Share：
Share this content in WeChat
X
[Chinese][PDF]381
• Review •
Deep learning-based multimodal magnetic resonance imaging techniques and their research progress in depression diagnosis and treatment
WANG Zihao  SONG Yuru  SU Hongxiao  SUN Jinnan  YI Wei  REN Rui 

Cite this article as: WANG Z H, SONG Y R, SU H X, et al. Deep learning-based multimodal magnetic resonance imaging techniques and their research progress in depression diagnosis and treatment[J]. Chin J Magn Reson Imaging, 2025, 16(5): 184-189, 228 DOI:10.12015/issn.1674-8034.2025.05.028.


[Abstract] In recent years, multimodal magnetic resonance imaging technology has become a highly promising and widely applied frontier technology in medical imaging. By integrating various imaging modalities, it provides more comprehensive and richer diagnostic information than single-modality imaging. This has opened up a new pathway for the diagnosis of mental disorders such as depression, which lack objective biological markers. However, multimodal data are characterized by high dimensionality, heterogeneity, and complex associations between modalities, which pose challenges for traditional data analysis methods. Deep learning technology, with its powerful ability to process high-dimensional data, can automatically extract valuable diagnostic features from complex neuroimaging data, offering the potential for individualized diagnosis and treatment. This method provides a new perspective and development direction for efficiently and accurately processing complex multimodal magnetic resonance data. This review summarizes the integration strategies of commonly used deep learning network models with multimodal MRI sequences and their application value in depression, explores future research directions, and provides selection strategies for deep learning models in MRI research of depression.
[Keywords] deep learning；magnetic resonance imaging；multimodal；depression；brain networks

WANG Zihao1   SONG Yuru2   SU Hongxiao3   SUN Jinnan1   YI Wei1   REN Rui1*  

1 Department of Radiology, Binzhou Medical University Hospital, Binzhou 256603, China

2 Department of Pediatrics, Huantai County People's Hospital, Zibo 256400, China

3 Occupational Health Department, Binzhou Center for Disease Control and Prevention, BinZhou 256600, China

Corresponding author: REN R, E-mail: 865699206@qq.com

Conflicts of interest   None.

Received  2025-03-13
Accepted  2025-05-09
DOI: 10.12015/issn.1674-8034.2025.05.028
Cite this article as: WANG Z H, SONG Y R, SU H X, et al. Deep learning-based multimodal magnetic resonance imaging techniques and their research progress in depression diagnosis and treatment[J]. Chin J Magn Reson Imaging, 2025, 16(5): 184-189, 228 DOI:10.12015/issn.1674-8034.2025.05.028.

[1]
LI H, JIANG Z D, LI X L. Research progress of multimodal MRI in evaluating the rapid antidepressant effect of ketamine[J]. Chin J Magn Reson Imag, 2023, 14(4): 115-119. DOI: 10.12015/issn.1674-8034.2023.04.020.
[2]
MALHI G S, MANN J J. Depression[J]. Lancet, 2018, 392(10161): 2299-2312. DOI: 10.1016/S0140-6736(18)31948-2.
[3]
THAPAR A, EYRE O, PATEL V, et al. Depression in young people[J]. Lancet, 2022, 400(10352): 617-631. DOI: 10.1016/s0140-6736(22)01012-1.
[4]
PARK S H, YU H Y. How useful is the center for epidemiologic studies depression scale in screening for depression in adults? An updated systematic review and meta-analysis[J/OL]. Psychiatry Res, 2021, 302: 114037 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/34098160/. DOI: 10.1016/j.psychres.2021.114037.
[5]
PILMEYER J, LAMERICHS R, SCHIELEN S, et al. Multi-modal MRI for objective diagnosis and outcome prediction in depression[J/OL]. Neuroimage Clin, 2024, 44: 103682 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/39395373/. DOI: 10.1016/j.nicl.2024.103682.
[6]
CHU Z S, YUAN L J, LIAN K, et al. Reduced gray matter volume of the hippocampal tail in melancholic depression: evidence from an MRI study[J/OL]. BMC Psychiatry, 2024, 24(1): 183 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/38443878/. DOI: 10.1186/s12888-024-05630-5.
[7]
SUN F F, SHUAI Y F, WANG J R, et al. Hippocampal gray matter volume alterations in patients with first-episode and recurrent major depressive disorder and their associations with gene profiles[J/OL]. BMC Psychiatry, 2025, 25(1): 134 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/39955494/. DOI: 10.1186/s12888-025-06562-4.
[8]
BARCH D M, HUA X, KANDALA S, et al. White matter alterations associated with lifetime and current depression in adolescents: Evidence for Cingulum disruptions[J]. Depress Anxiety, 2022, 39(12): 881-890. DOI: 10.1002/da.23294.
[9]
GROGANS S E, FOX A S, SHACKMAN A J. The amygdala and depression: a sober reconsideration[J]. Am J Psychiatry, 2022, 179(7): 454-457. DOI: 10.1176/appi.ajp.20220412.
[10]
PILMEYER J, HUIJBERS W, LAMERICHS R, et al. Functional MRI in major depressive disorder: a review of findings, limitations, and future prospects[J]. J Neuroimaging, 2022, 32(4): 582-595. DOI: 10.1111/jon.13011.
[11]
CHAABENE S, BOUDAYA A, BOUAZIZ B, et al. An overview of methods and techniques in multimodal data fusion with application to healthcare[J/OL]. Int J Data Sci Anal, 2025 [2025-03-12]. https://link.springer.com/article/10.1007/s41060-025-00715-0. DOI: 10.1007/s41060-025-00715-0.
[12]
AHMED F, MAHMOOD T, RIAZ M, et al. Comprehensive review of high-dimensional monitoring methods: trends, insights, and interconnections[J/OL]. Qual Technol Quant Manag, 2024 [2025-03-12]. https://www.tandfonline.com/doi/full/10.1080/16843703.2024.2395745. DOI: 10.1080/16843703.2024.2395745.
[13]
LIAO L F, WANG O C, LIU Y. Current progress of machine learning combined with functional magnetic resonance imaging in depression's diagnosis[J]. Chin J Magn Reson Imag, 2021, 12(5): 107-109, 117. DOI: 10.12015/issn.1674-8034.2021.05.026.
[14]
QI N, ZHAO J. Research progresses of radiomics and machine learning based on MRI for diagnosing depressive disorder[J]. Chin J Med Imag Technol, 2024, 40(3): 455-458. DOI: 10.13929/j.issn.1003-3289.2024.03.028.
[15]
ARCHANA R, ELIAHIM JEEVARAJ P S. Deep learning models for digital image processing: a review[J/OL]. Artif Intell Rev, 2024, 57(1): 11 [2025-03-12]. https://link.springer.com/article/10.1007/s10462-023-10631-z. DOI: 10.1007/s10462-023-10631-z.
[16]
MAZUROWSKI M A, BUDA M, SAHA A, et al. Deep learning in radiology: an overview of the concepts and a survey of the state of the art with focus on MRI[J]. J Magn Reson Imaging, 2019, 49(4): 939-954. DOI: 10.1002/jmri.26534.
[17]
RASTOGI D, JOHRI P, DONELLI M, et al. Brain tumor detection and prediction in MRI images utilizing a fine-tuned transfer learning model integrated within deep learning frameworks[J/OL]. Life (Basel), 2025, 15(3): 327 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/40141673/. DOI: 10.3390/life15030327.
[18]
LEE S, CHO Y, JI Y, et al. Multimodal integration of neuroimaging and genetic data for the diagnosis of mood disorders based on computer vision models[J/OL]. J Psychiatr Res, 2024, 172: 144-155 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/38382238/. DOI: 10.1016/j.jpsychires.2024.02.036.
[19]
HU J L, LUO J M, XU Z Y, et al. Spatio-temporal learning and explaining for dynamic functional connectivity analysis: Application to depression[J/OL]. J Affect Disord, 2024, 364: 266-273 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/39137835/. DOI: 10.1016/j.jad.2024.08.014.
[20]
EL-LATIF A A A, CHELLOUG S A, ALABDULHAFITH M, et al. Accurate detection of Alzheimer's disease using lightweight deep learning model on MRI data[J/OL]. Diagnostics (Basel), 2023, 13(7): 1216 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/37046434/. DOI: 10.3390/diagnostics13071216.
[21]
LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553): 436-444. DOI: 10.1038/nature14539.
[22]
SALEHI A W, KHAN S, GUPTA G, et al. A study of CNN and transfer learning in medical imaging: advantages, challenges, future scope[J/OL]. Sustainability, 2023, 15(7): 5930 [2025-03-12]. https://www.mdpi.com/2071-1050/15/7/5930. DOI: 10.3390/su15075930.
[23]
RUBIN-FALCONE H, ZANDERIGO F, THAPA-CHHETRY B, et al. Pattern recognition of magnetic resonance imaging-based gray matter volume measurements classifies bipolar disorder and major depressive disorder[J/OL]. J Affect Disord, 2018, 227: 498-505 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/29156364/. DOI: 10.1016/j.jad.2017.11.043.
[24]
EL-ASSY A M, AMER H M, IBRAHIM H M, et al. A novel CNN architecture for accurate early detection and classification of Alzheimer's disease using MRI data[J/OL]. Sci Rep, 2024, 14: 3463 [2025-03-12]. https://www.nature.com/articles/s41598-024-53733-6. DOI: 10.1038/s41598-024-53733-6.
[25]
WANG Y, GONG N, FU C Y. Major depression disorder diagnosis and analysis based on structural magnetic resonance imaging and deep learning[J]. J Integr Neurosci, 2021, 20(4): 977-984. DOI: 10.31083/j.jin2004098.
[26]
ZHANG X L, CAO J, CHEN X R, et al. Neural correlates of severity in major depressive disorder: a combined structural and resting-state functional MRI study[J/OL]. Biomark Neuropsychiatry, 2023, 9: 100072 [2025-03-12]. https://www.sciencedirect.com/science/article/pii/S2666144623000126. DOI: 10.1016/j.bionps.2023.100072.
[27]
HONG J L, HUANG Y Q, YE J M, et al. 3D FRN-ResNet: an automated major depressive disorder structural magnetic resonance imaging data identification framework[J/OL]. Front Aging Neurosci, 2022, 14: 912283 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/35645776/. DOI: 10.3389/fnagi.2022.912283.
[28]
LI R P, HUANG Y Q, WANG Y B, et al. MRI-based deep learning for differentiating between bipolar and major depressive disorders[J/OL]. Psychiatry Res Neuroimaging, 2024, 345: 111907 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/39357171/. DOI: 10.1016/j.pscychresns.2024.111907.
[29]
HU J L, HUANG Y M, ZHANG X J, et al. Identifying suicide attempts, ideation, and non-ideation in major depressive disorder from structural MRI data using deep learning[J/OL]. Asian J Psychiatry, 2023, 82: 103511 [2025-03-12]. https://www.sciencedirect.com/science/article/abs/pii/S1876201823000655. DOI: 10.1016/j.ajp.2023.103511.
[30]
HARRIS M A, COX S R, DE NOOIJ L, et al. Structural neuroimaging measures and lifetime depression across levels of phenotyping in UK biobank[J/OL]. Transl Psychiatry, 2022, 12(1): 157 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/35418197/. DOI: 10.1038/s41398-022-01926-w.
[31]
HONG S, LIU Y S, CAO B, et al. Identification of suicidality in adolescent major depressive disorder patients using sMRI: a machine learning approach[J]. J Affect Disord, 2021, 280(Pt A): 72-76. DOI: 10.1016/j.jad.2020.10.077.
[32]
TZOURIO-MAZOYER N, LANDEAU B, PAPATHANASSIOU D, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain[J]. Neuroimage, 2002, 15(1): 273-289. DOI: 10.1006/nimg.2001.0978.
[33]
BIANCHI F M, GRATTAROLA D, LIVI L, et al. Graph neural networks with convolutional ARMA filters[J]. IEEE Trans Pattern Anal Mach Intell, 2022, 44(7): 3496-3507. DOI: 10.1109/TPAMI.2021.3054830.
[34]
SHAHHOSSEINI Y, MIRANDA M F. Functional connectivity methods and their applications in fMRI data[J/OL]. Entropy (Basel), 2022, 24(3): 390 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/35327901/. DOI: 10.3390/e24030390.
[35]
SUNIL G, GOWTHAM S, BOSE A, et al. Graph neural network and machine learning analysis of functional neuroimaging for understanding schizophrenia[J/OL]. BMC Neurosci, 2024, 25(1): 2 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/38166747/. DOI: 10.1186/s12868-023-00841-0.
[36]
QIN K, LEI D, PINAYA W H L, et al. Using graph convolutional network to characterize individuals with major depressive disorder across multiple imaging sites[J/OL]. EBioMedicine, 2022, 78: 103977 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/35367775/. DOI: 10.1016/j.ebiom.2022.103977.
[37]
CHAO-GAN YAN, CHEN X, LI L, et al. Data of the REST-meta-MDD project from DIRECT ConsortiumData of the REST-meta-MDD project from DIRECT consortium[Z/OL]. Science Data Bank, 2022 [2025-04-09]. https://www.scidb.cn/en/detaildataSetId=cbeb3c7124bf47a6af7b3236a3aaf3a8. DOI: 10.57760/SCIENCEDB.O00115.00013.
[38]
KAISER R H, ANDREWS-HANNA J R, WAGER T D, et al. Large-scale network dysfunction in major depressive disorder: a meta-analysis of resting-state functional connectivity[J]. JAMA Psychiatry, 2015, 72(6): 603-611. DOI: 10.1001/jamapsychiatry.2015.0071.
[39]
LISTON C, CHEN A C, ZEBLEY B D, et al. Default mode network mechanisms of transcranial magnetic stimulation in depression[J]. Biol Psychiatry, 2014, 76(7): 517-526. DOI: 10.1016/j.biopsych.2014.01.023.
[40]
MULDERS P C, VAN EIJNDHOVEN P F, SCHENE A H, et al. Resting-state functional connectivity in major depressive disorder: a review[J/OL]. Neurosci Biobehav Rev, 2015, 56: 330-344 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/26234819/. DOI: 10.1016/j.neubiorev.2015.07.014.
[41]
Anonymous. Resting-state functional MRI in depression unmasks increased connectivity between networks via the dorsal nexus - PMC[EB/OL]. [2025-02-24]. https://pmc.ncbi.nlm.nih.gov/articles/PMC2890754/. DOI: 10.1073/pnas.100044610.
[42]
MA W Z, WANG Y, MA N X, et al. Diagnosis of major depressive disorder using a novel interpretable GCN model based on resting state fMRI[J/OL]. Neuroscience, 2025, 566: 124-131 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/39730018/. DOI: 10.1016/j.neuroscience.2024.12.045.
[43]
ZAHEDI A, ARTIGAS S O, SWABODA N, et al. Neural correlates of changing food choices while bypassing values[J/OL]. Neuroimage, 2023, 274: 120134 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/37100103/. DOI: 10.1016/j.neuroimage.2023.120134.
[44]
ZHOU N, YUAN Z, ZHOU H Y, et al. Using dynamic graph convolutional network to identify individuals with major depression disorder[J/OL]. J Affect Disord, 2025, 371: 188-195 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/39566747/. DOI: 10.1016/j.jad.2024.11.035.
[45]
SUN J N, YI W, HUANG J L, et al. Magnetic resonance imaging research progress of effective brain network connectivity in patients with major depressive disorder[J]. Chin J Magn Reson Imag, 2024, 15(11): 153-159. DOI: 10.12015/issn.1674-8034.2024.11.024.
[46]
JUN E, NA K S, KANG W, et al. Identifying resting-state effective connectivity abnormalities in drug-naïve major depressive disorder diagnosis via graph convolutional networks[J]. Hum Brain Mapp, 2020, 41(17): 4997-5014. DOI: 10.1002/hbm.25175.
[47]
VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[Z/OL]. arXiv, 2023 [2025-04-13]. http://arxiv.org/abs/1706.03762. DOI: 10.48550/arXiv.1706.03762.
[48]
LOGOTHETIS N K, PAULS J, AUGATH M, et al. Neurophysiological investigation of the basis of the fMRI signal [J/OL]. Nature, 2001. [2025-02-24]. https://www.nature.com/articles/35084005. DOI: 10.1038/35084005.
[49]
HUANG L, MAO F, ZHANG K, et al. Spatial-temporal convolutional transformer network for multivariate time series forecasting[J/OL]. Sensors (Basel), 2022, 22(3): 841 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/35161585/. DOI: 10.3390/s22030841.
[50]
ZHENG G W, ZHENG W H, ZHANG Y, et al. An attention-based multi-modal MRI fusion model for major depressive disorder diagnosis[J/OL]. J Neural Eng, 2023, 20(6) [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/37844568/. DOI: 10.1088/1741-2552/ad038c.
[51]
ABBAOUI W, RETAL S, EL BHIRI B, et al. Towards revolutionizing precision healthcare: a systematic literature review of artificial intelligence methods in precision medicine[J/OL]. Inform Med Unlocked, 2024, 46: 101475 [2025-03-12]. https://www.sciencedirect.com/science/article/pii/S2352914824000315. DOI: 10.1016/j.imu.2024.101475.
[52]
INSEL T R, CUTHBERT B N. Brain disorders? precisely[J]. Science, 2015, 348(6234): 499-500. DOI: 10.1126/science.aab2358.
[53]
ZHANG W, ZENG W M, CHEN H Y, et al. STANet: a novel spatio-temporal aggregation network for depression classification with small and unbalanced FMRI data[J]. Tomography, 2024, 10(12): 1895-1914. DOI: 10.3390/tomography10120138.
[54]
WANG L, MA Q, SUN X Y, et al. Frequency-resolved connectome alterations in major depressive disorder: a multisite resting fMRI study[J/OL]. J Affect Disord, 2023, 328: 47-57 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/36781144/. DOI: 10.1016/j.jad.2023.01.104.
[55]
DABLAIN D, KRAWCZYK B, CHAWLA N V. DeepSMOTE: fusing deep learning and SMOTE for imbalanced data[J]. IEEE Trans Neural Netw Learn Syst, 2023, 34(9): 6390-6404. DOI: 10.1109/TNNLS.2021.3136503.
[56]
Anonymous. Power failure: Why small sample size undermines the reliability of neuroscience | nature reviews neuroscience[EB/OL]. [2025-02-25]. https://www.nature.com/articles/nrn3475. DOI: 10.1038/nrn3475
[57]
WANG J Y, LI T T, SUN Q, et al. Automatic diagnosis of major depressive disorder using a high- and low-frequency feature fusion framework[J/OL]. Brain Sci, 2023, 13(11): 1590 [2025-03-12]. https://pubmed.ncbi.nlm.nih.gov/38002549/. DOI: 10.3390/brainsci13111590.

PREV Combined multi-b-value DWI and DCE distributed parameter model in diagnosing radiation necrosis: One case report
NEXT Advances of medical imaging in Moyamoya disease
  


LINK


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