Share：
Share this content in WeChat
X
[Chinese] [PDF] 33 1
• Clinical Article •
Application value of artificial intelligence-based automatic brain region segmentation technology in the study of the correlation between the volumes of hippocampal and amygdalar subfields and cerebral small vessel disease-related cognitive impairment
LIANG Xiuqin  LIU Bo  HAO Xiaoyong  LI Weirong  WANG Jianhong  WU Jiang 

Cite this article as: LIANG X Q, LIU B, HAO X Y, et al. Application value of artificial intelligence-based automatic brain region segmentation technology in the study of the correlation between the volumes of hippocampal and amygdalar subfields and cerebral small vessel disease-related cognitive impairment[J]. Chin J Magn Reson Imaging, 2026, 17(10): 28-36, 101. DOI:10.12015/issn.1674-8034.2026.10.004.


[Abstract] Objective To explore the correlation between atrophy of specific subfields of the hippocampus and amygdala and cognitive impairment in patients with cerebral small vessel disease (CSVD) using artificial intelligence (AI)-based brain region segmentation technology combined with high-resolution 3D T1-weighted MRI, and to identify the relevant imaging markers.Materials and Methods A total of 76 patients diagnosed with CSVD at Shanxi Cardiovascular Hospital from February 2025 to October 2025 were consecutively enrolled, including 38 cases of cerebral small vessel disease with cognitive impairment (CSVD-CI) and 38 cases of cerebral small vessel disease without cognitive impairment (CSVD-nCI). Thirty-eight normal controls (NC) were also enrolled. The cognitive function of all participants was evaluated using the Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE) scales. All subjects underwent conventional MRI scans and 3D-T1 MRI thin-slice scans. Brain segmentation was performed using uAI (20250730-release). ANCOVA was applied to compare brain regional and white matter hyperintensities (WMH) volumes across the three groups. Main effects were corrected using FDR; significant main effects (P < 0.05) were followed by Bonferroni-adjusted pairwise comparisons (α = 0.05 / 3 ≈ 0.0167). After correcting for FDR, a partial correlation analysis was conducted between the volumes of the three groups of brain regions with differences and the WMH volume, respectively, and the MoCA and MMSE scores. Subsequently, a multiple linear regression analysis was performed. For factors that showed statistically differences in the multiple linear regression analysis, receiver operating characteristic (ROC) curves were calculated, and the area under the curve (AUC) was used to evaluate diagnostic performance. To further evaluate the predictive calibration performance and incremental clinical benefit of the combined model, we performed calibration curves and decision curve analysis (DCA), respectively.Results The significant results of post-hoc comparisons are presented below. Compared with the CSVD-nCI group, the volumes of the right hippocampal CA3 region and the right amygdalar cortical nucleus were decreased in the CSVD-CI group (all P < 0.05). Compared with the HC group, the CSVD-CI group exhibited reduced volumes in the right hippocampal molecular layer, right hippocampal CA3 region, right amygdalar cortical nucleus, right anterior amygdala, right accessory basal amygdala, and right amygdalocortical transition zone (all P < 0.05). The volume of the right amygdalar cortical nucleus and WMH volume remained statistically significant after FDR correction. Partial correlation analysis showed that reduced volume of the right amygdalar cortical nucleus (r = 0.346, P < 0.001; r = 0.330, P < 0.001) and increased WMH volume (r = -0.737, P < 0.001; r = -0.684, P < 0.001) were associated with decreased MoCA and MMSE scores in patients with CSVD. Multi-factor linear regression analysis revealed that after adjusting for confounding variables, the volume of the right amygdalar cortical nucleus was independently positively correlated with the MoCA score (B = 14.78, 95% CI: 6.33 to 23.20, t = 3.47, P < 0.001). It was also independently positively correlated with the MMSE score (B = 10.14, 95% CI: 3.62 to 16.67, t = 3.08, P = 0.003). The total WMH volume was independently negatively correlated with the MoCA score (B = -0.93, 95% CI: -1.10 to -0.75, t = -10.53, P < 0.001). It was also independently negatively correlated with the MMSE score (B = -0.61, 95% CI: -0.74 to -0.47, t = -8.94, P < 0.001). ROC curve analyses based on residual-adjusted volumes revealed that the AUC of the right amygdalar cortical nucleus for discriminating CSVD-CI from CSVD-nCI was 0.662 (95% CI: 0.539 to 0.785), with sensitivity of 29% and specificity of 97%. For raw volume of the right amygdalar cortical nucleus, the AUC, sensitivity and specificity for differentiating CSVD-CI and CSVD-nCI were 0.664 (95% CI: 0.541 to 0.788), 66% and 74%, respectively. The raw WMH volume, the AUC for discriminating CSVD-CI from CSVD-nCI was 0.855 (95% CI: 0.763 to 0.946), with sensitivity of 74% and specificity of 92%. A joint predictive model combining the volume of right amygdalar cortical nucleus and WMH achieved an AUC of 0.884 (95% CI: 0.800 to 0.967); the sensitivity of this combined model increased to 79% while the specificity stayed at 92%. The calibration curves revealed Brier scores of 0.150, 0.142, and 0.230 for the WMH volume model, combined model, and right amygdalar cortical nucleus volume model, respectively. Decision curve analysis (DCA) demonstrated that the combined model yielded higher net benefit than the WMH-only model across the low-to-moderate threshold probability range.Conclusions Patients with CSVD-CI exhibit atrophy of specific subfields of the hippocampus and amygdala. AI-based automatic subfield segmentation combined with high-resolution 3D T1-weighted MRI can extract neuroimaging biomarkers for CSVD-CI. Clinically, locating and quantifying these biomarkers can help identify high-risk populations and facilitate early personalized interventions.
[Keywords] cerebral small-vessel disease；cognitive impairment；artificial intelligence；magnetic resonance imaging；brain regions；hippocampus；amygdala；neuroimaging；biomarkers

LIANG Xiuqin1   LIU Bo2   HAO Xiaoyong1   LI Weirong3   WANG Jianhong3   WU Jiang1*  

1 Department of Magnetic Resonance, Shanxi Provincial Cardiovascular Hospital, Shanxi Key Laboratory for Precision Diagnosis and Treatment of Heart Failure, Affiliated Cardiovascular Hospital of Shanxi Medical University, Taiyuan 030024, China

2 School of Medical Imaging, Shanxi Medical University, Taiyuan 030001, China

3 Department of Neurology, Shanxi Provincial Cardiovascular Hospital, Shanxi Key Laboratory for Precision Diagnosis and Treatment of Heart Failure, Affiliated Cardiovascular Hospital of Shanxi Medical University, Taiyuan 030024, China

Corresponding author: WU J, E-mail: wujiang1024@sina.com

Conflicts of interest   None.

Received  2026-02-12
Accepted  2026-08-25
DOI: 10.12015/issn.1674-8034.2026.10.004
Cite this article as: LIANG X Q, LIU B, HAO X Y, et al. Application value of artificial intelligence-based automatic brain region segmentation technology in the study of the correlation between the volumes of hippocampal and amygdalar subfields and cerebral small vessel disease-related cognitive impairment[J]. Chin J Magn Reson Imaging, 2026, 17(10): 28-36, 101. DOI:10.12015/issn.1674-8034.2026.10.004.

[1]
INOUE Y, SHUE F, BU G, et al. Pathophysiology and probable etiology of cerebral small vessel disease in vascular dementia and Alzheimer's disease[J/OL]. Mol Neurodegener, 2023, 18(1): 46 [2026-02-12]. https://pubmed.ncbi.nlm.gov/37434208/. DOI: 10.1186/s13024-023-00640-5.
[2]
WALLIN A, ROMAN G C, ESIRI M, et al. Update on Vascular Cognitive Impairment Associated with Subcortical Small-Vessel Disease[J]. J Alzheimers Dis, 2018, 62(3): 1417-1441. DOI: 10.3233/JAD-170803.
[3]
JOKINEN H, LIPSANEN J, SCHMIDT R, et al. Brain atrophy accelerates cognitive decline in cerebral small vessel disease: the LADIS study[J]. Neurology, 2012, 78(22): 1785-1792. DOI: 10.1212/WNL.0b013e3182583070.
[4]
CUADRADO-GODIA E, DWIVEDI P, SHARMA S, et al. Cerebral Small Vessel Disease: A Review Focusing on Pathophysiology, Biomarkers, and Machine Learning Strategies[J]. J Stroke, 2018, 20(3): 302-320. DOI: 10.5853/jos.2017.02922.
[5]
FENG Q, ZHANG Z, DAI F, et al. Volumetric alterations of hippocampal and amygdala subfields in Azheimer's disease[J/OL]. BMC Med Imaging, 2025, 25(1): 378 [2026-02-12]. https://pubmed.ncbi.nlm.gov/41013385/. DOI: 10.1186/s12880-025-01923-7.
[6]
ECKERSTROM C, OLSSON E, BORGA M, et al. Small baseline volume of left hippocampus is associated with subsequent conversion of MCI into dementia: the Göteborg MCI study[J]. J Neurol Sci, 2008, 272(1-2): 48-59. DOI: 10.1016/j.jns.2008.04.024.
[7]
DU Y, YU J, LIU M, et al. The relationship between depressive symptoms and cognitive function in Alzheimer's disease: The mediating effect of amygdala functional connectivityand radiomic features[J]. J Affect Disord, 2023, 330: 101-109. DOI: 10.1016/j.jad.2023.02.129.
[8]
CHENG Z, NIE W, LENG J, et al. Amygdala and cognitive impairment in cerebral smallvessel disease: structural, functional, and metabolic changes[J/OL]. Front Neurol, 2024, 15: 1398009 [2026-02-12]. https://pubmed.ncbi.nlm.gov/39070051/. DOI: 10.3389/fneur.2024.1398009.
[9]
PEISENIECE E, ZDANOVSKIS N, ŠNEIDERE K, et al. Amygdala Nuclei Atrophy in Cognitive Impairment and Dementia: Insights from High-Resolution Magnetic Resonance Imaging[J/OL]. Medicina (Kaunas), 2025, 61(1): 130 [2026-02-12]. https://pubmed.ncbi.nlm.gov/39859112/. DOI: 10.3390/medicina61010130.
[10]
PADULO C, SESTIERI C, PUNZI M, et al. Atrophy of specific amygdala subfields in subjects converting to mild cognitive impairment[J/OL]. Alzheimers Dement (N Y), 2023, 9(4): e12436 [2026-02-12]. https://pubmed.ncbi.nlm.gov/38053753/. DOI: 10.1002/trc2.12436.
[11]
ZHANG C, KONG M, WEI H, et al. The effect of ApoE ε 4 on clinical and structural MRI markers in prodromal Alzheimer's disease[J]. Quant Imaging Med Surg, 2020, 10(2): 464-474. DOI: 10.21037/qims.2020.01.14.
[12]
PUNZI M, SESTIERI C, PICEMI E, et al. Atrophy of hippocampal subfields and amygdala nuclei in subjects with mild cognitive impairment progressing to Alzheimer's disease[J/OL]. Heliyon, 2024, 10(6): e27429 [2026-02-12]. https://pubmed.ncbi.nlm.gov/38509925/. DOI: 10.1016/j.heliyon.2024.e27429.
[13]
LI X, LI D, LI Q, et al. Hippocampal subfield volumetry in patients with subcortical vascular mild cognitive impairment[J/OL]. Sci Rep, 2016, 6: 20873 [2026-02-12]. https://pubmed.ncbi.nlm.gov/26876151/. DOI: 10.1038/srep20873.
[14]
WEI Y, HUANG N, LIU Y, et al. Hippocampal and Amygdalar Morphological Abnormalities in Alzheimer's Disease Based on Three Chinese MRI Datasets[J]. Curr Alzheimer Res, 2020, 17(13): 1221-1231. DOI: 10.2174/1567205018666210218150223.
[15]
GAO W, BISWAL B, CHEN S, et al. Functional coupling of the orbitofrontal cortex and the basolateral amygdala mediates the association between spontaneous reappraisal and emotional response[J/OL]. Neuroimage, 2021, 232: 117918 [2026-02-12]. https://pubmed.ncbi.nlm.gov/33652140/. DOI: 10.1016/j.neuroimage.2021.117918.
[16]
LA JOIE R, FOUQUET M, MEZENGE F, et al. Differential effect of age on hippocampal subfields assessed using a new high-resolution 3T MR sequence[J]. Neuroimage, 2010, 53(2): 506-514. DOI: 10.1016/j.neuroimage.2010.06.024.
[17]
DE FLORES R, LA JOIE R, LANDEAU B, et al. Effects of age and Alzheimer's disease on hippocampal subfields: comparison between manual and FreeSurfer volumetry[J]. Hum Brain Mapp, 2015, 36(2): 463-474. DOI: 10.1002/hbm.22640.
[18]
LA JOIE R, PERROTIN A, DE SAINT-VÉRON A, et al. Hippocampal subfield volumetry in mild cognitive impairment, Alzheimer's disease and semantic dementia[J]. Neuroimage Clin, 2013, 3: 155-162. DOI: 10.1016/j.nicl.2013.08.007.
[19]
Neuroradiology Group of Chinese Society of Radiology of Chinese Medical Association. Expert Consensus on the Standardized Application of MRI in Cerebral Small Vessel Disease[J]. Chin J Radiol, 2024, 58(1): 6-17. DOI: 10.3760/cma.j.cn112149-20231031-00334.
[20]
HAINSWORTH A H, MARKUS H S, SCHNEIDER J A. Cerebral Small Vessel Disease, Hypertension, and Vascular Contributions to Cognitive Impairment and Dementia[J]. Hypertension, 2024, 81(1): 75-86. DOI: 10.1161/HYPERTENSIONAHA.123.19943.
[21]
WONG F C C, YATAWARA C, LOW A, et al. Cerebral Small Vessel Disease Influences Hippocampal Subfield Atrophy in Mild Cognitive Impairment[J]. Transl Stroke Res, 2021, 12(2): 284-292. DOI: 10.1007/s12975-020-00847-4.
[22]
TULLY P J, YANO Y, LAUNER L J, et al. Association Between Blood Pressure Variability and Cerebral Small-Vessel Disease: A Systematic Review and Meta-Analysis[J/OL]. J Am Heart Assoc, 2020, 9(1): e13841 [2026-02-12]. https://pubmed.ncbi.nlm.gov/31870233/. DOI: 10.1161/JAHA.119.013841.
[23]
BEN-ZION Z, KOREM N, FINE N B, et al. Structural Neuroimaging of Hippocampus and Amygdala Subregions in Posttraumatic Stress Disorder: A Scoping Review[J]. Biol Psychiatry Glob Open Sci, 2023, 4(1): 120-134. DOI: 10.1016/j.bpsgos.2023.07.001.
[24]
GABERE M, THU P N T, GRAFF-RADFORD J, et al. Automated Hippocampal SubfieldVolumetric Analyses in Atypical Alzheimer's Disease[J]. J Alzheimers Dis, 2020, 78(3): 927-937. DOI: 10.3233/JAD-200625.
[25]
HUANG J, CHENG R, LIU X, et al. Unraveling the link: white matter damage, gray matter atrophy and memory impairment in patients with subcortical ischemic vascular disease[J/OL]. Front Neurosci, 2024, 18: 1355207 [2026-02-12]. https://pubmed.ncbi.nlm.gov/38362024/. DOI: 10.3389/fnins.2024.1355207.
[26]
KESNER R P. An analysis of the dentate gyrus function[J]. Behav Brain Res, 2013, 254: 1-7. DOI: 10.1016/j.bbr.2013.01.012.
[27]
CAREY D, NOLAN H, KENNY R A, et al. Dissociable age and memory relationships with hippocampal subfield volumes in vivo:Data from the Irish Longitudinal Study on Ageing (TILDA)[J/OL]. Sci Rep, 2019, 9(1): 10981 [2026-02-12]. https://pubmed.ncbi.nlm.gov/31358771/. DOI: 10.1038/s41598-019-46481-5.
[28]
HAINMUELLER T, BARTOS M. Dentate gyrus circuits for encoding, retrieval and discrimination of episodic memories[J]. Nat Rev Neurosci, 2020, 21(3): 153-168. DOI: 10.1038/s41583-019-0260-z.
[29]
CECCOM J, HALEEY H, DAUMAS S, et al. A specific role for hippocampal mossy fiber's zinc in rapid storage of emotional memories[J]. Learn Mem, 2014, 21(5): 287-297. DOI: 10.1101/lm.033472.113.
[30]
KARPF J, UNICHENKO P, CHALMERS N, et al. Dentate gyrus astrocytes exhibit layer-specific molecular, morphological and physiological features[J]. Nat Neurosci, 2022, 25(12): 1626-1638. DOI: 10.1038/s41593-022-01192-5.
[31]
PESSOA L. On the relationship between emotion and cognition[J]. Nat Rev Neurosci, 2008, 9(2): 148-158. DOI: 10.1038/nrn2317.
[32]
PESSOA L, ADOLPHS R. Emotion processing and the amygdala: from a 'low road' to 'many roads' of evaluating biological significance[J]. Nat Rev Neurosci, 2010, 11(11): 773-783. DOI: 10.1038/nrn2920.
[33]
DEN HEIJER T, LAUNER L J, PRINS N D, et al. Association between blood pressure, white matter lesions, and atrophy of the medial temporal lobe[J]. Neurology, 2005, 64(2): 263-267. DOI: 10.1212/01.WNL.0000150582.86455.8B.
[34]
STOUFFER K M, CHEN C, KULASON S, et al. Early amygdala and ERC atrophy linked to 3D reconstruction of rostral neurofibrillary tau tangle pathology in Alzheimer's disease[J/OL]. Neuroimage Clin, 2023, 38: 103374 [2026-02-12]. https://pubmed.ncbi.nlm.gov/36934675/. DOI: 10.1016/j.nicl.2023.103374.
[35]
BZDOK D, LAIRD A R, ZILLES K, et al. An investigation of the structural, connectional, and functional subspecialization in the human amygdala[J]. Hum Brain Mapp, 2013, 34(12): 3247-3266. DOI: 10.1002/hbm.22138.
[36]
INMAN C S, HOLLEARN M K, AUGUSTIN L, et al. Discovering how the amygdala shapes human behavior: From lesion studies to neuromodulation[J]. Neuron, 2023, 111(24): 3906-3910. DOI: 10.1016/j.neuron.2023.09.040.
[37]
CHENG Z, YANG L, LI J, et al. Cognitive impairment and amygdala subregion volumes in elderly with cerebral small vessel disease: A large prospective cohort study[J/OL]. Neurobiol Dis, 2024, 202: 106716 [2026-02-12]. https://pubmed.ncbi.nlm.gov/39490683/. DOI: 10.1016/j.nbd.2024.106716.
[38]
FENG W, LEI X, XU S, et al. Association between white matter hyperintensities and gray matter volume in cerebral small vessel disease: insights from periventricular and deep white matter lesions[J/OL]. Front Neurol, 2025, 16: 1590997 [2026-02-12]. https://pubmed.ncbi.nlm.gov/40703777/. DOI: 10.3389/fneur.2025.1590997.
[39]
LI H, JACOB M A, CAI M, et al. Regional cortical thinning, demyelination and iron loss in cerebral small vessel disease[J]. Brain, 2023, 146(11): 4659-4673. DOI: 10.1093/brain/awad220.
[40]
MAYER C, FREY B M, SCHLEMM E, et al. Linking cortical atrophy to white matter hyperintensities of presumed vascular origin[J]. J Cereb Blood Flow Metab, 2021, 41(7): 1682-1691. DOI: 10.1177/0271678X20974170.
[41]
FENG M, WEN H, XIN H, et al. Decreased Local Specialization of Brain Structural Networks Associated with Cognitive Dysfuntion Revealed by Probabilistic Diffusion Tractography for Different Cerebral Small Vessel Disease Burdens[J]. Mol Neurobiol, 2024, 61(1): 326-339. DOI: 10.1007/s12035-023-03597-0.
[42]
ZENG W, CHEN Y, ZHU Z, et al. Severity of white matter hyperintensities: Lesion patterns, cognition, and microstructural changes[J]. J Cereb Blood Flow Metab, 2020, 40(12): 2454-2463. DOI: 10.1177/0271678X19893600.
[43]
RIZVI B, NARKHEDE A, LAST B S, et al. The effect of white matter hyperintensities on cognition is mediated by cortical atrophy[J]. Neurobiol Aging, 2018, 64: 25-32. DOI: 10.1016/j.neurobiolaging.2017.12.006.
[44]
PHUAH C, CHEN Y, STRAIN J F, et al. Association of Data-Driven White Matter Hyperintensity Spatial Signatures With Distinct Cerebral Small Vessel Disease Etiologies[J/OL]. Neurology, 2022, 99(23): e2535-e2547 [2026-02-12]. https://pubmed.ncbi.nlm.gov/36123127/. DOI: 10.1212/WNL.0000000000201186.
[45]
LAMBERT C, BENJAMIN P, ZEESTRATEN E, et al. Longitudinal patterns of leukoaraiosis and brain atrophy in symptomatic small vessel disease[J/OL]. Brain, 2016, 139(Pt 4): 1136-1151. DOI: 10.1093/brain/aww009.
[46]
MARCHESI V T. Alzheimer's dementia begins as a disease of small blood vessels, damaged by oxidative-induced inflammation and dysregulated amyloid metabolism: implications for early detection and therapy[J]. FASEB J, 2011, 25(1): 5-13. DOI: 10.1096/fj.11-0102ufm.
[47]
WANG Y, YANG Y, WANG T, et al. Correlation between White Matter Hyperintensities Related Gray Matter Volume and Cognition in Cerebral Small Vessel Disease[J/OL]. J Stroke Cerebrovasc Dis, 2020, 29(12): 105275 [2026-02-12]. https://pubmed.ncbi.nlm.gov/32992182/. DOI: 10.1016/j.jstrokecerebrovasdis.2020.105275.

PREV Elevated free water imaging metrics mediate white matter injury and cognitive decline in elderly patients with hypertension: An MRI study
NEXT Topological robustness and degree distribution characteristics of structural covariance networks in frontotemporal dementia based on network attack and degree distribution analyses
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn