zheng, Y., Chen, X., Li, D., Liu, Y., Tan, X., Liang, Y., Zhang, H.*, Qiu, S.*, Shen, D.*, Treatment-Naïve First Episode Depression Classification Based on High-order Brain Functional Network, Journal of Affective Disorders, Accepted.

3/5/2019: Nine of our abstracts were accepted as posters in OHBM 2019 in Rome, Italy! Again, tremendous works done by awesome people here in the IDEA lab!

  1. “Large-scale Dynamic Causal Modeling of Major Depressive Disorder based on Resting-state fMRI”, OHBM, Rome, Italy, June 9-13, 2019. [Guoshi Li, Yujie Liu, Yanting Zheng, Li-Ming Hsu, Han Zhang*, Dinggang Shen]
  2. “Frequency specificity of spontaneous brain activity in developing infant brain”, OHBM, Rome, Italy, June 9-13, 2019. [Li-Ming Hsu, Han Zhang*, Xuyun Wen, Bin Jing, Tae-Eui Kam, Li Wang, Zhengwang Wu, Pew-Thian Yap, Kristine R. Baluyot, Howell, B.R., Styner, M.A., Yacoub, E., Chen, G., Potts, T., Gilmore, J.H., Piven, J., Smith, J.K., Ugurbil, K., Zhu, H., Heather Hazlett, Elison, J.T. , Weili Lin, Dinggang Shen, for UNC/UMN Baby Connectome Project Consortium]
  3. Diagnosing Major Depressive Disorder with High-order Local Functional Connectivity”, OHBM, Rome, Italy, June 9-13, 2019. [Yujie Liu, Li-Ming Hsu, Yanting Zheng, Shijun Qiu, Han Zhang*, Dinggang Shen]
  4. “Dynamic Neural Disruptions Associated with Antisocial Behavior”, OHBM, Rome, Italy, June 9-13, 2019. [Weixiong Jiang, Han Zhang*, Lingli Zeng, Hui Shen, Wei Wang, Dewen Hu, Dinggang Shen]
  5. Infant Age Prediction Based on Deep Granular Label Distribution Learning of Cortical Features”, OHBM, Rome, Italy, June 9-13, 2019. [Dan Hu, Zhengwang Wu, Han Zhang, Weili Lin, Gang Li , Dinggang Shen, for UNC/UMN Baby Connectome Project Consortium]
  6. “Infant Resting-state FMRI Analysis Pipeline for UNC/UMN Baby Connectome Project”, OHBM, Rome, Italy, June 9-13, 2019. [Han Zhang, Xuyun Wen, Bin Jing, Li-Ming Hsu, Tae-Eui Kam, Zhengwang Wu, Li Wang, Gang Li, Weili Lin, Dinggang Shen, for UNC/UMN Baby Connectome Project Consortium]
  7. “Month-to-month Development of Brain Functional Networks during Early Infancy”, OHBM, Rome, Italy, June 9-13, 2019. [Han Zhang, Gang Li, Xuyun Wen, Bin Jing, Li-Ming Hsu, Tae-Eui Kam, Weili Lin, Dinggang Shen, for UNC/UMN Baby Connectome Project Consortium]
  8. “Deep Learning-based Automatic Noisy Component Detection for Automatic Resting-state fMRI Denoising”, OHBM, Rome, Italy, June 9-13, 2019. [Tae-Eui Kam, Han Zhang*, Bin Jing, Xuyun Wen, Weili Lin, Dinggang Shen, for UNC/UMN Baby Connectome Project Consortium]
  9. “Prediction of 7-year progression from subjective cognitive decline to MCI”, OHBM, Rome, Italy, June 9-13, 2019. [Ling Yue, Dan Hu, Han Zhang, Junhao Wen, Ye Wu, Tao Wang, Dinggang Shen, Shifu Xiao]

2/27/2019: The collaborative paper got accepted by Frontiers in Human Neuroscience, which is about identifying co-developing patterns from structural covariance network.

  • Xu, X., He, P., Yap, P.-T., Zhang, H., Nie, J., Shen, D., Meta-network Analysis of Structural Correlation Networks Provides Insights into Brain Network Development, Frontiers in Human Neuroscience, Accepted.

2/25/2019: Our two ISMRM abstracts got accepted. They are:

  • Wen, X., Zhang, H., Wang, R., Lin, W., Shen, D., Increased Functional Connectivity Flexibility During Early Infancy, 27th ISMRM, Montreal, QC, Canada, May 11-16, 2019. (Oral presentation).
  • Hsu, L.-M., Liu, Y., Zhang, H., Qiu, S., Shen, D., Diagnosing first-episode depression with high-order regional homogeneity, 27th ISMRM, Montreal, QC, Canada, May 11-16, 2019.

1/22/2019: The collaborative work with radiologists in Guangzhou University of Chinese Medicine on white matter abnormality in Type 2 Diabetes Mellitus (T2DM) has been accepted by Frontiers in Neuroscience. Congratulations! In the paper, we investigate the feasibility of using a new technique, local homogeneity of diffusion weights, to investigate white matter abnormalities in diabetes, which reveals more potential abnormalities than the traditional metrics like FA. It could help better understanding why diabetes patients are likely to develop cognitive impairment.

12/14/2018: Our network-wise high-order functional connectivity paper for early Alzheimer’s disease detection has been accepted by Neuroinformatics. Congratulations!

  • Zhang, H., Giannakopoulos. P., Haller, S., Shen, D., Inter-Network High-Order Functional Connectivity (IN-HOFC) and its Alteration in Patients with Mild Cognitive Impairment, Neuroinformatics, Accepted.   

12/6/2018: Another collaborative work by the leading author Yu Zhang, now in Stanford, has been accepted by Pattern Recognition. A new brain network construction method was proposed to better model the brain functional connectivity networks for individual disease diagnosis. Previous methods based on sparse learning ignore the differences between patients and healthy controls, which leads to less sensitive disease detection. Our method can not only generate group consistent brain networks with sparse connections, but also lever the group difference information to make the constructed networks different between two groups. This method can detect early Alzheimer’s disease with fundamentally increased accuracy. We hope this method can be adopted by other researchers for better individual diagnosis.

  • Zhang, Y., Zhang, H., Chen, X., Liu, M., Zhu, X., Lee, S.-W., Shen, D., Strength and Similarity Guided Group-level Brain Functional Network Construction for MCI Diagnosis. Pattern Recognition, Accepted.

12/3/2018: We are back from RSNA 2018, Chicago! Dr. Han Zhang (Assistant Professor, Department of Radiology and BRIC) and his student Ms. Xuyun Wen (visiting student from Sun Yat Sen University) attended RSNA 2018, during which they gave two oral presentations about their works on early infant brain development. They are both from the IDEA Research Laboratory (director: Prof. Dinggang Shen).

Dr. Zhang’s research focuses on how large-scale brain functional networks change in the first five years of age. His group found that, when embedded in a low dimensional space, the whole brain “functional connectome” could clearly trace the development and consciousness level change with two “gradients”. This will help better understanding of the neural substrate of early brain development and could facilitate the detection of underdevelopmental disorders.

Photo shows Dr. Zhang is presenting his work on a scientific paper session at RSNA 2018 (photo courtesy of Feng Shi)

Ms. Wen’s work, on the other hand, focuses on more subtle brain network changes in a shorter temporal scale. With a proposed data-driven method, she and the colleagues revealed that the brain dynamically changes its functional network patterns even in second. With development, more high-order cognitive function-related networks emerge, and the brain spends longer time on these networks, possibly answering the call to conducting more complex tasks, such as adaptive control. This is the first dynamic brain functional connectivity work on early infancy development.

Photo shows Ms. Wen’s work and she enjoy a break after presentation in Chicago downtown (left: Xuyun Wen, middle: Han Zhang, right: Yujie Liu; all from IDEA lab. Photo courtesy of Han Zhang)

In this year’s RSNA, IDEA lab has 7 abstracts accepted, 3 of which are accepted as oral presentations. Their works have been covering medical image analysis, early brain development, computer-aided diagnosis, and so on. This year’s RSNA puts more attention at artificial intelligence such as deep learning, and how radiologists can benefit from it. IDEA lab is a leading medical image analysis team worldwide, having been developing novel and practical AI algorithms to help radiologists working more efficiently and fruitfully in their daily routine, including diagnosis, prognosis, treatment planning, early disease detection and screening. See more exciting works on the website of IDEA lab.

Photo shows Dr. Zhang was presenting one of the IDEA lab’s work on infant diffusion weighted image analysis to professors in the Martinos Center at the Harvard Medical School (photo courtesy of Xuyun Wen)

11/13/2018: Finally! Although significantly delayed, the collaborative paper on deep-learning-based brain tumor overall survival time prediction got accepted. In the paper, we show that deep learning model can “see” in advance the malignancy of the brain tumors. This is one of the fruits from 10+ years collaborations with the terrific neurosurgeons in Huashan Hospital, Shanghai. AI empowered by deep convolutional neural network is used for individualized  prognosis for patients with high-grade gliomas with > 90% accuracy, great for surgical planning and early treatment tailoring.

  • Nie, D.#, Lu, J. #, Zhang, H.#, Adeli, E., Wang, J., Yu, Z., Liu, L., Wang, Q., Wu, J., Shen, D., Multi-Channel 3D Deep Feature Learning for Survival Time Prediction of Brain Tumor Patients Using Multi-Modal Neuroimages, Scientific Report, Accepted.

10/25/2018: For those who are interested in how does human brain functional network’s modular structure develop from birth to one year old at every three months, our paper entitled “First-Year Development of Modules and Hubs in Infant Brain Functional Networks” can be downloaded here for free! It’s the first time that the whole brain network’s topography is investigated using longitudinal natural sleeping state fMRI.

10/5/2018: Our paper on early development study of functional modules and hubs in neonates and infants based on a longitudinal fMRI data has been recently accepted by NeuroImage. Nice work done by our visiting student, Xuyun. In the work, we found that the number of communities in brain functional network keep increasing at every 3 months, with splitting occurs in order of temporal, frontal, parietal, and back to frontal lobe between 0 and 12 months old. We further revealed that intra-modular connectivity is increasing rapidly in the first 3 months of age, then slowing down for the rest of months, while the inter-modular connections are keep (linearly) increasing. We step further and visit the functional hubs at each age and found that provincial hubs are decreasing and connector hubs are increasing from null. All of the findings are based on our much improved modular detection algorithm that makes advantage of each individual’s modular structure to “augment” the modular structure at the group level while preserving the temporal smoothness along development, termed “Module-Guided Group-Level Network Construction”.

8/14/2018: Our paper on malignant brain tumor classification has been accepted by Joint IAPR International Workshops on Statistical Techniques in Pattern Recognition (SPR) and Structural and Syntactic Pattern Recognition (SSPR), Beijing, China, Aug. 17-19, which has been published in X. Bai et al. (Eds.): S+SSPR 2018, LNCS 11004, pp. 14-21, 2018. See below:

8/8/2018: A collaborate paper on individualized presurgical prediction of overall survival time for high-grade glioma patients based on low-order and high-order whole-brain functional network has been accepted by Brain Imaging and Behavior. Congratulations to my colleague Luyan Liu, Qian Wang, Xiaobo Chen and others, as well as my 10-year+ collaborator, the neurosurgeons in Huashan Hospital, Shanghai. Especially thanks to Dr. Junfeng Lu and Dr. Jinsong Wu for the generous data contribution and very helpful suggestions during the study.  Predicting survival time preoperatively is very important to surgical planning for individualized optimized treatment. Previously, doctors can only based on limited clinical and demographic information, such as patient’s age, daily living performance, tumor size and shape, to roughly predict the survival time for this patient.  Some recent machine learning papers are using tumor MR images to predict survival time, but only focusing on the features extracted from very localized tumor regions. We propose that, a brain tumor may also affect whole-brain connectomics in a systemic way, and that different tumors affect it differently, leading to varied survival time even in the patients with same grade, similar age and tumor size. We propose a machine learning framework for long or short survival time prediction by using both low- and high-order functional connectivity networks and achieved an prognosis accuracy of ~87%! See details on:

7/30/2018: Our review paper on early brain functional development with resting-state fMRI has been online published on NeuroImage. Read for understanding this field and consistent early development findings, as well as the future trends.

7/24/2018: Four of our abstracts has been accepted by RSNA 2018 Annual Meeting, two of which are oral presentations! The full list are as below:

  1. Zhang, H., Yin, W., Shen, D., Lin, W., Development- and state-related gradients in infant brain functional connectome, 104th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Nov. 25-30, 2018. (Oral Presentation).
  2. Jing, B., Zhang, H., Lin, W., Shen, D., Age Prediction Using Resting-State Functional Connectivity Characteristics in Typically Developing Infants, 104th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Nov. 25-30, 2018.
  3. Wen, X., Zhang, H., Lin, W., Shen, D., Evolution of Brain Dynamics in the First 2 Years of Life, 104th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Nov. 25-30, 2018. (Oral Presentation).
  4. Zheng, Y., Chen, X., Li, D., Liu, Y., Liang, Y., Qin, C., Zeng, H., Chen, J., Liu, J., Zhang, H., Qiu, S., Shen, D., Treatment-Naïve First Episode Depression Diagnosis Based on Brain Functional Network, 104th RSNA Scientific Assembly and Annual Meeting, Chicago, USA, Nov. 25-30, 2018.

7/21/2018: Update: The collaborative paper using deep learning to segment  6-month-old infant MRI and for early diagnosis of Autism has been accepted by MICCAI 2018. Another collaborative paper using deep learning to segment 6- and 12-month-old infant cerebellum has been accepted by Machine Learning in Medical Imaging 2018, a MICCAI workshop. See below.

  • Wang, L., Li, G., Shi, F., Cao, X., Lian, C., Nie, D., Liu, M., Zhang, H., Li, G., Lin, W., Shen, D., Volume-based Analysis of 6-month-old Infant Brain MRI for Autism Biomarker Identification and Early Diagnosis. MICCAI 2018, Granada, Spain, Sep. 16-20, 2018.
  • Chen, J., Zhang, H., Nie, D., Wang, L., Li, G., Lin, W., Shen, D., Automatic Accurate Infant Cerebellar Tissue Segmentation with Densely Connected Convolutional Network. MLMI 2018, Accepted.

7/19/2018: UNC Department of Radiology 2018 Summer Newsletter highlight the significant contributions from our group (IDEA lab) in “Spotlight on Basic Science”, mentioning our prominent role in the “Baby Connectome Project” as well as the “Brain Research through Advancing Innovative Neurotechnologies (BRAIN)” grant. I am really proud to be the one of the faculties who has been contributing to the investigation of how “Alzheimer’s disease is associated with changes in higher-order dynamics of brain function”. The news noted that “In addition to Dr. Shen, faculty members of the IDEA team include Drs. Pew-Thian Yap, Gang Li, Li Wang, and Han Zhang. The team has been consistently and highly funded by the National Institutes of Health (NIH). Shen was ranked 4th nationwide among radiology faculty in terms of NIH funding in 2016.”

7/16/2018: It is more than a month since a talented high school student, Nikhil, has been studying in our lab. Nikhil is working on deep learning-based depression classification very hard, hoping to raise the performance of individual diagnosis for the depression disorders. As his co-supervisor (also Prof. Shen), I am really glad to see his progress and happy to be working with him.

7/9/2018: The collaborate paper entitled “Spatiotemporal Analysis of Developing Brain Networks” by [Ping He, Xiaohua Xu, Han Zhang, Gang Li, Jingxin Nie, Pew-Thian Yap, Dinggang Shen] has been accepted by Frontiers in Neuroinformatics.  For the first time, the early development of spatially and longitudinally covaried cortical thickness sub-networks have been identified based on infant developmental MRI. Congratulations!

7/2/2018: Our review paper on resting-state fMRI-based early development studies is accepted by NeuroImage! The title is “Resting-state Functional MRI Studies on Infant Brains: a Decade of Gap-Filling Efforts”, the authors are Han Zhang, Dinggang Shen* and Weili Lin* (co-corresponding authors). This is an informative review which highlight the recent advance in brain function studies at neonatal and early infancy stages (0-5 years old). In the paper, we summarized consistent results in brain functional connectivity and brain network analysis in early development, provided a data analysis pipeline for “baby connectome” analysis, and gave future directions in the field. The paper will be online soon!

6/27/2018: Had an invitational talk at Center for Cognitive and Brain Disorders at Hangzhou Normal University, my previous place. I gave a presentation on high-order functional connectivity for early AD detection. Another invitational talk was given to the employees at United Imaging Intelligence, which is about how could we neuroimage computing guys could collaborate with the neurosurgeons. After an oral presentation and three-day poster presentation, I felt that people in OHBM 2018 are very interested in our method for multi-layer functional network analysis.

5/31/2018: Behold our new slides that will be presented in the coming OHBM 2018 conference in Singapore!

  

5/25/2018: The “Medical Image Computing and Computer Assisted Intervention Society (MICCAI)” just released the accepted paper for this year’s conference (MICCAI 2018, Granada, Spain, Sep. 16-20, 2018). We have achieved great ones! My paper “Multi-layer large-scale functional connectome reveals infant brain developmental patterns”, authored by Han Zhang, Natalie Stanley, Peter J. Mucha, Weiyan Yin, Weili Lin, and Dinggang Shen, as well as two my collaborators’ papers: Weizheng Yan et al., “Deep Chronnectome Learning via Full Bidirectional Long Short-Term Memory Networks for MCI Diagnosis”, and Tae-Eui Kam et al., “A Novel Deep Learning Framework on Brain Functional Networks for Early MCI Diagnosis”, have all been accepted!  Congratulations for the excellent works to all authors!  In my paper, I proposed a multi-layer network analysis method for module detection and modular development analysis based on a longitudinal infant 0-1-2 years old resting-state fMRI data. For the first time, we are able to detect subject-consistent modules and, more importantly, subject-inconsistent modules that indicate individual variability.  In Weizheng’s paper, we propose a memory-based deep learning algorithm for analyzing spatiotemporal information in the dynamic brain functional networks. This is the first deep learning paper for dynamic FC-based disease diagnosis.  In Tae-Eui’s paper, we propose the first ICA component-based 3D-CNN framework for early MCI detection. Instead of traditional voxel-wise mass-univariate analysis, deep learning is for the first time trained to extract high-level intra- and inter-network FC information for multi-brain-network-based joint disease classification.  Please send email for questions and inquiries. 

5/11/2018: The collaborating paper with Jun Wang @ Jiangnan Univ (our lab’s previous member, first author) has been accepted by IEEE Transactions on Cybernetics. Congratulations! The paper “Sparse Multi-View Task-Centralized Ensemble Learning for ASD Diagnosis Based on Age- and Sex-related Functional Connectivity Patterns” uses traditional FC and our proposed High-order FC (HOFC) in a multi-task multi-view framework to diagnose autism based on resting-state brain FC networks.

5/9/2018: I became Senior member of IEEE today!

5/8/2018: The Flux Satellite Conference 2018 was successful. The workshop on “Infant brain processing” has received many positive feedback. We announced iBeat v2.0 for structure MRI and surface-based analysis (Li Wang and Gang Li) and the first version of resting-state fMRI minimal/extensive preprocessing pipeline (Han Zhang). Thanks for our team (Xuyun Wen and Bin Jing).  Thanks for the organizers and all lecturers.

5/6/2018: IDEA Research lab and BRIC held FLUX Satellite Conference 2018 and we faculties working on Baby Connectome Project (BCP) will all give a lecture on multimodal imaging analysis (structure MRI — by Gang and Li, diffusion MRI — by Yap, resting-state fMRI — by me) in the workshop “Infant Brain Processing“.  My talk will showcase an overview of BCP resting-state fMRI data, data preprocessing and post-processing with examples. The audience are from everywhere in the country including students, postdocs and professors who are interested in the brain imaging-based developmental study in the early life.

4/17/18: Another high-order functional connectivity paper “Diagnosis of Autism Spectrum Disorders Using Multi-level High-order Functional Networks Derived from Resting-State Functional MRI” for Autism Spectrum Disorder (ASD) diagnosis has been accepted by Frontiers in Human Neuroscience. The first author is Feng Zhao. It shows that, by combining low-order (BOLD-signal-synchronization-based) FC and high-order (FC-profile-similarity-based) FC, accuracy of ASD diagnosis can be lifted from 73% to 81%.

3/17/18: Luyan Liu’s paper “Exploring Diagnosis and Imaging Biomarkers of Parkinson’s Disease via Iterative Canonical Correlation Analysis Based Feature Selection” where I was coauthored has been accepted by Computerized Medical Imaging and Graphics. This paper is about Parkinson’s disease diagnosis based on iterative CCA-based feature selection.

3/8/18: The abstract on multi-layer functional connectome analysis with preserved individual variability has been accepted by OHBM 2018 as oral presentation. The title is “Multi-Layer Functional Connectome Reveals New Developmental Patterns of the Infant Brain”. The meeting will be held on 17-21 June 2018 in Singapore. I will present our work there!

2/25/18: The collaborative paper with Huashan Hospital and Nanjing University on jointly predicting genomic biomarkers for brain tumor patients has been accepted by IEEE Transactions on Medical Imaging. The title is “Multi-label Nonlinear Matrix Completion with Transductive Multi-task Feature Selection for Joint MGMT and IDH1 Status Prediction of Patient with High-Grade Gliomas”.

2/17/18: The collaborative paper with Southern Medical University on white matter functional connectivity-guided fMRI registration has been accepted by Human Brain Mapping. The title is “Functional MRI Registration with Tissue-Specific Patch-Based Functional Correlation Tensors”.

2/2/18: The abstract on robust multi-subject brain network analysis for developmental analysis, and another abstract on infant brain molecularity analysis and hub identification have been both accepted by ISMRM 2018 as oral presentation. The meeting will be held on 16-21 June 2018 at Paris, France (Joint Annual Meeting ISMRM-ESMRMB).

1/19/18: The collaborative paper with Huashan Hospital, No. 1 Affiliated Hospital of Fujian Medical University, Affiliated Hospital of Hangzhou Normal University, and The third Hospital of Hebei Medical University on using ICA and resting-state fMRI for glioma detection has been accepted by Scientific Reports. The title is “Tumor Tissue Detection using Blood-Oxygen-Level-Dependent Functional MRI based on Independent Component Analysis”.