Neurodegenerative diseases cause great medical and economic burdens for both patients and society;however, the complex molecular mechanisms thereof are not yet well understood. With the development of high-coverage se...Neurodegenerative diseases cause great medical and economic burdens for both patients and society;however, the complex molecular mechanisms thereof are not yet well understood. With the development of high-coverage sequencing technology, researchers have started to notice that genomic repeat regions, previously neglected in search of disease culprits, are active contributors to multiple neurodegenerative diseases. In this review, we describe the association between repeat element variants and multiple degenerative diseases through genome-wide association studies and targeted sequencing. We discuss the identification of disease-relevant repeat element variants, further powered by the advancement of long-read sequencing technologies and their related tools, and summarize recent findings in the molecular mechanisms of repeat element variants in brain degeneration, such as those causing transcriptional silencing or RNA-mediated gain of toxic function. Furthermore, we describe how in silico predictions using innovative computational models, such as deep learning language models, could enhance and accelerate our understanding of the functional impact of repeat element variants. Finally, we discuss future directions to advance current findings for a better understanding of neurodegenerative diseases and the clinical applications of genomic repeat elements.展开更多
Distinct brain remodeling has been found after different nerve reconstruction strategies,including motor representation of the affected limb.However,differences among reconstruction strategies at the brain network lev...Distinct brain remodeling has been found after different nerve reconstruction strategies,including motor representation of the affected limb.However,differences among reconstruction strategies at the brain network level have not been elucidated.This study aimed to explore intranetwork changes related to altered peripheral neural pathways after different nerve reconstruction surgeries,including nerve repair,endto-end nerve transfer,and end-to-side nerve transfer.Sprague–Dawley rats underwent complete left brachial plexus transection and were divided into four equal groups of eight:no nerve repair,grafted nerve repair,phrenic nerve end-to-end transfer,and end-to-side transfer with a graft sutured to the anterior upper trunk.Resting-state brain functional magnetic resonance imaging was obtained 7 months after surgery.The independent component analysis algorithm was utilized to identify group-level network components of interest and extract resting-state functional connectivity values of each voxel within the component.Alterations in intra-network resting-state functional connectivity were compared among the groups.Target muscle reinnervation was assessed by behavioral observation(elbow flexion)and electromyography.The results showed that alterations in the sensorimotor and interoception networks were mostly related to changes in the peripheral neural pathway.Nerve repair was related to enhanced connectivity within the sensorimotor network,while end-to-side nerve transfer might be more beneficial for restoring control over the affected limb by the original motor representation.The thalamic-cortical pathway was enhanced within the interoception network after nerve repair and end-to-end nerve transfer.Brain areas related to cognition and emotion were enhanced after end-to-side nerve transfer.Our study revealed important brain networks related to different nerve reconstructions.These networks may be potential targets for enhancing motor recovery.展开更多
Binaural rendering is of great interest to virtual reality and immersive media. Although humans can naturally use their two ears to perceive the spatial information contained in sounds, it is a challenging task for ma...Binaural rendering is of great interest to virtual reality and immersive media. Although humans can naturally use their two ears to perceive the spatial information contained in sounds, it is a challenging task for machines to achieve binaural rendering since the description of a sound field often requires multiple channels and even the metadata of the sound sources. In addition, the perceived sound varies from person to person even in the same sound field. Previous methods generally rely on individual-dependent head-related transferred function(HRTF)datasets and optimization algorithms that act on HRTFs. In practical applications, there are two major drawbacks to existing methods. The first is a high personalization cost, as traditional methods achieve personalized needs by measuring HRTFs. The second is insufficient accuracy because the optimization goal of traditional methods is to retain another part of information that is more important in perception at the cost of discarding a part of the information. Therefore, it is desirable to develop novel techniques to achieve personalization and accuracy at a low cost. To this end, we focus on the binaural rendering of ambisonic and propose 1) channel-shared encoder and channel-compared attention integrated into neural networks and 2) a loss function quantifying interaural level differences to deal with spatial information. To verify the proposed method, we collect and release the first paired ambisonic-binaural dataset and introduce three metrics to evaluate the content information and spatial information accuracy of the end-to-end methods. Extensive experimental results on the collected dataset demonstrate the superior performance of the proposed method and the shortcomings of previous methods.展开更多
Background Structural imaging holds great potential for precise targeting and stimulation for deep brain stimulation(DBS).The anatomical information it provides may serve as potential biomarkers for predicting the eff...Background Structural imaging holds great potential for precise targeting and stimulation for deep brain stimulation(DBS).The anatomical information it provides may serve as potential biomarkers for predicting the efficacy of DBS in treatment-resistant depression(TRD).Aims The primary aim is to identify preoperative imaging biomarkers that correlate with the efficacy of DBS in patients with TRD.Methods Preoperative imaging parameters were estimated and correlated with the 6-month clinical outcome of patients with TRD receiving combined bed nucleus of the stria terminalis(BNST)-nucleus accumbens(NAc)DBS.White matter(WM)properties were extracted and compared between the response/non-response and remission/non-remission groups.Structural connectome was constructed and analysed using graph theory.Distances of the volume of activated tissue(VAT)to the main modulating tracts were also estimated to evaluate the correlations.Results Differences in fibre bundle properties of tracts,including superior thalamic radiation and reticulospinal tract,were observed between the remission and nonremission groups.Distance of the centre of the VAT to tracts connecting the ventral tegmental area and the anterior limb of internal capsule on the left side varied between the remission and non-remission groups(p=0.010,t=3.07).The normalised clustering coefficient(γ)and the small-world property(σ)in graph analysis correlated with the symptom improvement after the correction of age.Conclusions Presurgical structural alterations in WM tracts connecting the frontal area with subcortical regions,as well as the distance of the VAT to the modulating tracts,may influence the clinical outcome of BNST-NAc DBS.These findings provide potential imaging biomarkers for the DBS treatment for patients with TRD.展开更多
Spectral computed tomography(CT)based on photon counting detectors can resolve the energy of every single photon interacting with the sensor layer and be used to analyze material attenuation information under differen...Spectral computed tomography(CT)based on photon counting detectors can resolve the energy of every single photon interacting with the sensor layer and be used to analyze material attenuation information under different energy ranges,which can be helpful for material decomposition studies.However,there is a considerable amount of inherent quantum noise in narrow energy bins,resulting in a low signal-to-noise ratio,which can consequently affect the material decomposition performance in the image domain.Deep learning technology is currently widely used in medical image segmentation,denoising,and recognition.In order to improve the results of material decomposition,we propose an attention-based global convolutional generative adversarial network(AGC-GAN)to decompose different materials for spectral CT.Specifically,our network is a global convolutional neural network based on an attention mechanism that is combined with a generative adversarial network.The global convolutional network based on the attention mechanism is used as the generator,and a patchGAN discriminant network is used as the discriminator.Meanwhile,a clinical spectral CT image dataset is used to verify the feasibility of our proposed approach.Extensive experimental results demonstrate that AGC-GAN achieves a better material decomposition performance than vanilla U-Net,fully convolutional network,and fully convolutional denseNet.Remarkably,the mean intersection over union,structural similarity,mean precision,PAcc,and mean F1-score of our method reach up to 87.31%,94.83%,93.22%,97.39%,and 93.05%,respectively.展开更多
Video-based action recognition is becoming a vital tool in clinical research and neuroscientific study for disorder detection and prediction.However,action recognition currently used in non-human primate(NHP)research ...Video-based action recognition is becoming a vital tool in clinical research and neuroscientific study for disorder detection and prediction.However,action recognition currently used in non-human primate(NHP)research relies heavily on intense manual labor and lacks standardized assessment.In this work,we established two standard benchmark datasets of NHPs in the laboratory:Monkeyin Lab(Mi L),which includes 13 categories of actions and postures,and MiL2D,which includes sequences of two-dimensional(2D)skeleton features.Furthermore,based on recent methodological advances in deep learning and skeleton visualization,we introduced the Monkey Monitor Kit(Mon Kit)toolbox for automatic action recognition,posture estimation,and identification of fine motor activity in monkeys.Using the datasets and Mon Kit,we evaluated the daily behaviors of wild-type cynomolgus monkeys within their home cages and experimental environments and compared these observations with the behaviors exhibited by cynomolgus monkeys possessing mutations in the MECP2 gene as a disease model of Rett syndrome(RTT).Mon Kit was used to assess motor function,stereotyped behaviors,and depressive phenotypes,with the outcomes compared with human manual detection.Mon Kit established consistent criteria for identifying behavior in NHPs with high accuracy and efficiency,thus providing a novel and comprehensive tool for assessing phenotypic behavior in monkeys.展开更多
Deep metric learning(DML)has achieved great results on visual understanding tasks by seamlessly integrating conventional metric learning with deep neural networks.Existing deep metric learning methods focus on designi...Deep metric learning(DML)has achieved great results on visual understanding tasks by seamlessly integrating conventional metric learning with deep neural networks.Existing deep metric learning methods focus on designing pair-based distance loss to decrease intra-class distance while increasing interclass distance.However,these methods fail to preserve the geometric structure of data in the embedding space,which leads to the spatial structure shift across mini-batches and may slow down the convergence of embedding learning.To alleviate these issues,by assuming that the input data is embedded in a lower-dimensional sub-manifold,we propose a novel deep Riemannian metric learning(DRML)framework that exploits the non-Euclidean geometric structural information.Considering that the curvature information of data measures how much the Riemannian(nonEuclidean)metric deviates from the Euclidean metric,we leverage geometry flow,which is called a geometric evolution equation,to characterize the relation between the Riemannian metric and its curvature.Our DRML not only regularizes the local neighborhoods connection of the embeddings at the hidden layer but also adapts the embeddings to preserve the geometric structure of the data.On several benchmark datasets,the proposed DRML outperforms all existing methods and these results demonstrate its effectiveness.展开更多
Modified constraint-induced movement therapy(mCIMT)has shown beneficial effects on motor function improvement after brain injury,but the exact mechanism remains unclear.In this study,amplitude of low frequency fluctua...Modified constraint-induced movement therapy(mCIMT)has shown beneficial effects on motor function improvement after brain injury,but the exact mechanism remains unclear.In this study,amplitude of low frequency fluctuation(ALFF)metrics measured by resting-state functional magnetic resonance imaging was obtained to investigate the efficacy and mechanism of mCIMT in a control co rtical impact(CCI)rat model simulating traumatic brain injury.At 3 days after control co rtical impact model establishment,we found that the mean ALFF(mALFF)signals were decreased in the left motor cortex,somatosensory co rtex,insula cortex and the right motor co rtex,and were increased in the right corpus callosum.After 3 weeks of an 8-hour daily mClMT treatment,the mALFF values were significantly increased in the bilateral hemispheres compared with those at 3 days postoperatively.The mALFF signal valu es of left corpus callosum,left somatosensory cortex,right medial prefro ntal cortex,right motor co rtex,left postero dorsal hippocampus,left motor cortex,right corpus callosum,and right somatosensory cortex were increased in the mCIMT group compared with the control cortical impact group.Finally,we identified brain regions with significantly decreased mALFF valu es at 3 days postoperatively.Pearson correlation coefficients with the right forelimb sliding score indicated that the improvement in motor function of the affected upper limb was associated with an increase in mALFF values in these brain regions.Our findings suggest that functional co rtical plasticity changes after brain injury,and that mCIMT is an effective method to improve affected upper limb motor function by promoting bilateral hemispheric co rtical remodeling.mALFF values correlate with behavio ral changes and can potentially be used as biomarkers to assess dynamic cortical plasticity after traumatic brain injury.展开更多
The performance of medical image classification has been enhanced by deep convolutional neural networks(CNNs),which are typically trained with cross-entropy(CE)loss.However,when the label presents an intrinsic ordinal...The performance of medical image classification has been enhanced by deep convolutional neural networks(CNNs),which are typically trained with cross-entropy(CE)loss.However,when the label presents an intrinsic ordinal property in nature,e.g.,the development from benign to malignant tumor,CE loss cannot take into account such ordinal information to allow for better generalization.To improve model generalization with ordinal information,we propose a novel meta ordinal regression forest(MORF)method for medical image classification with ordinal labels,which learns the ordinal relationship through the combination of convolutional neural network and differential forest in a meta-learning framework.The merits of the proposed MORF come from the following two components:A tree-wise weighting net(TWW-Net)and a grouped feature selection(GFS)module.First,the TWW-Net assigns each tree in the forest with a specific weight that is mapped from the classification loss of the corresponding tree.Hence,all the trees possess varying weights,which is helpful for alleviating the tree-wise prediction variance.Second,the GFS module enables a dynamic forest rather than a fixed one that was previously used,allowing for random feature perturbation.During training,we alternatively optimize the parameters of the CNN backbone and TWW-Net in the meta-learning framework through calculating the Hessian matrix.Experimental results on two medical image classification datasets with ordinal labels,i.e.,LIDC-IDRI and Breast Ultrasound datasets,demonstrate the superior performances of our MORF method over existing state-of-the-art methods.展开更多
BACKGROUND The pontic design of fixed dental prostheses(FDPs)is strongly associated with the phonetic function,and the phonetic function of anterior FDPs with different pontic designs remains understudied.AIM To inves...BACKGROUND The pontic design of fixed dental prostheses(FDPs)is strongly associated with the phonetic function,and the phonetic function of anterior FDPs with different pontic designs remains understudied.AIM To investigate the immediate and short-term influence of pontic design of anterior FDPs on Chinese speech in a clinical case using objective acoustic analysis.METHODS Two FDPs with two types of pontic design(saddle pontic and modified ridge lap pontic)were fabricated for one patient with maxillary anterior teeth missing.The acoustic analysis of patient’s articulation was conducted immediately after wearing the FDPs and 1 wk after wearing these FDPs.RESULTS The effect of FDP on Chinese vowels(/a/,/o/,/e/,/i/,/u/,and/ü/)was insignificant,because the recovery of vowel distortion occurred within 1 wk for both FDPs.Three(/f/,/s/,and/sh/)of eight Chinese fricative consonants were found to have obvious distortions,and the/s/sound distortion last for more than 1 wk for the patient wearing FDP with modified ridge lap pontic design.CONCLUSION The influence of anterior FDP on articulation of Chinese vowels is insignificant,while the articulation of Chinese fricative consonants is more susceptible.When fabricating anterior FDPs for patients with speech related professions,saddle pontic design can be an alternative option compared with modified ridge lap pontic design.展开更多
Nuclear reaction rate A is a significant factor in processes of nucleosyntheses.A multi-layer directed-weighted nuclear reaction network,in which the reaction rate is taken as the weight,and neutron,proton,4 He and th...Nuclear reaction rate A is a significant factor in processes of nucleosyntheses.A multi-layer directed-weighted nuclear reaction network,in which the reaction rate is taken as the weight,and neutron,proton,4 He and the remainder nuclei as the criteria for different reaction layers,is for the first time built based on all thermonuclear reactions in the JINA REACLIB database.Our results show that with the increase in the stellar temperature T9,the distribution of nuclear reaction rates on the R-layer network demonstrates a transition from unimodal to bimodal distributions.Nuclei on the R-layer in the region of A=[1,2.5×101]have a more complicated out-going degree distribution than that in the region of A=[1011,1013],and the number of involved nuclei at T9=1 is very different from the one at T9=3.The redundant nuclei in the region of A=[1,2.5×101]at T9=3 prefer(γ,p)and(γ,α)reactions to the ones at T9=1,which produce nuclei around theβstable line.This work offers a novel way to the big-data analysis on the nuclear reaction network at stellar temperatures.展开更多
Prompt learning has attracted broad attention in computer vision since the large pre-trained visionlanguagemodels (VLMs) exploded. Based on the close relationship between vision and language information builtby VLM, p...Prompt learning has attracted broad attention in computer vision since the large pre-trained visionlanguagemodels (VLMs) exploded. Based on the close relationship between vision and language information builtby VLM, prompt learning becomes a crucial technique in many important applications such as artificial intelligencegenerated content (AIGC). In this survey, we provide a progressive and comprehensive review of visual promptlearning as related to AIGC. We begin by introducing VLM, the foundation of visual prompt learning. Then, wereview the vision prompt learning methods and prompt-guided generative models, and discuss how to improve theefficiency of adapting AIGC models to specific downstream tasks. Finally, we provide some promising researchdirections concerning prompt learning.展开更多
Signal transduction plays important roles in biological systems. Unfortunately, our knowledge about signaling pathways is far from complete. Specifically, the direction of signaling flows is less known even though the...Signal transduction plays important roles in biological systems. Unfortunately, our knowledge about signaling pathways is far from complete. Specifically, the direction of signaling flows is less known even though the signaling molecules of some signaling pathways have been determined. In this paper, we propose a novel hybrid intelligent method, namely HISP （Hybrid Intelligent approach for identifying directed Signaling Pathways）, to determine both the topologies of signaling pathways and the direction of signaling flows within a pathway based on integer linear programming and genetic algorithm. By integrating the protein-protein interaction, gene expression, and gene knockout data, our HISP approach is able to determine the optimal topologies of signaling pathways in an accurate way. Benchmark results on yeast MAPK signaling pathways demonstrate the efficiency of our proposed approach. When applied to the EGFR/ErbB signaling pathway in human hepatocytes, HISP unveils a high-resolution signaling path- way, where many signaling interactions were missing by existing computational approaches.展开更多
Background The association between perivascular space(PVS)and white matter hyperintensity(WMH)has been unclear.Normal-appearing white matter(NAWM)around WMH is also found correlated with the development of focal WMH.T...Background The association between perivascular space(PVS)and white matter hyperintensity(WMH)has been unclear.Normal-appearing white matter(NAWM)around WMH is also found correlated with the development of focal WMH.This study aims to investigate the topological connections among PVS,deep WMH(dWMH)and NAWM around WMH using 7 Tesla(7T)MRI.Methods Thirty-two patients with non-confluent WMHs and 16 subjects without WMHs were recruited from our department and clinic.We compared the PVS burden between patients with and without WMHs using a 5-point scale.Then,the dilatation and the number of PVS within a radius of 1 cm around each dWMH were compared with those of a reference site(without WMH)in the contralateral hemisphere.In this study,we define NAWM as an area within the radius of 1 cm around each dWMH.Furthermore,we assessed the spatial relationship between dWMH and PVS.Results Higher PVS scores in the centrum semiovale were found in patients with>5 dWMHs(median 3)than subjects without dWMH(median 2,p=0.014).We found there was a greater dilatation and a higher number of PVS in NAWM around dWMH than at the reference sites(p<0.001,p<0.001).In addition,79.59%of the dWMHs were spatially connected with PVS.Conclusion dWMH,NAWM surrounding WMH and MRI-visible PVS are spatially correlated in the early stage of cerebral small vessel disease.Future study of WMH and NAWM should not overlook MRI-visible PVS.展开更多
Autism spectrum disorder(ASD)is a highly heritable neurodevelopmental disorder characterized by deficits in social interactions and repetitive behaviors.Although hundreds of ASD risk genes,implicated in synaptic forma...Autism spectrum disorder(ASD)is a highly heritable neurodevelopmental disorder characterized by deficits in social interactions and repetitive behaviors.Although hundreds of ASD risk genes,implicated in synaptic formation and transcriptional regulation,have been identified through human genetic studies,the East Asian ASD cohorts are still under-represented in genome-wide genetic studies.Here,we applied whole-exome sequencing to 369 ASD trios including probands and unaffected parents of Chinese origin.Using a joint-calling analytical pipeline based on GATK toolkits,we identified numerous de novo mutations including 55 high-impact variants and 165 moderate-impact variants,as well as de novo copy number variations containing known ASD-related genes.Importantly,combined with single-cell sequencing data from the developing human brain,we found that the expression of genes with de novo mutations was specifically enriched in the pre-,post-central gyrus(PRC,PC)and banks of the superior temporal(BST)regions in the human brain.By further analyzing the brain imaging data with ASD and healthy controls,we found that the gray volume of the right BST in ASD patients was significantly decreased compared to healthy controls,suggesting the potential structural deficits associated with ASD.Finally,we found a decrease in the seed-based functional connectivity between BST/PC/PRC and sensory areas,the insula,as well as the frontal lobes in ASD patients.This work indicated that combinatorial analysis with genome-wide screening,single-cell sequencing,and brain imaging data reveal the brain regions contributing to the etiology of ASD.展开更多
The complexity of the brain has attracted scientists from all over the world.Much effort has been paid to explore the mechanisms from genetics to molecules,from cells to circuits,and from Caenorhabditis elegans to hum...The complexity of the brain has attracted scientists from all over the world.Much effort has been paid to explore the mechanisms from genetics to molecules,from cells to circuits,and from Caenorhabditis elegans to humans.Brain research contributes to the development of new technologies,especially artificial intelligence(AI).According to the International Data Corporation,the global revenue of the AI market is expected to exceed$500 billion by 2023 and$900 billion by 2026,highlighting the great importance of brain research for social progress.展开更多
As one of the state-of-the-art automated function prediction(AFP)methods,NetGO 2.0 integrates multi-source information to improve the performance.However,it mainly utilizes the proteins with experimentally supported f...As one of the state-of-the-art automated function prediction(AFP)methods,NetGO 2.0 integrates multi-source information to improve the performance.However,it mainly utilizes the proteins with experimentally supported functional annotations without leveraging valuable information from a vast number of unannotated proteins.Recently,protein language models have been proposed to learn informative representations[e.g.,Evolutionary Scale Modeling(ESM)-1b embedding] from protein sequences based on self-supervision.Here,we represented each protein by ESM-1b and used logistic regression(LR)to train a new model,LR-ESM,for AFP.The experimental results showed that LR-ESM achieved comparable performance with the best-performing component of NetGO 2.0.Therefore,by incorporating LR-ESM into NetGO 2.0,we developed NetGO 3.0 to improve the performance of AFP extensively.展开更多
We have all seen the current attention,and even hype,on big models and artificial general intelligence(AGl).Indeed,many colleagues are arguing that we have achieved AGl with the current version of ChatGPT.Really?Looki...We have all seen the current attention,and even hype,on big models and artificial general intelligence(AGl).Indeed,many colleagues are arguing that we have achieved AGl with the current version of ChatGPT.Really?Looking back,it is not surprising to have a machine,even a normal calculator,which can outperform us.For example,a calculator can easily beat most of us on the multiplication of two large numbers with its speed,while a basic laptop can store many books,but humans have far weaker means of recall.So it is something that we have already got used to,that man-made machines can perform certain tasks far betterthan us.展开更多
Background Previous prediction algorithms for cardiovascular diseases(CVD)were established using risk factors retrieved largely based on empirical clinical knowledge.This study sought to identify predictors among a co...Background Previous prediction algorithms for cardiovascular diseases(CVD)were established using risk factors retrieved largely based on empirical clinical knowledge.This study sought to identify predictors among a comprehensive variable space,and then employ machine learning(ML)algorithms to develop a novel CVD risk prediction model.Methods From a longitudinal population-based cohort of UK Biobank,this study included 473611 CVD-free participants aged between 37 and 73 years old.We implemented an ML-based data-driven pipeline to identify predictors from 645 candidate variables covering a comprehensive range of health-related factors and assessed multiple ML classifiers to establish a risk prediction model on 10-year incident CVD.The model was validated through a leave-one center-out cross-validation.Results During a median follow-up of 12.2 years,31466 participants developed CVD within 10 years after baseline visits.A novel UK Biobank CVD risk prediction(UKCRP)model was established that comprised 10 predictors including age,sex,medication of cholesterol and blood pressure,cholesterol ratio(total/high-density lipoprotein),systolic blood pressure,previous angina or heart disease,number of medications taken,cystatin C,chest pain and pack-years of smoking.Our model obtained satisfied discriminative performance with an area under the receiver operating characteristic curve(AUC)of 0.762±0.010 that outperformed multiple existing clinical models,and it was well-calibrated with a Brier Score of 0.057±0.006.Further,the UKCRP can obtain comparable performance for myocardial infarction(AUC 0.774±0.011)and ischaemic stroke(AUC 0.730±0.020),but inferior performance for haemorrhagic stroke(AUC 0.644±0.026).Conclusion ML-based classification models can learn expressive representations from potential high-risked CVD participants who may benefit from earlier clinical decisions.展开更多
Quantification of brain oxygenation and metabolism,both of which are indicators of the level of brain activity,plays a vital role in understanding the cerebral perfusion and the pathophysiology of brain disorders.Magn...Quantification of brain oxygenation and metabolism,both of which are indicators of the level of brain activity,plays a vital role in understanding the cerebral perfusion and the pathophysiology of brain disorders.Magnetic resonance imaging(MRI),a widely used clinical imaging technique,which is very sensitive to magnetic susceptibility,has the possibility of substitut-ing positron emission tomography(PET)in measuring oxygen metabolism.This review mainly focuses on the quantitative blood oxygenation level-dependent(qBOLD)method for the evaluation of oxygen extraction fraction(OEF)in the brain.Here,we review the theoretic basis of qBOLD,as well as existing acquisition and quantification methods.Some published clinical studies are also presented,and the pros and cons of qBOLD method are discussed as well.展开更多
基金supported by the National Natural Science Foundation of China, No.61932008Natural Science Foundation of Shanghai, No.21ZR1403200 (both to JC)。
文摘Neurodegenerative diseases cause great medical and economic burdens for both patients and society;however, the complex molecular mechanisms thereof are not yet well understood. With the development of high-coverage sequencing technology, researchers have started to notice that genomic repeat regions, previously neglected in search of disease culprits, are active contributors to multiple neurodegenerative diseases. In this review, we describe the association between repeat element variants and multiple degenerative diseases through genome-wide association studies and targeted sequencing. We discuss the identification of disease-relevant repeat element variants, further powered by the advancement of long-read sequencing technologies and their related tools, and summarize recent findings in the molecular mechanisms of repeat element variants in brain degeneration, such as those causing transcriptional silencing or RNA-mediated gain of toxic function. Furthermore, we describe how in silico predictions using innovative computational models, such as deep learning language models, could enhance and accelerate our understanding of the functional impact of repeat element variants. Finally, we discuss future directions to advance current findings for a better understanding of neurodegenerative diseases and the clinical applications of genomic repeat elements.
基金supported by the National Natural Science Foundation of China,Nos.81871836(to MZ),82172554(to XH),and 81802249(to XH),81902301(to JW)the National Key R&D Program of China,Nos.2018YFC2001600(to JX)and 2018YFC2001604(to JX)+3 种基金Shanghai Rising Star Program,No.19QA1409000(to MZ)Shanghai Municipal Commission of Health and Family Planning,No.2018YQ02(to MZ)Shanghai Youth Top Talent Development PlanShanghai“Rising Stars of Medical Talent”Youth Development Program,No.RY411.19.01.10(to XH)。
文摘Distinct brain remodeling has been found after different nerve reconstruction strategies,including motor representation of the affected limb.However,differences among reconstruction strategies at the brain network level have not been elucidated.This study aimed to explore intranetwork changes related to altered peripheral neural pathways after different nerve reconstruction surgeries,including nerve repair,endto-end nerve transfer,and end-to-side nerve transfer.Sprague–Dawley rats underwent complete left brachial plexus transection and were divided into four equal groups of eight:no nerve repair,grafted nerve repair,phrenic nerve end-to-end transfer,and end-to-side transfer with a graft sutured to the anterior upper trunk.Resting-state brain functional magnetic resonance imaging was obtained 7 months after surgery.The independent component analysis algorithm was utilized to identify group-level network components of interest and extract resting-state functional connectivity values of each voxel within the component.Alterations in intra-network resting-state functional connectivity were compared among the groups.Target muscle reinnervation was assessed by behavioral observation(elbow flexion)and electromyography.The results showed that alterations in the sensorimotor and interoception networks were mostly related to changes in the peripheral neural pathway.Nerve repair was related to enhanced connectivity within the sensorimotor network,while end-to-side nerve transfer might be more beneficial for restoring control over the affected limb by the original motor representation.The thalamic-cortical pathway was enhanced within the interoception network after nerve repair and end-to-end nerve transfer.Brain areas related to cognition and emotion were enhanced after end-to-side nerve transfer.Our study revealed important brain networks related to different nerve reconstructions.These networks may be potential targets for enhancing motor recovery.
基金supported in part by the National Natural Science Foundation of China (62176059, 62101136)。
文摘Binaural rendering is of great interest to virtual reality and immersive media. Although humans can naturally use their two ears to perceive the spatial information contained in sounds, it is a challenging task for machines to achieve binaural rendering since the description of a sound field often requires multiple channels and even the metadata of the sound sources. In addition, the perceived sound varies from person to person even in the same sound field. Previous methods generally rely on individual-dependent head-related transferred function(HRTF)datasets and optimization algorithms that act on HRTFs. In practical applications, there are two major drawbacks to existing methods. The first is a high personalization cost, as traditional methods achieve personalized needs by measuring HRTFs. The second is insufficient accuracy because the optimization goal of traditional methods is to retain another part of information that is more important in perception at the cost of discarding a part of the information. Therefore, it is desirable to develop novel techniques to achieve personalization and accuracy at a low cost. To this end, we focus on the binaural rendering of ambisonic and propose 1) channel-shared encoder and channel-compared attention integrated into neural networks and 2) a loss function quantifying interaural level differences to deal with spatial information. To verify the proposed method, we collect and release the first paired ambisonic-binaural dataset and introduce three metrics to evaluate the content information and spatial information accuracy of the end-to-end methods. Extensive experimental results on the collected dataset demonstrate the superior performance of the proposed method and the shortcomings of previous methods.
基金supported by an unrestricted,investigator-initiated research grant by Scenery(BS),which provided the devices used.The project was sponsored by SJTU Trans-med Awards Research(2019015 to BS)Shanghai Clinical Research Centre for Mental Health(19MC191100 to BS)+3 种基金sponsored by the National Natural Science Foundation of China(81771482)supported by the Guangci Professorship Programme of Ruijin Hospital(N/A)and a Medical Research Council Senior Clinical Fellowship(MR/P008747/1)sponsored by the National Natural Science Foundation of China(82101546)the Shanghai Sailing Program(21YF1426700).The funding sources were not involved in the design and conduct of the study。
文摘Background Structural imaging holds great potential for precise targeting and stimulation for deep brain stimulation(DBS).The anatomical information it provides may serve as potential biomarkers for predicting the efficacy of DBS in treatment-resistant depression(TRD).Aims The primary aim is to identify preoperative imaging biomarkers that correlate with the efficacy of DBS in patients with TRD.Methods Preoperative imaging parameters were estimated and correlated with the 6-month clinical outcome of patients with TRD receiving combined bed nucleus of the stria terminalis(BNST)-nucleus accumbens(NAc)DBS.White matter(WM)properties were extracted and compared between the response/non-response and remission/non-remission groups.Structural connectome was constructed and analysed using graph theory.Distances of the volume of activated tissue(VAT)to the main modulating tracts were also estimated to evaluate the correlations.Results Differences in fibre bundle properties of tracts,including superior thalamic radiation and reticulospinal tract,were observed between the remission and nonremission groups.Distance of the centre of the VAT to tracts connecting the ventral tegmental area and the anterior limb of internal capsule on the left side varied between the remission and non-remission groups(p=0.010,t=3.07).The normalised clustering coefficient(γ)and the small-world property(σ)in graph analysis correlated with the symptom improvement after the correction of age.Conclusions Presurgical structural alterations in WM tracts connecting the frontal area with subcortical regions,as well as the distance of the VAT to the modulating tracts,may influence the clinical outcome of BNST-NAc DBS.These findings provide potential imaging biomarkers for the DBS treatment for patients with TRD.
基金supported by National Natural Science Foundation of China (No.62101136)Shanghai Sailing Program (No.21YF1402800)+3 种基金Shanghai Municipal Science and Technology Major Project (No.2018SHZDZX01)ZJLab,Shanghai Municipal of Science and Technology Project (No.20JC1419500)Natural Science Foundation of Chongqing (No.CSTB2022NSCQ-MSX0360)Shanghai Center for Brain Science and Brain-inspired Technology.
文摘Spectral computed tomography(CT)based on photon counting detectors can resolve the energy of every single photon interacting with the sensor layer and be used to analyze material attenuation information under different energy ranges,which can be helpful for material decomposition studies.However,there is a considerable amount of inherent quantum noise in narrow energy bins,resulting in a low signal-to-noise ratio,which can consequently affect the material decomposition performance in the image domain.Deep learning technology is currently widely used in medical image segmentation,denoising,and recognition.In order to improve the results of material decomposition,we propose an attention-based global convolutional generative adversarial network(AGC-GAN)to decompose different materials for spectral CT.Specifically,our network is a global convolutional neural network based on an attention mechanism that is combined with a generative adversarial network.The global convolutional network based on the attention mechanism is used as the generator,and a patchGAN discriminant network is used as the discriminator.Meanwhile,a clinical spectral CT image dataset is used to verify the feasibility of our proposed approach.Extensive experimental results demonstrate that AGC-GAN achieves a better material decomposition performance than vanilla U-Net,fully convolutional network,and fully convolutional denseNet.Remarkably,the mean intersection over union,structural similarity,mean precision,PAcc,and mean F1-score of our method reach up to 87.31%,94.83%,93.22%,97.39%,and 93.05%,respectively.
基金supported by the National Key R&D Program of China (2021ZD0202805,2019YFA0709504,2021ZD0200900)National Defense Science and Technology Innovation Special Zone Spark Project (20-163-00-TS-009-152-01)+4 种基金National Natural Science Foundation of China (31900719,U20A20227,82125008)Innovative Research Team of High-level Local Universities in Shanghai,Science and Technology Committee Rising-Star Program (19QA1401400)111 Project (B18015)Shanghai Municipal Science and Technology Major Project (2018SHZDZX01)Shanghai Center for Brain Science and Brain-Inspired Technology。
文摘Video-based action recognition is becoming a vital tool in clinical research and neuroscientific study for disorder detection and prediction.However,action recognition currently used in non-human primate(NHP)research relies heavily on intense manual labor and lacks standardized assessment.In this work,we established two standard benchmark datasets of NHPs in the laboratory:Monkeyin Lab(Mi L),which includes 13 categories of actions and postures,and MiL2D,which includes sequences of two-dimensional(2D)skeleton features.Furthermore,based on recent methodological advances in deep learning and skeleton visualization,we introduced the Monkey Monitor Kit(Mon Kit)toolbox for automatic action recognition,posture estimation,and identification of fine motor activity in monkeys.Using the datasets and Mon Kit,we evaluated the daily behaviors of wild-type cynomolgus monkeys within their home cages and experimental environments and compared these observations with the behaviors exhibited by cynomolgus monkeys possessing mutations in the MECP2 gene as a disease model of Rett syndrome(RTT).Mon Kit was used to assess motor function,stereotyped behaviors,and depressive phenotypes,with the outcomes compared with human manual detection.Mon Kit established consistent criteria for identifying behavior in NHPs with high accuracy and efficiency,thus providing a novel and comprehensive tool for assessing phenotypic behavior in monkeys.
基金supported in part by the Young Elite Scientists Sponsorship Program by CAST(2022QNRC001)the National Natural Science Foundation of China(61621003,62101136)+2 种基金Natural Science Foundation of Shanghai(21ZR1403600)Shanghai Municipal Science and Technology Major Project(2018SHZDZX01)ZJLab,and Shanghai Municipal of Science and Technology Project(20JC1419500)。
文摘Deep metric learning(DML)has achieved great results on visual understanding tasks by seamlessly integrating conventional metric learning with deep neural networks.Existing deep metric learning methods focus on designing pair-based distance loss to decrease intra-class distance while increasing interclass distance.However,these methods fail to preserve the geometric structure of data in the embedding space,which leads to the spatial structure shift across mini-batches and may slow down the convergence of embedding learning.To alleviate these issues,by assuming that the input data is embedded in a lower-dimensional sub-manifold,we propose a novel deep Riemannian metric learning(DRML)framework that exploits the non-Euclidean geometric structural information.Considering that the curvature information of data measures how much the Riemannian(nonEuclidean)metric deviates from the Euclidean metric,we leverage geometry flow,which is called a geometric evolution equation,to characterize the relation between the Riemannian metric and its curvature.Our DRML not only regularizes the local neighborhoods connection of the embeddings at the hidden layer but also adapts the embeddings to preserve the geometric structure of the data.On several benchmark datasets,the proposed DRML outperforms all existing methods and these results demonstrate its effectiveness.
基金supported by the National Key R&D Program of China,Nos.2020YFC2004202(to DSX),2018 YFC2001600(to XYH)the National Natural Science Foundation of China,Nos.81974358(to DSX),81802249(to XYH)and 82172554(to XYH)。
文摘Modified constraint-induced movement therapy(mCIMT)has shown beneficial effects on motor function improvement after brain injury,but the exact mechanism remains unclear.In this study,amplitude of low frequency fluctuation(ALFF)metrics measured by resting-state functional magnetic resonance imaging was obtained to investigate the efficacy and mechanism of mCIMT in a control co rtical impact(CCI)rat model simulating traumatic brain injury.At 3 days after control co rtical impact model establishment,we found that the mean ALFF(mALFF)signals were decreased in the left motor cortex,somatosensory co rtex,insula cortex and the right motor co rtex,and were increased in the right corpus callosum.After 3 weeks of an 8-hour daily mClMT treatment,the mALFF values were significantly increased in the bilateral hemispheres compared with those at 3 days postoperatively.The mALFF signal valu es of left corpus callosum,left somatosensory cortex,right medial prefro ntal cortex,right motor co rtex,left postero dorsal hippocampus,left motor cortex,right corpus callosum,and right somatosensory cortex were increased in the mCIMT group compared with the control cortical impact group.Finally,we identified brain regions with significantly decreased mALFF valu es at 3 days postoperatively.Pearson correlation coefficients with the right forelimb sliding score indicated that the improvement in motor function of the affected upper limb was associated with an increase in mALFF values in these brain regions.Our findings suggest that functional co rtical plasticity changes after brain injury,and that mCIMT is an effective method to improve affected upper limb motor function by promoting bilateral hemispheric co rtical remodeling.mALFF values correlate with behavio ral changes and can potentially be used as biomarkers to assess dynamic cortical plasticity after traumatic brain injury.
基金This work was supported in part by the Natural Science Foundation of Shanghai(21ZR1403600)the National Natural Science Foundation of China(62176059)+3 种基金Shanghai Municipal Science and Technology Major Project(2018SHZDZX01)Zhang Jiang Laboratory,Shanghai Sailing Program(21YF1402800)Shanghai Municipal of Science and Technology Project(20JC1419500)Shanghai Center for Brain Science and Brain-inspired Technology.
文摘The performance of medical image classification has been enhanced by deep convolutional neural networks(CNNs),which are typically trained with cross-entropy(CE)loss.However,when the label presents an intrinsic ordinal property in nature,e.g.,the development from benign to malignant tumor,CE loss cannot take into account such ordinal information to allow for better generalization.To improve model generalization with ordinal information,we propose a novel meta ordinal regression forest(MORF)method for medical image classification with ordinal labels,which learns the ordinal relationship through the combination of convolutional neural network and differential forest in a meta-learning framework.The merits of the proposed MORF come from the following two components:A tree-wise weighting net(TWW-Net)and a grouped feature selection(GFS)module.First,the TWW-Net assigns each tree in the forest with a specific weight that is mapped from the classification loss of the corresponding tree.Hence,all the trees possess varying weights,which is helpful for alleviating the tree-wise prediction variance.Second,the GFS module enables a dynamic forest rather than a fixed one that was previously used,allowing for random feature perturbation.During training,we alternatively optimize the parameters of the CNN backbone and TWW-Net in the meta-learning framework through calculating the Hessian matrix.Experimental results on two medical image classification datasets with ordinal labels,i.e.,LIDC-IDRI and Breast Ultrasound datasets,demonstrate the superior performances of our MORF method over existing state-of-the-art methods.
基金the National Natural Science Foundation of China,No.81901060the China Postdoctoral Science Foundation Funded Project,No.2018M640931the Science and Technology Key Research and Development Program of Sichuan Province,No.2019YFS0142.
文摘BACKGROUND The pontic design of fixed dental prostheses(FDPs)is strongly associated with the phonetic function,and the phonetic function of anterior FDPs with different pontic designs remains understudied.AIM To investigate the immediate and short-term influence of pontic design of anterior FDPs on Chinese speech in a clinical case using objective acoustic analysis.METHODS Two FDPs with two types of pontic design(saddle pontic and modified ridge lap pontic)were fabricated for one patient with maxillary anterior teeth missing.The acoustic analysis of patient’s articulation was conducted immediately after wearing the FDPs and 1 wk after wearing these FDPs.RESULTS The effect of FDP on Chinese vowels(/a/,/o/,/e/,/i/,/u/,and/ü/)was insignificant,because the recovery of vowel distortion occurred within 1 wk for both FDPs.Three(/f/,/s/,and/sh/)of eight Chinese fricative consonants were found to have obvious distortions,and the/s/sound distortion last for more than 1 wk for the patient wearing FDP with modified ridge lap pontic design.CONCLUSION The influence of anterior FDP on articulation of Chinese vowels is insignificant,while the articulation of Chinese fricative consonants is more susceptible.When fabricating anterior FDPs for patients with speech related professions,saddle pontic design can be an alternative option compared with modified ridge lap pontic design.
基金Supported by the National Natural Science Foundation of China(Grant Nos.11890714,11421505,11875133,and 11075057)the National Key R&D Program of China(Grant No.2018YFB2101302)+1 种基金the Key Research Program of Frontier Sciences of the CAS(Grant No.QYZDJ-SSW-SLH002)the Strategic Priority Research Program of the Chinese Academy of Sciences(Grant No.XDB34030200)。
文摘Nuclear reaction rate A is a significant factor in processes of nucleosyntheses.A multi-layer directed-weighted nuclear reaction network,in which the reaction rate is taken as the weight,and neutron,proton,4 He and the remainder nuclei as the criteria for different reaction layers,is for the first time built based on all thermonuclear reactions in the JINA REACLIB database.Our results show that with the increase in the stellar temperature T9,the distribution of nuclear reaction rates on the R-layer network demonstrates a transition from unimodal to bimodal distributions.Nuclei on the R-layer in the region of A=[1,2.5×101]have a more complicated out-going degree distribution than that in the region of A=[1011,1013],and the number of involved nuclei at T9=1 is very different from the one at T9=3.The redundant nuclei in the region of A=[1,2.5×101]at T9=3 prefer(γ,p)and(γ,α)reactions to the ones at T9=1,which produce nuclei around theβstable line.This work offers a novel way to the big-data analysis on the nuclear reaction network at stellar temperatures.
基金Project supported by the National Natural Science Foundation of China(Nos.62306075 and 62101136)the China Postdoctoral Science Foundation(No.2022TQ0069)+2 种基金the Natural Science Foundation of Shanghai,China(No.21ZR1403600)the Shanghai Municipal of Science and Technology Project,China(No.20JC1419500)the Shanghai Center for Brain Science and Brain-Inspired Technology,China。
文摘Prompt learning has attracted broad attention in computer vision since the large pre-trained visionlanguagemodels (VLMs) exploded. Based on the close relationship between vision and language information builtby VLM, prompt learning becomes a crucial technique in many important applications such as artificial intelligencegenerated content (AIGC). In this survey, we provide a progressive and comprehensive review of visual promptlearning as related to AIGC. We begin by introducing VLM, the foundation of visual prompt learning. Then, wereview the vision prompt learning methods and prompt-guided generative models, and discuss how to improve theefficiency of adapting AIGC models to specific downstream tasks. Finally, we provide some promising researchdirections concerning prompt learning.
文摘Signal transduction plays important roles in biological systems. Unfortunately, our knowledge about signaling pathways is far from complete. Specifically, the direction of signaling flows is less known even though the signaling molecules of some signaling pathways have been determined. In this paper, we propose a novel hybrid intelligent method, namely HISP （Hybrid Intelligent approach for identifying directed Signaling Pathways）, to determine both the topologies of signaling pathways and the direction of signaling flows within a pathway based on integer linear programming and genetic algorithm. By integrating the protein-protein interaction, gene expression, and gene knockout data, our HISP approach is able to determine the optimal topologies of signaling pathways in an accurate way. Benchmark results on yeast MAPK signaling pathways demonstrate the efficiency of our proposed approach. When applied to the EGFR/ErbB signaling pathway in human hepatocytes, HISP unveils a high-resolution signaling path- way, where many signaling interactions were missing by existing computational approaches.
基金support for the research,authorship,and/or publication of this article:This study was funded by the National Key Research and Development Project(NO.2019YFC1711600,2019YFC1711603)National Natural Science Foundation of China(No.81771288)Clinical Research Plan of Shanghai Hospital Development Center(No.SHDC2020CR2046B).
文摘Background The association between perivascular space(PVS)and white matter hyperintensity(WMH)has been unclear.Normal-appearing white matter(NAWM)around WMH is also found correlated with the development of focal WMH.This study aims to investigate the topological connections among PVS,deep WMH(dWMH)and NAWM around WMH using 7 Tesla(7T)MRI.Methods Thirty-two patients with non-confluent WMHs and 16 subjects without WMHs were recruited from our department and clinic.We compared the PVS burden between patients with and without WMHs using a 5-point scale.Then,the dilatation and the number of PVS within a radius of 1 cm around each dWMH were compared with those of a reference site(without WMH)in the contralateral hemisphere.In this study,we define NAWM as an area within the radius of 1 cm around each dWMH.Furthermore,we assessed the spatial relationship between dWMH and PVS.Results Higher PVS scores in the centrum semiovale were found in patients with>5 dWMHs(median 3)than subjects without dWMH(median 2,p=0.014).We found there was a greater dilatation and a higher number of PVS in NAWM around dWMH than at the reference sites(p<0.001,p<0.001).In addition,79.59%of the dWMHs were spatially connected with PVS.Conclusion dWMH,NAWM surrounding WMH and MRI-visible PVS are spatially correlated in the early stage of cerebral small vessel disease.Future study of WMH and NAWM should not overlook MRI-visible PVS.
基金This work was supported by the National Natural Science Foundation of China(31625013,81941015,32000726,and 61973086)the Shanghai Brain-Intelligence Project from STCSM(16JC1420501)+2 种基金the Strategic Priority Research Program of the Chinese Academy of Sciences(XDBS01060200)the Program of Shanghai Academic Research LeaderThe Open Large Infrastructure Research of the Chinese Academy of Sciences,and the Shanghai Municipal Science and Technology Major Project(2018SHZDZX01).
文摘Autism spectrum disorder(ASD)is a highly heritable neurodevelopmental disorder characterized by deficits in social interactions and repetitive behaviors.Although hundreds of ASD risk genes,implicated in synaptic formation and transcriptional regulation,have been identified through human genetic studies,the East Asian ASD cohorts are still under-represented in genome-wide genetic studies.Here,we applied whole-exome sequencing to 369 ASD trios including probands and unaffected parents of Chinese origin.Using a joint-calling analytical pipeline based on GATK toolkits,we identified numerous de novo mutations including 55 high-impact variants and 165 moderate-impact variants,as well as de novo copy number variations containing known ASD-related genes.Importantly,combined with single-cell sequencing data from the developing human brain,we found that the expression of genes with de novo mutations was specifically enriched in the pre-,post-central gyrus(PRC,PC)and banks of the superior temporal(BST)regions in the human brain.By further analyzing the brain imaging data with ASD and healthy controls,we found that the gray volume of the right BST in ASD patients was significantly decreased compared to healthy controls,suggesting the potential structural deficits associated with ASD.Finally,we found a decrease in the seed-based functional connectivity between BST/PC/PRC and sensory areas,the insula,as well as the frontal lobes in ASD patients.This work indicated that combinatorial analysis with genome-wide screening,single-cell sequencing,and brain imaging data reveal the brain regions contributing to the etiology of ASD.
基金supported by the National Programs for Brain Science and Brain-like Intelligence Technology of China(2021ZD0200800).
文摘The complexity of the brain has attracted scientists from all over the world.Much effort has been paid to explore the mechanisms from genetics to molecules,from cells to circuits,and from Caenorhabditis elegans to humans.Brain research contributes to the development of new technologies,especially artificial intelligence(AI).According to the International Data Corporation,the global revenue of the AI market is expected to exceed$500 billion by 2023 and$900 billion by 2026,highlighting the great importance of brain research for social progress.
基金supported by the National Natural Science Foundation of China(Grant Nos.61872094 and 62272105)the Shanghai Municipal Science and Technology Major Project(Grant No.2018SHZDZX01)+2 种基金the ZJ Lab,and the Shanghai Research Center for Brain Science and Brain-Inspired Intelligence Technology.Shaojun Wang and Ronghui You have been supported by the lll Project(Grant No.B18015)the Shanghai Municipal Science and Technology Major Project(Grant No.2017SHZDZX01)the Information Technology Facility,CAS-MPG Partner Institute for Computational Biology,Shanghai Institute for Biological Sciences,Chinese Academy of Sciences.Yi Xiong has been supported by the National Natural Science Foundation of China(Grant Nos.61832019 and 62172274).
文摘As one of the state-of-the-art automated function prediction(AFP)methods,NetGO 2.0 integrates multi-source information to improve the performance.However,it mainly utilizes the proteins with experimentally supported functional annotations without leveraging valuable information from a vast number of unannotated proteins.Recently,protein language models have been proposed to learn informative representations[e.g.,Evolutionary Scale Modeling(ESM)-1b embedding] from protein sequences based on self-supervision.Here,we represented each protein by ESM-1b and used logistic regression(LR)to train a new model,LR-ESM,for AFP.The experimental results showed that LR-ESM achieved comparable performance with the best-performing component of NetGO 2.0.Therefore,by incorporating LR-ESM into NetGO 2.0,we developed NetGO 3.0 to improve the performance of AFP extensively.
基金Shanghai Municipal Science and Technology Major Project,Grant/Award Number:2018SHZDZX01ZJ Lab+1 种基金Shanghai Center for Brain Science and Brain-Inspired Technology111Project,Grant/Award Number:B18015。
文摘We have all seen the current attention,and even hype,on big models and artificial general intelligence(AGl).Indeed,many colleagues are arguing that we have achieved AGl with the current version of ChatGPT.Really?Looking back,it is not surprising to have a machine,even a normal calculator,which can outperform us.For example,a calculator can easily beat most of us on the multiplication of two large numbers with its speed,while a basic laptop can store many books,but humans have far weaker means of recall.So it is something that we have already got used to,that man-made machines can perform certain tasks far betterthan us.
基金the National Natural Science Foundation of China(82071997,82071201)National Key R&D Program of China(2018YFC1312904,2019YFA0709502)+6 种基金Science and Technology Innovation 2030 Major Projects(2022ZD0211600)Shanghai Municipal Science and Technology Major Project(2018SHZDZX01)the 111 Project(B18015)hanghai Rising-Star Program(21QA1408700)Research Start-up Fund of Huashan Hospital(2022QD002)Excellence 2025 Talent Cultivation Program at Fudan University(3030277001)Shanghai Municipal Health Commission New Interdisciplinary Research Project(2022JC014).
文摘Background Previous prediction algorithms for cardiovascular diseases(CVD)were established using risk factors retrieved largely based on empirical clinical knowledge.This study sought to identify predictors among a comprehensive variable space,and then employ machine learning(ML)algorithms to develop a novel CVD risk prediction model.Methods From a longitudinal population-based cohort of UK Biobank,this study included 473611 CVD-free participants aged between 37 and 73 years old.We implemented an ML-based data-driven pipeline to identify predictors from 645 candidate variables covering a comprehensive range of health-related factors and assessed multiple ML classifiers to establish a risk prediction model on 10-year incident CVD.The model was validated through a leave-one center-out cross-validation.Results During a median follow-up of 12.2 years,31466 participants developed CVD within 10 years after baseline visits.A novel UK Biobank CVD risk prediction(UKCRP)model was established that comprised 10 predictors including age,sex,medication of cholesterol and blood pressure,cholesterol ratio(total/high-density lipoprotein),systolic blood pressure,previous angina or heart disease,number of medications taken,cystatin C,chest pain and pack-years of smoking.Our model obtained satisfied discriminative performance with an area under the receiver operating characteristic curve(AUC)of 0.762±0.010 that outperformed multiple existing clinical models,and it was well-calibrated with a Brier Score of 0.057±0.006.Further,the UKCRP can obtain comparable performance for myocardial infarction(AUC 0.774±0.011)and ischaemic stroke(AUC 0.730±0.020),but inferior performance for haemorrhagic stroke(AUC 0.644±0.026).Conclusion ML-based classification models can learn expressive representations from potential high-risked CVD participants who may benefit from earlier clinical decisions.
基金supported by the National Natural Science Foundation of China(No.81971583)National Key R&D Program of China(No.2018YFC1312900)+1 种基金Shanghai Natural Science Foundation(No.20ZR1406400)Shanghai Municipal Science and Technology Major Project(No.2017SHZDZX01,No.2018SHZDZX01)and ZJLab.
文摘Quantification of brain oxygenation and metabolism,both of which are indicators of the level of brain activity,plays a vital role in understanding the cerebral perfusion and the pathophysiology of brain disorders.Magnetic resonance imaging(MRI),a widely used clinical imaging technique,which is very sensitive to magnetic susceptibility,has the possibility of substitut-ing positron emission tomography(PET)in measuring oxygen metabolism.This review mainly focuses on the quantitative blood oxygenation level-dependent(qBOLD)method for the evaluation of oxygen extraction fraction(OEF)in the brain.Here,we review the theoretic basis of qBOLD,as well as existing acquisition and quantification methods.Some published clinical studies are also presented,and the pros and cons of qBOLD method are discussed as well.
