Parkinson′s disease (PD), the second most prevalent neurodegenerative disease, has shown an increasing incidence in recent years, significantly impacting the quality of life of elderly individuals and their families. The onset of cognitive impairment and Parkinson′s disease dementia (PDD) are critical milestones in the progression of PD. Imaging examinations, serving as essential detection and evaluation tools for neurodegenerative diseases, have been increasingly applied to PD-related research. Different imaging techniques demonstrate distinct advantages and features in screening cognitive dysfunction in PD. Choosing a reliable imaging method not only maximizes the advantages of the examination, enhancing the sensitivity and specificity in identifying PD patients at risk of cognitive impairment, but also reduces the frequency of examinations and the radiation dose received by patients. This review summarizes the existing findings on imaging markers of cognitive impairment in PD from three aspects, structural imaging, functional imaging, and multimodal techniques. Furthermore, it explores future research directions, aiming to provide powerful imaging support for the clinical diagnosis and treatment of PD.
CAO Mingming, WANG Hui, YIN Yafu. Current research status of imaging markers for cognitive impairment in Parkinson′s disease. Journal of Shanghai Jiao Tong University (Medical Science)[J], 2025, 45(5): 646-652 doi:10.3969/j.issn.1674-8115.2025.05.014
帕金森病（Parkinson′s disease，PD）是一种常见的神经系统退行性疾病，严重影响中老年人健康。其临床表现包括运动系统症状和非运动系统症状，而非运动系统症状中最重要的表现之一则是认知功能障碍。病程大于10年的PD患者中，有超过75%的患者会出现认知功能障碍，且这部分患者往往会发展为帕金森病痴呆（Parkinson′s disease dementia，PDD）［1］。帕金森病认知功能障碍（Parkinson′s disease with cognitive impairment，PD-CI）不仅影响患者及其家属的生活质量，还对社会医疗资源和经济发展造成巨大负担［2］。影像学检查作为PD诊断及病情评估的重要手段，近年来得到广泛应用。特别是针对多巴胺系统开发的各种特异性探针，与PET/CT及PET/MRI等多模态影像技术相结合，能精准描绘大脑多巴胺耗竭情况［3］。目前PD-CI的诊断主要依赖于临床症状及神经功能评估量表，如蒙特利尔认知量表（Montreal Cognitive Assessment，MoCA）和迷你精神状态检查量表（Mini-Mental State Examination，MMSE）等，然而量表的评估过程不仅耗时，且难以及时、准确地反映疾病进展［4］。因此，选择合适的影像标志物进行评估不仅可以及时筛选出具有PD-CI风险的患者，同时还能有效减少患者不必要的检查。因此，本文拟从结构影像、功能影像和其他影像技术3个方面对PD-CI影像标志物的研究现状进行梳理与总结。
除了PD特异性病理蛋白α-突触核蛋白（α-synuclein，α-syn）沉积外，β-淀粉样蛋白（amyloid-β，Aβ）和tau蛋白的沉积也是PDD的重要病理改变。一项采用11C-PiB和18F-AV1451双显像剂的研究［39］表明，楔前叶和颞下回tau蛋白的沉积与路易体痴呆（dementia with Lewy body，DLB）以及PD-CI有关；除此之外，研究者还发现Aβ斑块和tau蛋白沉积之间具有协同作用，较高的Aβ斑块沉积可能是驱动tau蛋白沉积至皮质区域的关键因素。另一项对PD-CI患者大脑tau蛋白沉积的研究［40］则发现，PD-CI患者颞叶下部的tau蛋白沉积高于健康对照，而内侧嗅区的tau蛋白沉积高于认知功能正常的PD患者。Aβ斑块沉积方面，有研究［41］发现特定脑区的Aβ斑块沉积与PD-CI相关，如左前扣带回和右顶叶等，因此在对PD-CI和PDD患者的Aβ斑块负荷进行评估时，选择某些脑区而非全脑皮层的Aβ沉积评分更有临床意义。不过也有一些研究得出了相悖的结论。一项横断面研究［42］并未发现Aβ沉积在认知功能正常的PD患者和轻度PD-CI患者间存在差异。目前Aβ斑块和tau蛋白沉积在PD中的具体机制尚未明确，但其对于PD-CI的诊断已经展现出一定的临床价值。因此探讨其作用机制，以及如何选择合适的时机应用病理蛋白影像检查更能有助于患者的诊疗将是下一步的研究方向。
机器学习及影像组学是医学技术和人工智能发展的产物，可以通过提取不同的影像信息构建模型。其模型可以单独应用，也可与其他代谢组学、检验学及流行病学等领域信息相结合，更好地服务于临床诊疗和科研［59］。目前影像组学已经在神经退行性疾病诊疗实践中得到了广泛开展，可用于鉴别诊断PD及非典型帕金森综合征［60］。除此之外，将T1加权和DTI序列相结合，还能实现早期PD运动亚型的分型，尤其是针对震颤为主型和姿势不稳定型的PD患者［61］。另外，还有研究［62］利用机器学习研究PD患者的嗓音变化特征，结果发现计算机语音分析可以捕捉到患者的嗓音改变情况，对于PD病情评估具有重要潜力。但目前关于机器学习和影像组学在PD-CI中的应用相对较少，不过已有研究发现了积极的结果。ZHU等［63］将临床症状、血清学数据和影像数据相结合，通过机器学习构建轻度PD-CI患者的诊断模型，该模型平均准确率可达76%，平均受试者操作特征曲线的曲线下面积（area under the curve，AUC）达到了0.84，进一步证实了多模态组学和临床数据整合在PD-CI评估中的重要潜力。随着技术的不断发展，机器学习和影像组学势必成为未来的研究热点，并且有望在PD诊疗领域中发挥更重要的作用。
The topic selection and writing instruction were performed by WANG Hui and YIN Yafu. The manuscript was drafted and revised by CAO Mingming. All authors have read the last version of paper and consented to submission.
利益冲突声明
所有作者声明不存在利益冲突。
COMPETING INTERESTS
All authors disclose no relevant conflict of interests.
WEINTRAUB D, CASPELL-GARCIA C, SIMUNI T, et al. Neuropsychiatric symptoms and cognitive abilities over the initial quinquennium of Parkinson disease[J]. Ann Clin Transl Neurol, 2020, 7(4): 449-461.
PIGOTT J S, DAVIES N, CHESTERMAN E, et al. Delivering optimal care to people with cognitive impairment in Parkinson′s disease: a qualitative study of patient, caregiver, and professional perspectives[J]. Parkinsons Dis, 2023, 2023: 9732217.
NIU J Q, ZHONG Y, JIN C T, et al. Positron emission tomography imaging of synaptic dysfunction in Parkinson′s disease[J]. Neurosci Bull, 2024, 40(6): 743-758.
JOHAR I, MOLLENHAUER B, AARSLAND D. Cerebrospinal fluid biomarkers of cognitive decline in Parkinson′s disease[J]. Int Rev Neurobiol, 2017, 132: 275-294.
FILIPPI M, CANU E, DONZUSO G, et al. Tracking cortical changes throughout cognitive decline in Parkinson′s disease[J]. Mov Disord, 2020, 35(11): 1987-1998.
SARASSO E, AGOSTA F, PIRAMIDE N, et al. Progression of grey and white matter brain damage in Parkinson′s disease: a critical review of structural MRI literature[J]. J Neurol, 2021, 268(9): 3144-3179.
GAO Y Y, NIE K, HUANG B, et al. Changes of brain structure in Parkinson′s disease patients with mild cognitive impairment analyzed via VBM technology[J]. Neurosci Lett, 2017, 658: 121-132.
JIANG Y Q, CHEN Q Z, YANG Y, et al. White matter lesions contribute to motor and non-motor disorders in Parkinson′s disease: a critical review[J]. Geroscience, 2025, 47(1): 591-609.
DADAR M, GEE M, SHUAIB A, et al. Cognitive and motor correlates of grey and white matter pathology in Parkinson′s disease[J]. Neuroimage Clin, 2020, 27: 102353.
JEONG S H, LEE H S, JUNG J H, et al. Associations between white matter hyperintensities, striatal dopamine loss, and cognition in drug-naïve Parkinson′s disease[J]. Parkinsonism Relat Disord, 2022, 97: 1-7.
HANNING U, TEUBER A, LANG E, et al. White matter hyperintensities are not associated with cognitive decline in early Parkinson′s disease: the DeNoPa cohort[J]. Parkinsonism Relat Disord, 2019, 69: 61-67.
DE BARTOLO M I, OJHA A, LEODORI G, et al. Association of early fMRI connectivity alterations with different cognitive phenotypes in patients with newly diagnosed parkinson disease[J]. Neurology, 2025, 104(1): e210192.
ZARIFKAR P, KIM J, LA C, et al. Cognitive impairment in Parkinson′s disease is associated with Default Mode Network subsystem connectivity and cerebrospinal fluid Aβ[J]. Parkinsonism Relat Disord, 2021, 83: 71-78.
WANG Z J, JIA X Q, CHEN H M, et al. Abnormal spontaneous brain activity in early Parkinson′s disease with mild cognitive impairment: a resting-state fMRI study[J]. Front Physiol, 2018, 9: 1093.
CAMPBELL M C, JACKSON J J, KOLLER J M, et al. Proteinopathy and longitudinal changes in functional connectivity networks in Parkinson disease[J]. Neurology, 2020, 94(7): e718-e728.
XING Y L, FU S S, LI M, et al. Regional neural activity changes in Parkinson′s disease-associated mild cognitive impairment and cognitively normal patients[J]. Neuropsychiatr Dis Treat, 2021, 17: 2697-2706.
WANG Q G, HE W, LIU D H, et al. Functional connectivity in Parkinson′s disease patients with mild cognitive impairment[J]. Int J Gen Med, 2021, 14: 2623-2630.
CHEN L, HUANG T, MA D, et al. Altered default mode network functional connectivity in Parkinson′s disease: a resting-state functional magnetic resonance imaging study[J]. Front Neurosci, 2022, 16: 905121.
IMARISIO A, PILOTTO A, PREMI E, et al. Atypical brain FDG-PET patterns increase the risk of long-term cognitive and motor progression in Parkinson′s disease[J]. Parkinsonism Relat Disord, 2023, 115: 105848.
BOOTH S, KO J H. Radionuclide imaging of the neuroanatomical and neurochemical substrate of cognitive decline in Parkinson′s disease[J]. Nucl Med Mol Imaging, 2024, 58(4): 213-226.
MEYER P T, FRINGS L, RÜCKER G, et al. 18F-FDG PET in Parkinsonism: differential diagnosis and evaluation of cognitive impairment[J]. J Nucl Med, 2017, 58(12): 1888-1898.
BABA T, HOSOKAI Y, NISHIO Y, et al. Longitudinal study of cognitive and cerebral metabolic changes in Parkinson′s disease[J]. J Neurol Sci, 2017, 372: 288-293.
YOO H S, KIM H K, NA H K, et al. Association of striatal dopamine depletion and brain metabolism changes with motor and cognitive deficits in patients with parkinson disease[J]. Neurology, 2024, 103(12): e210105.
CHUNG S J, YOO H S, OH J S, et al. Effect of striatal dopamine depletion on cognition in de novo Parkinson′s disease[J]. Parkinsonism Relat Disord, 2018, 51: 43-48.
CHRISTOPHER L, MARRAS C, DUFF-CANNING S, et al. Combined insular and striatal dopamine dysfunction are associated with executive deficits in Parkinson′s disease with mild cognitive impairment[J]. Brain, 2014, 137(Pt 2): 565-575.
HONG Y J, CHOI S H, KIM S, et al. Cognitive and neurodegenerative trajectories of subjective cognitive decline according to baseline biomarkers: results of the CoSCo study[J]. Alzheimers Dement, 2025, 21(2): e14473.
CHUN M Y, CHUNG S J, KIM S H, et al. Hippocampal perfusion affects motor and cognitive functions in parkinson disease: an early phase 18F-FP-CIT positron emission tomography study[J]. Ann Neurol, 2024, 95(2): 388-399.
WITZIG V, PJONTEK R, TAN S H, et al. Modulating the cholinergic system: novel targets for deep brain stimulation in Parkinson′s disease[J]. J Neurochem, 2025, 169(2): e16264.
KALBE E, FOLKERTS A K, WITT K, et al. German Society of Neurology guidelines for the diagnosis and treatment of cognitive impairment and affective disorders in people with Parkinson′s disease: new spotlights on diagnostic procedures and non-pharmacological interventions[J]. J Neurol, 2024, 271(11): 7330-7357.
SLATER N M, MELZER T R, MYALL D J, et al. Cholinergic basal forebrain integrity and cognition in Parkinson′s disease: a reappraisal of magnetic resonance imaging evidence[J]. Mov Disord, 2024, 39(12): 2155-2172.
VAN DER ZEE S, MÜLLER M L T M, KANEL P, et al. Cholinergic denervation patterns across cognitive domains in Parkinson′s disease[J]. Mov Disord, 2021, 36(3): 642-650.
SCHUMACHER J, RAY N, TEIPEL S, et al. Associations of cholinergic system integrity with cognitive decline in GBA1 and LRRK2 mutation carriers[J]. NPJ Parkinsons Dis, 2024, 10(1): 127.
BOHNEN N I, ALBIN R L, MÜLLER M L T M, et al. Frequency of cholinergic and caudate nucleus dopaminergic deficits across the predemented cognitive spectrum of Parkinson disease and evidence of interaction effects[J]. JAMA Neurol, 2015, 72(2): 194-200.
SOARES É N, COSTA A C D S, FERROLHO G J, et al. Nicotinic acetylcholine receptors in glial cells as molecular target for Parkinson′s disease[J]. Cells, 2024, 13(6): 474.
GOMPERTS S N, LOCASCIO J J, MAKARETZ S J, et al. Tau positron emission tomographic imaging in the Lewy body diseases[J]. JAMA Neurol, 2016, 73(11): 1334-1341.
MIHAESCU A S, VALLI M, URIBE C, et al. Beta amyloid deposition and cognitive decline in Parkinson′s disease: a study of the PPMI cohort[J]. Mol Brain, 2022, 15(1): 79.
THEIS H, PAVESE N, REKTOROVÁ I, et al. Imaging biomarkers in prodromal and earliest phases of Parkinson′s disease[J]. J Parkinsons Dis, 2024, 14(s2): S353-S365.
EDISON P, AHMED I, FAN Z, et al. Microglia, amyloid, and glucose metabolism in Parkinson′s disease with and without dementia[J]. Neuropsychopharmacology, 2013, 38(6): 938-949.
PERERA MOLLIGODA ARACHCHIGE A S, GARNER A K. Seven tesla MRI in Alzheimer′s disease research: state of the art and future directions: a narrative review[J]. AIMS Neurosci, 2023, 10(4): 401-422.
KHAN M A, HAIDER N, SINGH T, et al. Promising biomarkers and therapeutic targets for the management of Parkinson′s disease: recent advancements and contemporary research[J]. Metab Brain Dis, 2023, 38(3): 873-919.
WIELER M, GEE M, WAYNE MARTIN W R. Longitudinal midbrain changes in early Parkinson′s disease: iron content estimated from R2*/MRI[J]. Parkinsonism Relat Disord, 2015, 21(3): 179-183.
WELTON T, HARTONO S, SHIH Y C, et al. Ultra-high-field 7T MRI in Parkinson′s disease: ready for clinical use? A narrative review[J]. Quant Imaging Med Surg, 2023, 13(11): 7607-7620.
GRIMALDI S, EL MENDILI M M, ZAARAOUI W, et al. Increased sodium concentration in substantia nigra in early Parkinson′s disease: a preliminary study with ultra-high field (7T) MRI[J]. Front Neurol, 2021, 12: 715618.
XIAO K M, LI J L, ZHOU L Y, et al. Retinopathy in Parkinson′s disease: a potential biomarker for early diagnosis and clinical assessment[J]. Neuroscience, 2025, 565: 202-210.
HANNAWAY N, ZARKALI A, LEYLAND L A, et al. Visual dysfunction is a better predictor than retinal thickness for dementia in Parkinson′s disease[J]. J Neurol Neurosurg Psychiatry, 2023, 94(9): 742-750.
CHRYSOU A, HEIKKA T, VAN DER ZEE S, et al. Reduced thickness of the retina in de novo Parkinson′s disease shows a distinct pattern, different from glaucoma[J]. J Parkinsons Dis, 2024, 14(3): 507-519.
ZHANG J R, CAO Y L, LI K, et al. Correlations between retinal nerve fiber layer thickness and cognitive progression in Parkinson′s disease: a longitudinal study[J]. Parkinsonism Relat Disord, 2021, 82: 92-97.
MURUETA-GOYENA A, ROMERO-BASCONES D, TEIJEIRA-PORTAS S, et al. Association of retinal neurodegeneration with the progression of cognitive decline in Parkinson′s disease[J]. NPJ Parkinsons Dis, 2024, 10(1): 26.
BU S T, PANG H Z, LI X L, et al. Multi-parametric radiomics of conventional T1 weighted and susceptibility-weighted imaging for differential diagnosis of idiopathic Parkinson′s disease and multiple system atrophy[J]. BMC Med Imaging, 2023, 23(1): 204.
PANAHI M, HOSSEINI M S. Multi-modality radiomics of conventional T1 weighted and diffusion tensor imaging for differentiating Parkinson′s disease motor subtypes in early-stages[J]. Sci Rep, 2024, 14(1): 20708.
OLIVEIRA G C, PAH N D, NGO Q C, et al. A pilot study for speech assessment to detect the severity of Parkinson′s disease: an ensemble approach[J]. Comput Biol Med, 2025, 185: 109565.
ZHU Y Y, WANG F, NING P P, et al. Multimodal neuroimaging-based prediction of Parkinson′s disease with mild cognitive impairment using machine learning technique[J]. NPJ Parkinsons Dis, 2024, 10(1): 218.
... 机器学习及影像组学是医学技术和人工智能发展的产物，可以通过提取不同的影像信息构建模型.其模型可以单独应用，也可与其他代谢组学、检验学及流行病学等领域信息相结合，更好地服务于临床诊疗和科研［59］.目前影像组学已经在神经退行性疾病诊疗实践中得到了广泛开展，可用于鉴别诊断PD及非典型帕金森综合征［60］.除此之外，将T1加权和DTI序列相结合，还能实现早期PD运动亚型的分型，尤其是针对震颤为主型和姿势不稳定型的PD患者［61］.另外，还有研究［62］利用机器学习研究PD患者的嗓音变化特征，结果发现计算机语音分析可以捕捉到患者的嗓音改变情况，对于PD病情评估具有重要潜力.但目前关于机器学习和影像组学在PD-CI中的应用相对较少，不过已有研究发现了积极的结果.ZHU等［63］将临床症状、血清学数据和影像数据相结合，通过机器学习构建轻度PD-CI患者的诊断模型，该模型平均准确率可达76%，平均受试者操作特征曲线的曲线下面积（area under the curve，AUC）达到了0.84，进一步证实了多模态组学和临床数据整合在PD-CI评估中的重要潜力.随着技术的不断发展，机器学习和影像组学势必成为未来的研究热点，并且有望在PD诊疗领域中发挥更重要的作用. ...
1
... 机器学习及影像组学是医学技术和人工智能发展的产物，可以通过提取不同的影像信息构建模型.其模型可以单独应用，也可与其他代谢组学、检验学及流行病学等领域信息相结合，更好地服务于临床诊疗和科研［59］.目前影像组学已经在神经退行性疾病诊疗实践中得到了广泛开展，可用于鉴别诊断PD及非典型帕金森综合征［60］.除此之外，将T1加权和DTI序列相结合，还能实现早期PD运动亚型的分型，尤其是针对震颤为主型和姿势不稳定型的PD患者［61］.另外，还有研究［62］利用机器学习研究PD患者的嗓音变化特征，结果发现计算机语音分析可以捕捉到患者的嗓音改变情况，对于PD病情评估具有重要潜力.但目前关于机器学习和影像组学在PD-CI中的应用相对较少，不过已有研究发现了积极的结果.ZHU等［63］将临床症状、血清学数据和影像数据相结合，通过机器学习构建轻度PD-CI患者的诊断模型，该模型平均准确率可达76%，平均受试者操作特征曲线的曲线下面积（area under the curve，AUC）达到了0.84，进一步证实了多模态组学和临床数据整合在PD-CI评估中的重要潜力.随着技术的不断发展，机器学习和影像组学势必成为未来的研究热点，并且有望在PD诊疗领域中发挥更重要的作用. ...
1
... 机器学习及影像组学是医学技术和人工智能发展的产物，可以通过提取不同的影像信息构建模型.其模型可以单独应用，也可与其他代谢组学、检验学及流行病学等领域信息相结合，更好地服务于临床诊疗和科研［59］.目前影像组学已经在神经退行性疾病诊疗实践中得到了广泛开展，可用于鉴别诊断PD及非典型帕金森综合征［60］.除此之外，将T1加权和DTI序列相结合，还能实现早期PD运动亚型的分型，尤其是针对震颤为主型和姿势不稳定型的PD患者［61］.另外，还有研究［62］利用机器学习研究PD患者的嗓音变化特征，结果发现计算机语音分析可以捕捉到患者的嗓音改变情况，对于PD病情评估具有重要潜力.但目前关于机器学习和影像组学在PD-CI中的应用相对较少，不过已有研究发现了积极的结果.ZHU等［63］将临床症状、血清学数据和影像数据相结合，通过机器学习构建轻度PD-CI患者的诊断模型，该模型平均准确率可达76%，平均受试者操作特征曲线的曲线下面积（area under the curve，AUC）达到了0.84，进一步证实了多模态组学和临床数据整合在PD-CI评估中的重要潜力.随着技术的不断发展，机器学习和影像组学势必成为未来的研究热点，并且有望在PD诊疗领域中发挥更重要的作用. ...
1
... 机器学习及影像组学是医学技术和人工智能发展的产物，可以通过提取不同的影像信息构建模型.其模型可以单独应用，也可与其他代谢组学、检验学及流行病学等领域信息相结合，更好地服务于临床诊疗和科研［59］.目前影像组学已经在神经退行性疾病诊疗实践中得到了广泛开展，可用于鉴别诊断PD及非典型帕金森综合征［60］.除此之外，将T1加权和DTI序列相结合，还能实现早期PD运动亚型的分型，尤其是针对震颤为主型和姿势不稳定型的PD患者［61］.另外，还有研究［62］利用机器学习研究PD患者的嗓音变化特征，结果发现计算机语音分析可以捕捉到患者的嗓音改变情况，对于PD病情评估具有重要潜力.但目前关于机器学习和影像组学在PD-CI中的应用相对较少，不过已有研究发现了积极的结果.ZHU等［63］将临床症状、血清学数据和影像数据相结合，通过机器学习构建轻度PD-CI患者的诊断模型，该模型平均准确率可达76%，平均受试者操作特征曲线的曲线下面积（area under the curve，AUC）达到了0.84，进一步证实了多模态组学和临床数据整合在PD-CI评估中的重要潜力.随着技术的不断发展，机器学习和影像组学势必成为未来的研究热点，并且有望在PD诊疗领域中发挥更重要的作用. ...
1
... 机器学习及影像组学是医学技术和人工智能发展的产物，可以通过提取不同的影像信息构建模型.其模型可以单独应用，也可与其他代谢组学、检验学及流行病学等领域信息相结合，更好地服务于临床诊疗和科研［59］.目前影像组学已经在神经退行性疾病诊疗实践中得到了广泛开展，可用于鉴别诊断PD及非典型帕金森综合征［60］.除此之外，将T1加权和DTI序列相结合，还能实现早期PD运动亚型的分型，尤其是针对震颤为主型和姿势不稳定型的PD患者［61］.另外，还有研究［62］利用机器学习研究PD患者的嗓音变化特征，结果发现计算机语音分析可以捕捉到患者的嗓音改变情况，对于PD病情评估具有重要潜力.但目前关于机器学习和影像组学在PD-CI中的应用相对较少，不过已有研究发现了积极的结果.ZHU等［63］将临床症状、血清学数据和影像数据相结合，通过机器学习构建轻度PD-CI患者的诊断模型，该模型平均准确率可达76%，平均受试者操作特征曲线的曲线下面积（area under the curve，AUC）达到了0.84，进一步证实了多模态组学和临床数据整合在PD-CI评估中的重要潜力.随着技术的不断发展，机器学习和影像组学势必成为未来的研究热点，并且有望在PD诊疗领域中发挥更重要的作用. ...
