浏览全部资源
扫码关注微信
收稿日期：2025-01-17，
修回日期：2025-02-25，
录用日期：2025-03-03，
网络出版日期：2025-03-04，
移动端阅览
目的
2
阿尔茨海默症（Alzheimer's disease， AD）作为一种常见的老年性痴呆疾病，近年来已成为全球公共卫生面临的重大挑战，设计一种有效且精确的阿尔茨海默症早期诊断模型具有重要的临床意义和迫切需求。目前，阿尔茨海默症的临床诊断通常依赖于正电子发射断层扫描（Positron Emission Tomography， PET）和核磁共振成像（Magnetic Resonance Imaging， MRI）两种医学影像数据。然而，由于这两种模态间存在信息差异大，未精确配准等问题，现有的基于人工智能（Artificial Intelligence， AI）的诊断模型大多仅使用单一的MRI数据。这在一定程度上限制了多模态影像信息的充分利用和分类性能的进一步提升，制约了其临床实用性。针对上述问题，提出一种结合DenseNet和Mamba的多模态医学脑影像阿尔茨海默症早期诊断模型——DenseMamba。
方法
2
该方法以经过标准预处理流程后的PET和MRI数据为输入，预处理流程包括：颅骨剥离、配准、偏置场校正、归一化。MRI和PET级联后首先经过卷积层和激活层进行初步特征提取，提取到的特征再依次经过若干个交替的Denseblock和TransMamba模块分别进行局部和全局的特征提取，Denseblock内的密集连接结构，增强了局部特征的提取和传播，从而能够捕捉影像中的细节信息；而TransMamba模块则基于状态空间模型，高效地建模全局依赖关系，交替的Denseblock和TransMamba使得模型能够更全面地理解多模态数据信息，充分挖掘多模态数据在临床诊断任务上的潜力。最后，将提取到的特征送入分类器得到疾病预测结果。
结果
2
为验证方法的有效性，实验在公开的ADNI（Alzheimer’s Disease Neuroimaging Initiative）数据集上对其进行了评估。最终模型的准确率（accuracy）、精确度（precision）、召回率（recall）和F1值分别为92.42%、92.5%、92.42%、92.21%。DenseMamba在阿尔茨海默症分类任务中较其他算法表现优异，与现有先进的方法相比准确率提升0.42%。
结论
2
实验结果表明，与现有的基于单模态影像数据的分类方法相比，DenseMamba能够充分发挥PET和MRI数据的潜力，显著提升分类性能，为阿尔茨海默症的早期诊断提供更精准的支持。
Objective
2
Alzheimer's disease（AD） is the most common form of dementia in the elderly and has become a major global public health challenge in recent years. With the acceleration of global aging， the incidence of Alzheimer's disease continues to rise， putting tremendous pressure on medical treatment and caregiving. The main characteristic of Alzheimer's disease is the gradual decline in cognitive functions， particularly memory， language， and decision-making abilities， often accompanied by irreversible neuronal loss and brain atrophy. As the disease progresses， it becomes increasingly difficult to differentiate Alzheimer's disease from other types of dementia， making early diagnosis more challenging. Therefore， early diagnosis is crucial for timely intervention and personalized treatment plans， as it can effectively delay disease progression. Currently， clinical diagnosis of Alzheimer's disease typically relies on neuropsychological assessments， biomarker detection， and neuroimaging techniques. Among these， Positron Emission Tomography （PET） and Magnetic Resonance Imaging （MRI） are the most commonly used imaging methods to evaluate the brain structure and function of AD patients. PET provides information on brain metabolism and amyloid plaques， while MRI offers high-resolution structural images that can reveal brain atrophy and other neuroanatomical changes related to the disease. However， there are significant differences between these two imaging methods， such as resolution， contrast， and spatial alignment issues. As a result， existing artificial intelligence （AI）-based diagnostic models often rely on single-modality data， typically using only MRI images. This limits the full utilization of multimodal information and restricts further improvements in classification performance. To address this issue， this paper proposes an early diagnosis model for Alzheimer's disease based on multimodal imaging data （PET and MRI） — DenseMamba. This model combines the advantages of both imaging modalities， overcoming the limitations of traditional single-modality methods， and adopts a state-space model-based framework （Mamba） to effectively extract and fuse heterogeneous modal features， achieving more robust classification performance.
Method
2
The DenseMamba model takes multimodal PET and MRI imaging data as input. To effectively integrate these two types of imaging data， the model first performs initial feature extraction through a convolutional layer and an activation layer. The extracted features are then passed through several alternating Denseblock and TransMamba modules for further local and global feature extraction. The Denseblock module， with its dense connection design， enhances feature learning capability by facilitating feature reuse across layers， which effectively mitigates the vanishing gradient problem in deep neural networks. This dense connectivity helps capture complex patterns and dependencies within the data. In contrast， the TransMamba module focuses on capturing global feature representations， modeling long-range dependencies within the data， and effectively integrating cross-space and cross-time related information. After multiple rounds of local and global feature extraction， the final features are processed by a classification head （composed of fully connected layers and a Softmax output layer） to perform the multi-class classification task. The final output is the predicted category for each input image， indicating whether the patient is classified as healthy control （CN）， mild cognitive impairment （MCI）， or Alzheimer's disease （AD）.
Results
2
To validate the effectiveness of the DenseMamba model， extensive experiments were conducted using the publicly available Alzheimer's Disease Neuroimaging Initiative （ADNI） dataset. This dataset includes multimodal imaging data （PET and MRI）， and the imaging data have been preprocessed to ensure alignment between PET and MRI images. Experimental results show that the DenseMamba model significantly outperforms traditional single-modality methods， achieving an accuracy of 92.42%， precision of 92.5%， recall of 92.42%， and an F1 score of 92.21%.
Conclusion
2
DenseMamba provides an efficient and accurate solution for the early diagnosis of Alzheimer's disease. By combining the complementary advantages of PET and MRI， DenseMamba overcomes some of the limitations of traditional single-modality methods， such as insufficient feature representation and data misalignment issues. Experimental results demonstrate that DenseMamba performs excellently on the ADNI dataset， significantly improving classification performance and far surpassing existing single-modality imaging classification methods.
Aderghal K ， Benois-Pineau J and Afdel K . 2017 . Classification of smri for alzheimer's disease diagnosis with CNN： single siamese networks with 2D+？ approach and fusion on ADNI // Proceedings of the 2017 ACM on International Conference on Multimedia Retrieval . Bucharest， Romania ： Association for Computing Machinery： 494 - 498 ［ DOI： 10.1145/3078971.3079010 http://dx.doi.org/10.1145/3078971.3079010 ］
Ahmad W M ， Wonsik J ， Kwanseok O ， Jee S Y ， Kun H L and Heung-Il S . 2023 . Estimating explainable alzheimer’s disease likelihood map via clinically-guided prototype learning . NeuroImage ， 273 ： 120073 ［ DOI： 10.1016/j.neuroimage.2023.120073 http://dx.doi.org/10.1016/j.neuroimage.2023.120073 ］
Cui Z ， Gao Z ， Leng J ， Zhang T ， Quan P and Zhao W . 2019 . Alzheimer's disease diagnosis using enhanced inception network based on brain magnetic resonance image // 2019 IEEE International Conference on Bioinformatics and Biomedicine （BIBM） . IEEE ： 2324 - 2330 ［ DOI： 10.1109/BIBM47256.2019.8983046 http://dx.doi.org/10.1109/BIBM47256.2019.8983046 ］
Billones C D ， Demetria O J L D ， Hostallero D E D and Naval P C . 2016 . DemNet： A convolutional neural network for the detection of alzheimer's disease and mild cognitive impairment // 2016 IEEE Region 10 Conference （TENCON） . IEEE ： 3724 - 3727 ［ DOI： 10.1109/TENCON.2016.7848755 http://dx.doi.org/10.1109/TENCON.2016.7848755 ］
Saratxaga C L ， Moya I ， Picón A ， Acosta M ， Moreno-Fernandez-de-Leceta A ， Garrote E and Bereciartúa-Pérez A . 2021 . MRI deep learning-based solution for alzheimer’s disease prediction . Journal of Personalized Medicine ， 11 ［ DOI： 10.3390/jpm11090902 http://dx.doi.org/10.3390/jpm11090902 ］
Chen H ， Guo H R ， Xing L Q ， Chen D ， Yuan T ， Zhang Y P and Zhang X D . 2023 . Multimodal predictive classification of alzheimer's disease based on attention-combined fusion network： Integrated Neuroimaging Modalities And Medical Examination data . IET Image Processing ， 17 （ 11 ）： 3153 - 3164 ［ DOI： 10.1049/ipr2.12841 http://dx.doi.org/10.1049/ipr2.12841 ］
Deng S ， HE X ， Qing L B ， Chen H G ， Teng Q Z . 2022 . Weakly supervised fine-grained classification method of alzheimer’s disease based on improved visual geometry group network . Journal of Computer Applications ， 42 （ 1 ）： 302 - 309
邓爽 ， 何小海 ， 卿粼波 ， 陈洪刚 ， 滕奇志 . 2022 . 基于改进VGG网络的弱监督细粒度阿尔茨海默症分类方法 . 计算机应用 ， 42 （ 1 ）： 302 - 309 ［ DOI： 10.11772/j.issn.1001-9081.2021020258 http://dx.doi.org/10.11772/j.issn.1001-9081.2021020258 ］
Huang G， Liu Z， L. Van Der Maaten and K. Q. Weinberger . 2017 . Densely Connected Convolutional Networks //2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. IEEE ： 2261 - 2269 ［ DOI： 10.1109/CVPR.2017.243 http://dx.doi.org/10.1109/CVPR.2017.243 ］
Hu Z Y ， Wang Y H and Xiao L . 2025 . Alzheimer’s disease diagnosis by 3D-SEConvNeXt . Journal of Big Data ， 12 （ 15 ） ［DOI： 10.1186/s40537-025-01088-8］
Jin D ， Zhou B ， Han Y ， Ren J ， Han T ， Liu B ， Lu J ， Song C ， Wang P ， Wang D ， Xu J ， Yang Z ， Yao H ， Yu C ， Zhao K ， Wintermark ， M ， Zuo N ， Zhang X ， Zhou Y ， Zhang X ， Jiang T ， Wang Q and Liu Y . 2020 . Generalizable ， reproducible， and neuroscientifically interpretable imaging biomarkers for alzheimer's disease. Adv Sci （Weinh）， 7 （ 14 ）： 1992 - 2003 ［ DOI： 10.1002/advs.202000675 http://dx.doi.org/10.1002/advs.202000675 ］
Kong Z K ， Zhang M Y ， Zhu W J ， Yi Y ， Wang T and Zhang B C . 2022 . Multi-modal data alzheimer’s disease detection based on 3D convolution . Biomedical Signal Processing and Control ， 75 ： 103565 ［ DOI： 10.1016/j.bspc.2022.103565 http://dx.doi.org/10.1016/j.bspc.2022.103565 ］
Kruthika K. R. ， Rajeswari ， Maheshappa H. D. . 2019 . CBIR system using capsule networks and 3D CNN for alzheimer's disease diagnosis . Informatics In Medicine Unlocked ， 14 ： 59 - 68 ［ DOI： 10.1016/j.imu.2018.12.001 http://dx.doi.org/10.1016/j.imu.2018.12.001 ］
Lian C ， Liu M ， Pan Y and Shen D . 2022 . Attention-guided hybrid network for dementia diagnosis with structural MR images . IEEE Transactions on Cybernetics ， 52 （ 4 ）： 1992 - 2003 ［ DOI： 10.1109/TCYB.2020.3005859 http://dx.doi.org/10.1109/TCYB.2020.3005859 ］
Liao W B ， Zhu Y H ， Wang X Y ， Pan C W ， Wang Y S and Ma L T . 2024 . LightM-UNet： mamba assists in lightweight unet for medical image segmentation［EB\OL］.［2024.03 . 11 ］. https：//arxiv.org/abs/2403.05246 https://arxiv.org/abs/2403.05246
Liu S ， Chhavi Y ， Carlos F G and Narges R . 2019 . On the design of convolutional neural networks for automatic detection of alzheimer's disease . Proceedings of Machine Learning Research ， 116 ： 184 - 201
Liu M ， Zhang J ， Adeli . E and Shen D . 2019 . Joint classification and regression via deep multi-task multi-channel learning for alzheimer's disease diagnosis. IEEE Transactions on Biomedical Engineering ， 66 （ 5 ）： 1195 - 1206 ［ DOI： 10.1109/TBME.2018.2869989 http://dx.doi.org/10.1109/TBME.2018.2869989 ］
Li Y M ， Tang J H ， Ding W W ， Luo J ， Ahmad N and Kumar R . 2024 . An assisted diagnosis of alzheimer’s disease incorporating attention mechanisms Med-3D transfer modeling . Computers， Materials and Continua ， 78 （ 1 ）： 713 - 733 ［ DOI： 10.32604/cmc.2023.046872 http://dx.doi.org/10.32604/cmc.2023.046872 ］
Liu X ， Li W M ， Miao S ， Liu F Y ， Han K and Bezabih T T . 2024 . Hammf： hierarchical attention-based multi-task and multi-modal fusion model for computer-aided diagnosis of alzheimer’s disease . Computers In Biology And Medicine ， 176 ： 108564 ［ DOI： 10.1016/j.compbiomed.2024.108564 http://dx.doi.org/10.1016/j.compbiomed.2024.108564 ］
Pan Q ， Ding K and Chen D . 2021 . Multi-Classification Prediction of Alzheimer’s Disease based on Fusing Multi-modal Features // 2021 IEEE International Conference on Data Mining （ICDM） . IEEE ： 1270 - 1275 ［ DOI： 10.1109/ICDM51629.2021.00156 http://dx.doi.org/10.1109/ICDM51629.2021.00156 ］
Puente-Castro ， Alejandro Fernandez-Blanco ， Enrique Pazos ， Alejandro Munteanu and Cristian R. . 2020 . Automatic assessment of alzheimer’s disease diagnosis based on deep learning techniques . Computers In Biology And Medicine ， 120 ： 103764 ［ DOI： 10.1016/j.compbiomed.2020.103764 http://dx.doi.org/10.1016/j.compbiomed.2020.103764 ］
Saleem T. J. ， Zahra S. R. ， Wu F. ， Alwakeel A. ， Alwakeel M. ， Jeribi F . and Hijji， M .. 2022 . Deep Learning-based diagnosis of alzheimer's disease. J Pers Med ， 12 （ 5 ） ［ DOI： 10.3390/jpm12050815 http://dx.doi.org/10.3390/jpm12050815 ］
Shi J ， Wang L L ， Wang S S ， Chen Y X ， Wang Q ， Wei D M ， Liang S J ， Peng J L ， Yi J J ， Liu S F ， Ni D ， Wang M L ， Zhang D Q ， Shen D G . 2020 . Applications of deep learning in medical imaging： a survey . Journal of Image and Graphics ， 25 （ 10 ）： 1953 - 1981
施俊 ， 汪琳琳 ， 王珊珊 ， 陈艳霞 ， 王乾 ， 魏冬铭 ， 梁淑君 ， 彭佳林 ， 易佳锦 ， 刘盛锋 ， 倪东 ， 王明亮 ， 张道强 ， 沈定刚 . 2020 . 深度学习在医学影像中的应用综述 . 中国图象图形学报 ， 25 （ 10 ）： 1953 - 1981 ［ DOI： 10.11834/jig.200255 http://dx.doi.org/10.11834/jig.200255 ］
Tang Y ， Xiong X ， Tong G ， Yang Y and Zhang H . 2024 . Multimodal diagnosis model of alzheimer’s disease based on improved transformer . BioMedical Engineering OnLine ， 23 （ 8 ） ［DOI： 10.1186/s12938-024-01204-4］
Venugopalan ， Janani Tong ， Li Hassanzadeh ， Hamid Reza Wang and May D. . 2021 . Multimodal deep learning models for early detection of alzheimer’s disease stage . Scientific Reports ， 11 （ 1 ）： 3254 ［ DOI： 10.1038/s41598-020-74399-w http://dx.doi.org/10.1038/s41598-020-74399-w ］
Wang H F ， Shen Y Y ， Wang S Q ， Xiao T F ， Deng L M ， Wang X Y and Zhao X Y . 2019 . Ensemble of 3D densely connected convolutional network for diagnosis of mild cognitive impairment and alzheimer’s disease . Neurocomputing ， 333 ： 145 - 156 ［ DOI： 10.1016/j.neucom.2018.12.018 http://dx.doi.org/10.1016/j.neucom.2018.12.018 ］
Wu R K ， Liu Y H ， Liang P C and Chang Q . 2024 . Ultralight vm-unet： parallel vision mamba significantly reduces parameters for skin lesion segmentation ［EB/OL］.［2024. 04 . 09 ］. https://arxiv.org/abs/2403.20035 https://arxiv.org/abs/2403.20035
Wang B ， Wu X H ， Gu R ， Qing L B ， He X H . 2023 . Alzheimer's disease classification network based on improved residual network . Intelligent Computer and Applications ， 13 （ 3 ）： 69 - 76+82
王斌 ， 吴晓红 ， 辜蕊 ， 卿粼波 ， 何小海 . 2023 . 基于改进ResNet的阿尔茨海默症分类网络 . 智能计算机与应用 ， 13 （ 3 ）： 69 - 76+82
Xia Z ， Yue G ， Xu Y ， Feng C ， Yang M ， Wang T and Lei B . 2020 . A Novel End-to-End Hybrid Network for Alzheimer's Disease Detection Using 3D CNN and 3D CLSTM // 2020 IEEE 17th International Symposium on Biomedical Imaging （ISBI） . lowa city ： IEEE： 1 - 4 ［ DOI： 10.1109/ISBI45749.2020.9098621 http://dx.doi.org/10.1109/ISBI45749.2020.9098621 ］
Xin H ， Huang K F ， Yang C H . 2023 . Temporal series features extraction based on bi-convlstm of alzheimer’s disease pridiction CTISS model . Journal of Image and Graphics ， 28 （ 04 ）： 1146 - 1156
洪欣 ， 黄铠沣 ， 杨晨晖 . 2023 . 基于Bi-ConvLSTM时序特征提取的阿尔茨海默症预测CTISS模型 . 中国图象图形学报 ， 28 （ 04 ）： 1146 - 1156 ［ DOI： 10.11834/jig.211186 http://dx.doi.org/10.11834/jig.211186 ］
Zeng A ， Jia L F ， Pan D ， Song X W . 2019 . Early prognosis of alzheimer's disease based on convolutional neural networks and ensemble learning . Journal of Biomedical Engineering ， 36 （ 5 ）： 711 - 719
曾安 ， 贾龙飞 ， 潘丹 ， Song Xiaowei . 2019 . 基于卷积神经网络和集成学习的阿尔茨海默症早期诊断 . 生物医学工程学杂志 ， 36 （ 5 ）： 711 - 719 ［ DOI： 10.7507/1001-5515.201809040 http://dx.doi.org/10.7507/1001-5515.201809040 ］
相关文章
相关作者
相关机构
京公网安备11010802024621