Arrow Research search

Author name cluster

Yun Zhou

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

26 papers
2 author rows

Possible papers

26

AAAI Conference 2026 Conference Paper

Collaborative LLM Numerical Reasoning with Local Data Protection

  • Min Zhang
  • Yuzhe Lu
  • Yun Zhou
  • Panpan Xu
  • Lin Lee Cheong
  • Chang-Tien Lu
  • Haozhu Wang

Numerical reasoning over documents, which demands both contextual understanding and logical inference, is challenging for low-capacity local models deployed on computation-constrained devices. Although such complex reasoning queries could be routed to powerful remote models like GPT-4, exposing local data raises significant data leakage concerns. Existing mitigation methods generate problem descriptions or examples for remote assistance. However, the inherent complexity of numerical reasoning hinders the local model from generating logically equivalent queries and accurately inferring answers with remote guidance. In this paper, we present a model collaboration framework with two key innovations: (1) a context-aware synthesis strategy that shifts the query topics while preserving reasoning patterns; and (2) a tool-based answer reconstruction approach that reuses the remote-generated plug-and-play solution with code snippets. Experimental results demonstrate that our method achieves better reasoning accuracy than solely using local models while providing stronger data protection than fully relying on remote models. Furthermore, our method improves accuracy by 16.2% - 43.6% while reducing data leakage by 2.3% - 44.6% compared to existing data protection approaches.

PDF Details DOI

EAAI Journal 2026 Journal Article

Multi-view feature learning and enhanced hypergraph neural networks for synergistic prediction of drug combination

  • Wei Wang
  • Mengyi Ma
  • Hongjun Zhang
  • Yun Zhou
  • Guangsheng Wu

Drug combination therapy demonstrates more significant efficacy than monotherapy in cancer treatment. Despite the proposal of several computational approaches aimed at effectively identifying synergistic drug combinations, challenges persist due to inadequate multi-level learning within multimodal data. Furthermore, existing models still struggle to adequately capture the complex biological network interactions between drug combinations and cell lines. To overcome these issues, we propose a novel hypergraph neural network method for synergistic drug combination prediction. This method integrates multi-view feature learning and enhanced hypergraph neural networks to improve drug combination prediction. First, multi-view learning is independently applied to the multimodal data of drugs and cell lines. This framework employs a fine-tuned ChemBERTa model enhanced by contrastive learning to effectively capture the contextual information of drug SMILES. Second, enhanced hypergraph neural networks equipped with a multi-head attention mechanism are designed to capture the complex topological information between drugs and cell lines and to address the limited ability of the hypergraph to capture global information. Third, the similarity-based multi-task supervision module further stabilizes the model. The experimental results show that our method outperforms state-of-the-art methods in various scenarios, including leave-drug-combination-out, leave-cell-out, and leave-drug-out scenarios. Specifically, in the leave-drug combination-out scenario, our method achieves a Mean Squared Error of 163. 635, a Root Mean Squared Error of 12. 792, and a Pearson Correlation Coefficient of 0. 751. Finally, a case study demonstrates the efficacy of the model in predicting novel synergistic drug combinations.

Details DOI

AAAI Conference 2026 Conference Paper

PQDA:Policy-Aligned Q-Consistency Meets Decoupled Augmentation for Generalizable Visual RL

  • Yun Zhou
  • Yuqiang Wu
  • Chunyu Tan

A fundamental challenge in visual reinforcement learning (RL) is achieving robust generalization across environments with varying visual distractions. Current RL methods struggle with generalization due to their inability to differentiate foreground and background features during augmentation,while their Q-consistency mechanisms rely on outdated actions from replay buffers that drift from the current policy.In this paper, we present PQDA, a novel framework that addresses generalization challenges in RL through two key innovations: (1) Foreground-Background Decoupled Augmentation leverages Gaussian mixture model-based segmentation to efficiently generate and cache masks in replay buffers, applying differentiated augmentation strategies to foreground and background regions, thereby enhancing data diversity while maintaining task-relevant features. (2) Policy-Aligned Q-Consistency enforces policy alignment by sampling actions from the current policy for Q-regularization, achieving faster and more stable convergence. Notably, PQDA eliminates auxiliary tasks entirely through a unified architecture that co-optimizes the encoder and RL components directly. Extensive experiments on DMControl benchmarks (including our newly proposed CVDMC benchmark) and robotic manipulation tasks demonstrate PQDA's superior generalization performance, outperforming state-of-the-art methods.

PDF Details DOI

ICML Conference 2025 Conference Paper

COSDA: Counterfactual-based Susceptibility Risk Framework for Open-Set Domain Adaptation

  • Wenxu Wang
  • Rui Zhou
  • Jing Wang
  • Yun Zhou
  • Cheng Zhu
  • Ruichun Tang
  • Bo Han
  • Nevin L. Zhang

Open-Set Domain Adaptation (OSDA) aims to transfer knowledge from the labeled source domain to the unlabeled target domain that contains unknown categories, thus facing the challenges of domain shift and unknown category recognition. While recent works have demonstrated the potential of causality for domain alignment, little exploration has been conducted on causal-inspired theoretical frameworks for OSDA. To fill this gap, we introduce the concept of Susceptibility and propose a novel C ounterfactual-based susceptibility risk framework for OSDA, termed COSDA. Specifically, COSDA consists of three novel components: (i) a Susceptibility Risk Estimator (SRE) for capturing causal information, along with comprehensive derivations of the computable theoretical upper bound, forming a risk minimization framework under the OSDA paradigm; (ii) a Contrastive Feature Alignment (CFA) module, which is theoretically proven based on mutual information to satisfy the Exogeneity assumption and facilitate cross-domain feature alignment; (iii) a Virtual Multi-unknown-categories Prototype (VMP) pseudo-labeling strategy, providing label information by measuring how similar samples are to known and multiple virtual unknown category prototypes, thereby assisting in open-set recognition and intra-class discriminative feature learning. Extensive experiments demonstrate that our approach achieves state-of-the-art performance.

Details

ICML Conference 2025 Conference Paper

ReferSplat: Referring Segmentation in 3D Gaussian Splatting

  • Shuting He
  • Guangquan Jie
  • Changshuo Wang 0001
  • Yun Zhou
  • Shuming Hu
  • Guanbin Li
  • Henghui Ding

We introduce Referring 3D Gaussian Splatting Segmentation (R3DGS), a new task that aims to segment target objects in a 3D Gaussian scene based on natural language descriptions, which often contain spatial relationships or object attributes. This task requires the model to identify newly described objects that may be occluded or not directly visible in a novel view, posing a significant challenge for 3D multi-modal understanding. Developing this capability is crucial for advancing embodied AI. To support research in this area, we construct the first R3DGS dataset, Ref-LERF. Our analysis reveals that 3D multi-modal understanding and spatial relationship modeling are key challenges for R3DGS. To address these challenges, we propose ReferSplat, a framework that explicitly models 3D Gaussian points with natural language expressions in a spatially aware paradigm. ReferSplat achieves state-of-the-art performance on both the newly proposed R3DGS task and 3D open-vocabulary segmentation benchmarks. Dataset and code are available at https: //github. com/heshuting555/ReferSplat.

Details

AAAI Conference 2025 Conference Paper

Test-Time Adaptation on Noisy Data via Model-Pruning-Based Filtering and Flatness-Aware Entropy Minimization

  • Xingzhi Zhou
  • Zhiliang Tian
  • Boyang Zhang
  • Yibo Zhang
  • Ka Chun Cheung
  • Simon See
  • Hao Yang
  • Yun Zhou

Test-time adaptation (TTA) deals with domain shifts during inference by training models based on only unlabeled test samples. Test samples may include noisy samples, which degrade domain adaptation. Existing methods rely on the model's output prediction to detect and filter noisy samples, and further search for flat regions during optimization, which makes the optimization more robust on noisy samples. However, there are two issues: (1) the output prediction tends to be inaccurate due to domain shifts, weakening noisy-sample detection; (2) current approaches for searching flat regions focus on optimization to enhance the worst case, which ignores achieving flatness by avoiding the quick changing of losses. To address these challenges, we propose a model pruning-based test-time adaptation model for noisy data streams, named MoTTA, which leverages a new proposed filtering, output difference under pruning (ODP)-based filtering, and a flatness-aware entropy minimization (FlatEM). Specifically, to reduce the impact of inaccurate output predictions, ODP-based filtering measures the output difference of a sample before and after model pruning, which works even under inaccurate output. To improve the search for flat loss surfaces, FlatEM integrates zeroth-order flatness and first-order flatness (minimize the maximal gradient normalization with a weight perturbation constrained in a small Euclidean ball) on entropy minimization. To solve these hard maximum problems, we leverage Taylor expansion to obtain approximated results for optimization. FlatEM also adopts a parameter regularization to mitigate incorrect updates from noisy samples. The experiments show our advantages in dealing with noisy data streams at TTA comparable to existing baselines.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Point-to-Spike Residual Learning for Energy-Efficient 3D Point Cloud Classification

  • Qiaoyun Wu
  • Quanxiao Zhang
  • Chunyu Tan
  • Yun Zhou
  • Changyin Sun

Spiking neural networks (SNNs) have revolutionized neural learning and are making remarkable strides in image analysis and robot control tasks with ultra-low power consumption advantages. Inspired by this success, we investigate the application of spiking neural networks to 3D point cloud processing. We present a point-to-spike residual learning network for point cloud classification, which operates on points with binary spikes rather than floating-point numbers. Specifically, we first design a spatial-aware kernel point spiking neuron to relate spiking generation to point position in 3D space. On this basis, we then design a 3D spiking residual block for effective feature learning based on spike sequences. By stacking the 3D spiking residual blocks, we build the point-to-spike residual classification network, which achieves low computation cost and low accuracy loss on two benchmark datasets, ModelNet40 and ScanObjectNN. Moreover, the classifier strikes a good balance between classification accuracy and biological characteristics, allowing us to explore the deployment of 3D processing to neuromorphic chips for developing energy-efficient 3D robotic perception systems.

PDF Details DOI

JBHI Journal 2023 Journal Article

A Two-Branch Neural Network for Short-Axis PET Image Quality Enhancement

  • Minghan Fu
  • Meiyun Wang
  • Yaping Wu
  • Na Zhang
  • Yongfeng Yang
  • Haining Wang
  • Yun Zhou
  • Yue Shang

The axial field of view (FOV) is a key factor that affects the quality of PET images. Due to hardware FOV restrictions, conventional short-axis PET scanners with FOVs of 20 to 35 cm can acquire only low-quality PET (LQ-PET) images in fast scanning times (2–3 minutes). To overcome hardware restrictions and improve PET image quality for better clinical diagnoses, several deep learning-based algorithms have been proposed. However, these approaches use simple convolution layers with residual learning and local attention, which insufficiently extract and fuse long-range contextual information. To this end, we propose a novel two-branch network architecture with swin transformer units and graph convolution operation, namely SW-GCN. The proposed SW-GCN provides additional spatial- and channel-wise flexibility to handle different types of input information flow. Specifically, considering the high computational cost of calculating self-attention weights in full-size PET images, in our designed spatial adaptive branch, we take the self-attention mechanism within each local partition window and introduce global information interactions between nonoverlapping windows by shifting operations to prevent the aforementioned problem. In addition, the convolutional network structure considers the information in each channel equally during the feature extraction process. In our designed channel adaptive branch, we use a Watts Strogatz topology structure to connect each feature map to only its most relevant features in each graph convolutional layer, substantially reducing information redundancy. Moreover, ensemble learning is adopted in our SW-GCN for mapping distinct features from the two well-designed branches to the enhanced PET images. We carried out extensive experiments on three single-bed position scans for 386 patients. The test results demonstrate that our proposed SW-GCN approach outperforms state-of-the-art methods in both quantitative and qualitative evaluations.

Details DOI

AAAI Conference 2020 Conference Paper

Relation-Guided Spatial Attention and Temporal Refinement for Video-Based Person Re-Identification

  • Xingze Li
  • Wengang Zhou
  • Yun Zhou
  • Houqiang Li

Video-based person re-identification has received considerable attention in recent years due to its significant application in video surveillance. Compared with image-based person reidentification, video-based person re-identification is characterized by a much richer context, which raises the significance of identifying informative regions and fusing the temporal information across frames. In this paper, we propose two relation-guided modules to learn reinforced feature representations for effective re-identification. First, a relation-guided spatial attention (RGSA) module is designed to explore the discriminative regions globally. The weight at each position is determined by its feature as well as the relation features from other positions, revealing the dependence between local and global contents. Based on the adaptively weighted frame-level feature, then, a relation-guided temporal refinement (RGTR) module is proposed to further refine the feature representations across frames. The learned relation information via the RGTR module enables the individual frames to complement each other in an aggregation manner, leading to robust video-level feature representations. Extensive experiments on four prevalent benchmarks verify the state-of-theart performance of the proposed method.

PDF Details

AAAI Conference 2020 Conference Paper

Spatial-Temporal Multi-Cue Network for Continuous Sign Language Recognition

  • Hao Zhou
  • Wengang Zhou
  • Yun Zhou
  • Houqiang Li

Despite the recent success of deep learning in continuous sign language recognition (CSLR), deep models typically focus on the most discriminative features, ignoring other potentially non-trivial and informative contents. Such characteristic heavily constrains their capability to learn implicit visual grammars behind the collaboration of different visual cues (i, e. , hand shape, facial expression and body posture). By injecting multi-cue learning into neural network design, we propose a spatial-temporal multi-cue (STMC) network to solve the vision-based sequence learning problem. Our STMC network consists of a spatial multi-cue (SMC) module and a temporal multi-cue (TMC) module. The SMC module is dedicated to spatial representation and explicitly decomposes visual features of different cues with the aid of a self-contained pose estimation branch. The TMC module models temporal correlations along two parallel paths, i. e. , intra-cue and intercue, which aims to preserve the uniqueness and explore the collaboration of multiple cues. Finally, we design a joint optimization strategy to achieve the end-to-end sequence learning of the STMC network. To validate the effectiveness, we perform experiments on three large-scale CSLR benchmarks: PHOENIX-2014, CSL and PHOENIX-2014-T. Experimental results demonstrate that the proposed method achieves new state-of-the-art performance on all three benchmarks.

PDF Details

YNICL Journal 2019 Journal Article

The effect of ApoE ε4 on longitudinal brain region-specific glucose metabolism in patients with mild cognitive impairment: a FDG-PET study

  • Manish D. Paranjpe
  • Xueqi Chen
  • Min Liu
  • Ishan Paranjpe
  • Jeffrey P. Leal
  • Rongfu Wang
  • Martin G. Pomper
  • Dean F. Wong

While the ApoE ε4 allele is a known risk factor for mild cognitive impairment (MCI) and Alzheimer's disease, brain region specific effects remain elusive. In this study, we investigate whether the ApoE ε4 allele exhibits brain region specific effects in longitudinal glucose uptake among patients with MCI from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Preprocessed FDG PET images, MRIs, and demographic information were downloaded from the ADNI database. An iterative reblurred Van Cittertiteration method was used for partial volume correction (PVC) on all PET images. Structural MRIs were used for PET spatial normalization and region of interest (ROI) definition in standard space. Longitudinal changes in ROI FDG standardized uptake value ratio (SUVR) relative to cerebellum in 24 ApoE ε4 carriers and 24 age-matched ApoE ε4 non-carriers were measured for up to 84-months (median 72 months, SD = 11.2 months) and compared using a generalized linear mixed effects model controlling for gender, education, baseline age, and follow-up period. Additionally, voxelwise analysis was performed by implementing a paired t-test comparing matched baseline and 72 month FDG SUVR images in ApoE carriers and non-carriers separately. Results with PVC were compared with ones from non-PVC based analysis. After applying PVC, the superior fontal, parietal, lateral temporal, medial temporal, caudate, thalamus, and post-cingulate, and amygdala regions had greater longitudinal decreases in FDG uptake in ApoE ε4 carriers with MCI compared to non-carriers with MCI. Similar forebrain and limbic clusters were found through voxelwise analysis. Compared to the PVC based analysis, fewer significant ApoE-associated regions and clusters were found in the non-PVC based PET analysis. Our findings suggest that the ApoE ε4 genotype is associated with a longitudinal decline in glucose uptake in 8 forebrain and limbic brain regions in the context of MCI. In conclusion, this 84-months longitudinal FDG PET study demonstrates a novel ApoE ε4-associated brain-region specific glucose metabolism pattern in patients with MCI. Partial volume correction improved FDG PET quantification.

Details DOI

IJCAI Conference 2017 Conference Paper

Adversarial Generation of Real-time Feedback with Neural Networks for Simulation-based Training

  • Xingjun Ma
  • Sudanthi Wijewickrema
  • Shuo Zhou
  • Yun Zhou
  • Zakaria Mhammedi
  • Stephen O'Leary
  • James Bailey

Simulation-based training (SBT) is gaining popularity as a low-cost and convenient training technique in a vast range of applications. However, for a SBT platform to be fully utilized as an effective training tool, it is essential that feedback on performance is provided automatically in real-time during training. It is the aim of this paper to develop an efficient and effective feedback generation method for the provision of real-time feedback in SBT. Existing methods either have low effectiveness in improving novice skills or suffer from low efficiency, resulting in their inability to be used in real-time. In this paper, we propose a neural network based method to generate feedback using the adversarial technique. The proposed method utilizes a bounded adversarial update to minimize a L1 regularized loss via back-propagation. We empirically show that the proposed method can be used to generate simple, yet effective feedback. Also, it was observed to have high effectiveness and efficiency when compared to existing methods, thus making it a promising option for real-time feedback generation in SBT.

PDF Details

AAAI Conference 2015 Conference Paper

Tackling Mental Health by Integrating Unobtrusive Multimodal Sensing

  • Dawei Zhou
  • Jiebo Luo
  • Vincent Silenzio
  • Yun Zhou
  • Jile Hu
  • Glenn Currier
  • Henry Kautz

Mental illness is becoming a major plague in modern societies and poses challenges to the capacity of current public health systems worldwide. With the widespread adoption of social media and mobile devices, and rapid advances in artificial intelligence, a unique opportunity arises for tackling mental health problems. In this study, we investigate how users’ online social activities and physiological signals detected through ubiquitous sensors can be utilized in realistic scenarios for monitoring their mental health states. First, we extract a suite of multimodal time-series signals using modern computer vision and signal processing techniques, from recruited participants while they are immersed in online social media that elicit emotions and emotion transitions. Next, we use machine learning techniques to build a model that establishes the connection between mental states and the extracted multimodal signals. Finally, we validate the effectiveness of our approach using two groups of recruited subjects.

PDF Details

YNIMG Journal 2013 Journal Article

Characterization of [11C]RO5013853, a novel PET tracer for the glycine transporter type 1 (GlyT1) in humans

  • Dean F. Wong
  • Susanne Ostrowitzki
  • Yun Zhou
  • Vanessa Raymont
  • Carsten Hofmann
  • Edilio Borroni
  • Anil Kumar
  • Nikhat Parkar

We characterize a novel radioligand for the glycine transporter type 1 (GlyT1), [11 C]RO5013853, in humans. Ten healthy male volunteers, 23–60years of age, were enrolled in this PET study; seven subjects participated in the evaluation of test-retest reliability and three subjects in whole body dosimetry. Subjects were administered intravenous bolus injections of approximately 1100MBq (30mCi) [11 C]RO5013853 with a high specific activity of about 481GBq (13Ci)/μmol. Standard compartmental model analysis with arterial plasma input function, and an alternative noninvasive analysis method which was evaluated and validated by occupancy studies in both baboons and humans, were performed. Mean parameter estimates of the volumes of distribution (VT) obtained by a 2-tissue 5-parameter model were higher in the cerebellum, pons, and thalamus (1. 99 to 2. 59mL/mL), and lower in the putamen, caudate, and cortical areas (0. 86 to 1. 13mL/mL), with estimates showing less than 10% difference between test and retest scans. Tracer retention was effectively blocked by the specific glycine reuptake inhibitor (GRI), bitopertin (RG1678). [11 C]RO5013853 was safe and well tolerated. Human dosimetry studies showed that the effective dose was approximately 0. 0033mSv/MBq, with the liver receiving the highest absorbed dose. In conclusion, quantitative dynamic PET of the human brain after intravenous injection of [11C]RO5013853 attains reliable measurements of GlyT1 binding in accordance with the expected transporter distribution in the human brain. [11C]RO5013853 is a radioligand suitable for further clinical PET studies. Full characterization of a novel radiotracer for GlyT1 in humans is provided. The tracer has subsequently been used to assess receptor occupancy in healthy volunteers and to estimate occupancy at doses associated with best efficacy in a clinical trial with schizophrenic patients with predominantly negative symptoms.

Details DOI

YNIMG Journal 2013 Journal Article

Pre-clinical characterization of [11C]R05013853 as a novel radiotracer for imaging of the glycine transporter type 1 by positron emission tomography

  • Edilio Borroni
  • Yun Zhou
  • Susanne Ostrowitzki
  • Daniela Alberati
  • Anil Kumar
  • Dominik Hainzl
  • Thomas Hartung
  • John Hilton

A specific positron emission tomography (PET) radiotracer for the glycine transporter type 1 (GlyT1) would constitute an imaging biomarker to investigate the distribution of GlyT1 in normal individuals and those with neuropsychiatric disorders. In addition it could demonstrate the ability of a novel drug to reach its target in the brain and enable receptor occupancy studies, thus facilitating drug development. In this article we describe the evaluation in non-human primates of two candidate PET radiotracers ([11C]RO5013852 and [11C]RO5013853) previously characterized in the rat. Both radiotracers showed acceptable uptake in the baboon brain and heterogeneous distribution consistent with that reported for GlyT1. In vivo blockade studies with two specific glycine reuptake inhibitors (GRIs), RO5013853 and bitopertin (RG1678, reduced uptake of both tracers to homogenous levels across brain regions and demonstrated specificity of the signal. [11C]RO5013853 showed a larger specific signal and slightly higher brain uptake and was therefore selected for further characterization. Quantitative compartmental analysis of PET data showed that the 2-tissue compartment model with 5 parameters was the most appropriate to describe the kinetics of [11C]RO5013853. Two additional methods were used: a) the Logan graphical analysis using plasma input and, b) a linear parametric imaging approach with the 2-tissue compartmental model. These produced VT estimates of comparable magnitude, namely, pons, thalamus and cerebellum>caudate, putamen and cortical regions. High resolution autoradiography with tritiated RO5013853 was used to confirm the binding pattern observed by PET. In vivo metabolism studies in the baboon demonstrated the formation of a single, radiolabeled metabolite more polar than the parent compound. Finally, [11C]RO5013853 was used to quantify the degree of cerebral GlyT1 occupancy observed in the baboon following oral administration of bitopertin, a selective GRI presently in Phase III clinical trial. Plasma concentrations of approximately 150–300ng/mL were estimated to produce 50% GlyT1 occupancy in the thalamus, the cerebellum and the pons. [11C]RO5013853 is a promising radiotracer for in vivo imaging of the GlyT1. It can be easily radiolabeled, exhibits moderate metabolism, displays a good specific signal, and is suitable for receptor occupancy studies of therapeutic compounds that target the GlyT1. The successful characterization of [11C]RO5013853 in healthy volunteers is presented in this NeuroImage issue (Wong et al. , 2013).

Details DOI

YNIMG Journal 2010 Journal Article

Multi-graphical analysis of dynamic PET

  • Yun Zhou
  • Weiguo Ye
  • James R. Brašić
  • Dean F. Wong

In quantitative dynamic PET studies, graphical analysis methods including the Gjedde–Patlak plot, the Logan plot, and the relative equilibrium-based graphical plot (RE plot) (Zhou Y. , Ye W. , Brašić J. R. , Crabb A. H. , Hilton J. , Wong D. F. 2009b. A consistent and efficient graphical analysis method to improve the quantification of reversible tracer binding in radioligand receptor dynamic PET studies. Neuroimage 44(3): 661–670) are based on the theory of a compartmental model with assumptions on tissue tracer kinetics. If those assumptions are violated, then the resulting estimates may be biased. In this study, a multi-graphical analysis method was developed to characterize the non-relative equilibrium effects on the estimates of total distribution volume (DVT) from the RE plot. A novel bi-graphical analysis method using the RE plot with the Gjedde–Patlak plot (RE-GP plots) was proposed to estimate DVT for the quantification of reversible tracer kinetics that may not be at relative equilibrium states during PET study period. The RE-GP plots and the Logan plot were evaluated by 19 [11C]WIN35, 428 and 10 [11C]MDL100, 907 normal human dynamic PET studies with brain tissue tracer kinetics measured at both region of interest (ROI) and pixel levels. A 2-tissue compartment model (2TCM) was used to fit ROI time activity curves (TACs). By applying multi-graphical plots to the 2TCM fitted ROI TACs which were considered as the noise-free tracer kinetics, the estimates of DVT from the RE-GP plots, the Logan plot, and the 2TCM fitting were equal to each other. For the measured ROI TACs, there was no significant difference between the estimates of the DVT from the RE-GP plots and those from 2TCM fitting (p =0. 77), but the estimates of the DVT from the Logan plot were significantly (p <0. 001) lower, 2. 3% on average, than those from 2TCM fitting. There was a highly linear correlation between the ROI DVT from the parametric images (Y) and those from the ROI kinetics (X) by using the RE-GP plots (Y =1. 01X +0. 23, R 2 =0. 99). For the Logan plot, the ROI estimates from the parametric images were 13% to 83% lower than those from ROI kinetics. The computational time for generating parametric images was reduced by 69% on average by the RE-GP plots in contrast to the Logan plot. In conclusion, the bi-graphical analysis method using the RE-GP plots was a reliable, robust and computationally efficient kinetic modeling approach to improve the quantification of dynamic PET.

Details DOI

YNIMG Journal 2009 Journal Article

A consistent and efficient graphical analysis method to improve the quantification of reversible tracer binding in radioligand receptor dynamic PET studies

  • Yun Zhou
  • Weiguo Ye
  • James R. Brašić
  • Andrew H. Crabb
  • John Hilton
  • Dean F. Wong

The widely used Logan plot in radioligand receptor dynamic PET studies produces marked noise-induced negative biases in the estimates of total distribution volume (DVT) and binding potential (BP). To avoid the inconsistencies in the estimates from the Logan plot, a new graphical analysis method was proposed and characterized in this study. The new plot with plasma input and with reference tissue input was first derived to estimate DVT and BP. A condition was provided to ensure that the estimate from the new plot equals DVT or BP. It was demonstrated theoretically that 1) the statistical expectations of the estimates from the new plot with given input are independent of the noise of the target tissue concentration measured by PET; and 2) the estimates from the time activity curves of regions of interest are identical to those from the parametric images for the new plot. The theoretical results of the new plot were also confirmed by computer simulations and fifty-five human [11C]raclopride dynamic PET studies. By contrast, the marked noise-induced underestimation in the DVT and BP images and noise-induced negative bias in the estimates from the Logan plot were demonstrated by the same data sets used for the new plot. The computational time for generating DVT or BP images in the human studies was reduced by 80% on average by the new plot in contrast to the Logan plot. In conclusion, the new plot is a consistent and computationally efficient graphical analysis method to improve the quantification of reversible tracer binding in radioligand receptor dynamic PET studies.

Details DOI

YNIMG Journal 2007 Journal Article

Impulsivity and chronic stress are associated with amphetamine-induced striatal dopamine release

  • Lynn M. Oswald
  • Dean F. Wong
  • Yun Zhou
  • Anil Kumar
  • James Brasic
  • Mohab Alexander
  • Weiguo Ye
  • Hiroto Kuwabara

A challenging question that continues to plague the field of addiction is why some individuals are more vulnerable for substance use disorders than others. Several important risk factors for substance abuse have been identified in clinical studies, including trait impulsivity and environmental stress. However, the neurobiological mechanisms that underlie the relationships remain poorly understood. The purpose of this study was to examine associations among impulsivity, stress, and striatal dopamine (DA) responses to amphetamine (AMPH) in humans. Forty healthy M, F adults, ages 18–29 years, completed self-report measures of trait impulsivity, life events stress, and perceived stress. Subjects subsequently underwent two consecutive 90-min positron emission tomography (PET) studies with high specific activity [11C]raclopride. The first scan was preceded by an intravenous injection of saline; the second was preceded by 0. 3 mg/kg AMPH. Findings showed that high impulsivity was associated with blunted right ventral striatal DA release. However, effects were modified by a significant interaction with life events stress. Dopamine release was greater in low vs. high impulsivity subjects under conditions of low or moderate stress. Under conditions of high stress, both groups had low DA release. Subjects with high impulsivity reported more pleasant effects with AMPH than subjects with low impulsivity. In contrast, stress was negatively associated with pleasant drug effects. No associations were observed between impulsivity or stress and cortisol responses to AMPH. The findings are consistent with notions that blunted DA responses represent an endophenotype for substance use disorders.

Details DOI

YNIMG Journal 2007 Journal Article

Using a reference tissue model with spatial constraint to quantify [11C]Pittsburgh compound B PET for early diagnosis of Alzheimer's disease

  • Yun Zhou
  • Susan M. Resnick
  • Weiguo Ye
  • Hong Fan
  • Daniel P. Holt
  • William E. Klunk
  • Chester A. Mathis
  • Robert Dannals

Introduction: Reference tissue model (RTM) is a compartmental modeling approach that uses reference tissue time activity curve (TAC) as input for quantification of ligand–receptor dynamic PET without blood sampling. There are limitations in applying the RTM for kinetic analysis of PET studies using [11C]Pittsburgh compound B ([11C]PIB). For region of interest (ROI) based kinetic modeling, the low specific binding of [11C]PIB in a target ROI can result in a high linear relationship between the output and input. This condition may result in amplification of errors in estimates using RTM. For pixel-wise quantification, due to the high noise level of pixel kinetics, the parametric images generated by RTM with conventional linear or nonlinear regression may be too noisy for use in clinical studies. Methods: We applied RTM with parameter coupling and a simultaneous fitting method as a spatial constraint for ROI kinetic analysis. Three RTMs with parameter coupling were derived from a classical compartment model with plasma input: an RTM of 4 parameters (R 1, k′2R, k 4, BP) (RTM4P); an RTM of 5 parameters (R 1, k 2R, NS, k 6, BP) (RTM5P); and a simplified RTM (SRTM) of 3 parameters (R 1, k′2R, BP) (RTM3P). The parameter sets [k′2R, k 4], [k 2R, NS, k 6], and k′2R are coupled among ROIs for RTM4P, RTM5P, and RTM3P, respectively. A linear regression with spatial constraint (LRSC) algorithm was applied to the SRTM for parametric imaging. Logan plots were used to estimate the distribution volume ratio (DVR) (=1+BP (binding potential)) in ROI and pixel levels. Ninety-minute [11C]PIB dynamic PET was performed in 28 controls and 6 individuals with mild cognitive impairment (MCI) on a GE Advance scanner. ROIs of cerebellum (reference tissue) and 15 other regions were defined on coregistered MRIs. Results: The coefficients of variation of DVR estimates from RTM3P obtained by the simultaneous fitting method were lower by 77–89% (in striatum, frontal, occipital, parietal, and cingulate cortex) as compared to that by conventional single ROI TAC fitting method. There were no significant differences in both TAC fitting and DVR estimates between the RTM3P and the RTM4P or RTM5P. The DVR in striatum, lateral temporal, frontal and cingulate cortex for MCI group was 25% to 38% higher compared to the control group (p ≤0. 05), even in this group of individuals with generally low PIB retention. The DVR images generated by the SRTM with LRSC algorithm had high linear correlations with those from the Logan plot (R 2 =0. 99). Conclusion: In conclusion, the RTM3P with simultaneous fitting method is shown to be a robust compartmental modeling approach that may be useful in [11C]PIB PET studies to detect early markers of Alzheimer's disease where specific ROIs have been hypothesized. In addition, the SRTM with LRSC algorithm may be useful in generating R 1 and DVR images for pixel-wise quantification of [11C]PIB dynamic PET.

Details DOI

YNIMG Journal 2006 Journal Article

An extended simplified reference tissue model for the quantification of dynamic PET with amphetamine challenge

  • Yun Zhou
  • Ming-Kai Chen
  • Christopher J. Endres
  • Weiguo Ye
  • James R. Brašić
  • Mohab Alexander
  • Andrew H. Crabb
  • Tomás R. Guilarte

Background: Equilibrium analysis to quantify dynamic positron emission tomography (PET) with bolus followed by continuous tracer infusion and acute amphetamine challenge assumes that all tissue kinetics attain steady states during pre- and post-challenge phases. Violations of this assumption may result in unreliable estimation of the amphetamine-induced percent change in the binding potential (ΔBP%). Method: We derived an extended simplified reference tissue model (ESRTM) for modeling tracer kinetics in the pre- and post-challenge phases. Ninety-minute [11C]raclopride PET studies with bolus injection followed by continuous tracer infusion were performed on 18 monkeys and 2 baboons. Forty minutes after the bolus injection, a single acute intravenous amphetamine administration was given of 2. 0 mg/kg to monkeys and of 0. 05, 0. 1, 0. 5, and 1. 5 mg/kg to baboons. Computer simulations further evaluated and characterized the ESRTM. Results: In monkey studies, the ΔBP% estimated by the ESRTM was 32±11, whereas, the ΔBP% obtained using the equilibrium methods was 32% to 81% lower. In baboon studies, the ΔBP% values estimated with the ESRTM showed a linear relationship between the ΔBP% and the natural logarithm of amphetamine dose (R 2 =0. 96), where the ΔBP%=10. 67Ln(dose)+33. 79 (0. 05≤dose in mg/kg≤1. 5). At 1. 5 mg/kg amphetamine, the ΔBP% estimates from equilibrium methods were 18% to 40% lower than those estimated by the ESRTM. Results showed that the nonsteady state of tracer kinetics produced an underestimation of the ΔBP% from the equilibrium analysis. The accuracy of the ΔBP% estimates from the equilibrium analysis was significantly improved by the ESRTM. The ΔBP% estimated by the ESRTM in the study was consistent with that from previous [11C]raclopride PET with amphetamine challenge. Conclusion: In conclusion, the ESRTM is a robust kinetic modeling approach and is proposed for the quantification of dynamic PET with acute amphetamine stimulation.

Details DOI

YNIMG Journal 2003 Journal Article

Linear regression with spatial constraint to generate parametric images of ligand-receptor dynamic PET studies with a simplified reference tissue model

  • Yun Zhou
  • Christopher J Endres
  • James Robert Brašić
  • Sung-Cheng Huang
  • Dean F Wong

For the quantitative analysis of ligand-receptor dynamic positron emission tomography (PET) studies, it is often desirable to apply reference tissue methods that eliminate the need for arterial blood sampling. A common technique is to apply a simplified reference tissue model (SRTM). Applications of this method are generally based on an analytical solution of the SRTM equation with parameters estimated by nonlinear regression. In this study, we derive, based on the same assumptions used to derive the SRTM, a new set of operational equations of integral form with parameters directly estimated by conventional weighted linear regression (WLR). In addition, a linear regression with spatial constraint (LRSC) algorithm is developed for parametric imaging to reduce the effects of high noise levels in pixel time activity curves that are typical of PET dynamic data. For comparison, conventional weighted nonlinear regression with the Marquardt algorithm (WNLRM) and nonlinear ridge regression with spatial constraint (NLRRSC) were also implemented using the nonlinear analytical solution of the SRTM equation. In contrast to the other three methods, LRSC reduces the percent root mean square error of the estimated parameters, especially at higher noise levels. For estimation of binding potential (BP), WLR and LRSC show similar variance even at high noise levels, but LRSC yields a smaller bias. Results from human studies demonstrate that LRSC produces high-quality parametric images. The variance of R 1 and k 2 images generated by WLR, WNLRM, and NLRRSC can be decreased 30%–60% by using LRSC. The quality of the BP images generated by WLR and LRSC is visually comparable, and the variance of BP images generated by WNLRM can be reduced 10%–40% by WLR or LRSC. The BP estimates obtained using WLR are 3%–5% lower than those estimated by LRSC. We conclude that the new linear equations yield a reliable, computationally efficient, and robust LRSC algorithm to generate parametric images of ligand-receptor dynamic PET studies.

Details DOI

YNIMG Journal 2002 Journal Article

Improved Parametric Image Generation Using Spatial-Temporal Analysis of Dynamic PET Studies

  • Yun Zhou
  • Sung-Cheng Huang
  • Marvin Bergsneider
  • Dean F. Wong

The value of parametric images that represent both spatial distribution and quantification of the physiological parameters of tracer kinetics has long been recognized. However, the inherent high noise level of pixel kinetics of dynamic PET makes it unsuitable to generate parametric images of the microparameters of tracer kinetic model by conventional weighted nonlinear least squares (WNLS) fitting. Based on the concept that both spatial and temporal information should be integrated to improve parametric image quality, a nonlinear ridge regression with spatial constraint (NLRRSC) parametric imaging algorithm was proposed in this study. For NLRRSC, a term that penalizes local spatial variation of parameters was added to the cost function of WNLS fitting. The initial estimates and spatial constraint were estimated by component representation model (CRM) with cluster analysis. A hierarchical cluster with average linkage method was used to extract components. The ridge parameter was determined by linear ridge regression theory at each iteration, and a modified Gauss–Newton algorithm was used for minimizing the cost function. Results from a computer simulation showed that the percent mean square error of estimates obtained by NLRRSC can be decreased by 60–80% compared to that of WNLS. The parametric images estimated by NLRRSC are significantly better than the ones generated by WNLS. A highly correlated linear relationship was found between the ROI values calculated from the microparametric images generated by NLRRSC and estimates from ROI kinetic fitting. NLRRSC provided a reliable estimate of glucose metabolite uptake rate with a comparable image quality compared to Patlak analysis. In conclusion, NLRRSC is a reliable and robust parametric imaging algorithm for dynamic PET studies.

Details DOI
v2026.09.13