Table 3.

Summary of AD-relevant effects based on candidate imaging biomarkers and association studies

Author	Dataset	Genes included	Model	Methods	Imaging phenotypes	Neural location	Results
Pathophysiological pathway: Brain Aβ accumulation (Aβ PET)
2009 Drzezga et al. [70]	32 AD	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Aβ plaque deposition	Bilateral temporoparietal, frontal cortex	The ɛ4-positive patients with AD had higher levels of Aβ plaque deposition compared to age-matched ɛ4-negative patients with similar levels of cognitive impairment and brain atrophy
2009 Reiman et al. [71]	28 AD	APOE	Univariate imaging—Univariate genetic	Candidate-based association	PiB DVR fibrillar Aβ burden	Frontal, temporal, parietal, posterior cingulate-precuneus,basal ganglia ROIs	Fibrillar Aβ burden in cognitively normal older people was associated with APOE ɛ4 gene dose
2011 Chibnik et al. [72]	n = 1666	CR1, CLU, PICALM	Univariate imaging—Multivariate genetic	Candidate-based association	Pathology score of neuritic plaques	Whole brain cortex	Common variation at the CR1 locus had a broad impact on cognition and this effect was mediated by an individual’s amyloid plaque burden
2012 Thambisetty et al. [73]	57 HC	CR1, APOE	Univariate imaging—Multivariate genetic	Candidate-based association	PIB DVR	Orbitofrontal, prefrontal, superior frontal, posterior cingulate, lateral temporal, occipital cortices	There was a greater variability in brain amyloid deposition in the CLU rs3818361 noncarrier group relative to risk carriers, an effect explained partly by APOE genotype
2012 Swaminathan et al. [74]	ADNI (22 HC, 25 AD, 56 MCI)	15 amyloid candidate genes (DNCR24, NCSTN, SOAT1, BCHE, etc.)	Multivariate imaging—Multivariate genetic	Candidate-based association	Normalized PiB uptake value	Anterior cingulate, frontal cortex, parietal cortex, precuneus	The minor allele of an intronic SNP within DHCR24 was identified and associated with a lower average PiB uptake, and non-carriers of the minor allele had higher PiB uptake in frontal regions compared to carriers
2013 Shulman et al. [75]	Multiple cohorts (n = 725/ 56/58)	ABCA7, MS4A6A/MS4A4E, EPHA1, CD3, CR1, CD2AP, CLU, BIN1, PICALM	Univariate imaging—Multivariate genetic	Candidate-based association	Pathology score of neuritic plaques	Midfrontal, middle temporal, inferior parietal, entorhinal, hippocampal cortex	Besides the previously reported APOE and CR1 loci, ABCA7 (rs3764650) and CD2AP (rs9349407) were associated with neuritic plaque burden
2013 Shulman et al. [75]	Multiple cohorts (n = 725/ 56/58)	Genome-wide genotyping	Univariate imaging—Multivariate genetic	GWAS	Pathology score of neuritic plaques	Midfrontal, middle temporal, inferior parietal, entorhinal, hippocampal cortex	The finding discovered a novel variant near the amyloid precursor protein gene (APP, rs2829887) that is associated with neuritic plaques
2013 Hohman et al. [76]	ADNI (174 HC, 64 AD, 292 MCI)	PICALM, BIN1, CR1, CLU, MS4A6A, EPHA1, CD33, ABCA7, CD2AP	Multivariate imaging—Univariate genetic	Candidate-based association	Aβ PET SUVR	Cingulate, frontal, temporal, lateral parietal cortices	Two SNP-SNP interactions (BIN1 (rs7561528, rs744373) × PICALM (rs7851179)) reached significance when correcting for multiple comparisons
2014 Lehmann et al. [77]	52 AD	APOE	Multivariate imaging—Univariate genetic	Candidate-based association	PIB DVR, FDG SUVR	Frontal, lateral parietal/temporal, occipital cortices, precuneus, posterior cingulate gyrus, hippocampus	APOE ε4+ AD patients showed lower global amyloid burden and greater medial temporal hypometabolism compared with matched APOE  ε4- patients
2014 Ramanan et al. [78]	ADNI (n = 555)	Genome-wide genotyping	Univariate imaging—Multivariate genetic	GWAS	Aβ PET brain amyloid burden	Frontal, parietal, temporal, limbic, occipital lobes	A novel association with higher rates of amyloid load independent from APOE ε4 status was identified in IL1RAP (rs12053868-G)
2018 Apostolova et al. [17]	ADNI (322 HC, 159 AD, 496 MCI)	The top 20 AD risk variants (ABCA7,CLU, SORL1, DSG2, etc.)	Univariate imaging—Multivariate genetic	Candidate-based association	Florbetapir mean SUVR	Frontal, anterior–posterior cingulate, lateral-parietal, lateral-temporal cortices	ABCA7 gene had the strongest association with amyloid deposition, after APOE ε4. FERMT2 gene had a stage-dependent association with brain amyloidosis
2018 Scelsi et al. [79]	ADNI (226 HC, 125 AD, 92 SMC, 501 MCI)	Genome-wide genotyping	Multivariate imaging—Multivariate genetic	PGS-based association	Aβ PET SUVR, HV	Hippocampus	The finding identified a genome-wide significant locus implicating LCORL rs6850306. The possession of a minor allele at rs6850306 was protective against conversion from MCI to AD
2019 Li et al. [80]	ADNI (155 HC, 125 AD, 72 SMC, 422 MCI)	Genome-wide genotyping	Univariate imaging—Multivariate genetic	GWAS	Florbetapir composite SUVR	Frontal, anterior/ posterior cingulate, lateral parietal/ temporal regions	The study identified 24 consensus modules enriched by robust genetic signals in a genome wide association analysis, including a few novel genes (ABL1, ABLIM2)
2021 Kim et al [81]	Korean cohort (n = 1474)	Genome- wide genotyping	Univariate imaging—multivariate genetic	GWAS	Aβ PET SUVR	Whole brain	In addition to APOE, nine SNPs of FGL2 gene on chromosome 7 were identified, which were associated with a decreased risk of Aβ positivity at a genome-wide suggestive level
2021 Liu et al. [82]	Multiple cohorts (n = 767/ 1373)	Summary statistics	Multivariate imaging—Multivariate genetic	PGS-based association	Aβ PET SUVR, HV, entorhinal, middle temporal gyrus volumes	Whole brain cortex, Hippocampus, entorhinal cortex	PGS was associated with the increased cortical amyloid burdens (PiB-PET and AV45-PET), but decreased hippocampus and entorhinal cortex volumes
Pathophysiological pathway: Tau hyperphosphorylation (Tau PET)
2016 Smith et al. [83]	4 HC, 3 AD	MAPT	Univariate imaging—Univariate genetic	Candidate-based association	Tau PET SUVR, GM volume	Global AD pathology	18F-AV1451 tau PET imaging correlated with tau pathology in MAPT mutation carriers
2018 Mattsson et al. [84]	65 Aβ + patients	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Tau PET SUVR, GM volume	Parietal, entorhinal cortex	APOE ε4-negative patients had greater tau load and reduced cortical thickness, with the most pronounced effects for both in the parietal cortex
2019 Shen et al. [85]	ADNI (90 HC)	MAPT rs242557	Univariate imaging—Univariate genetic	Candidate-based association	Tau PET SUVR	Hippocampus	The finding confirmed the significant correlation of MAPT rs242557 risk variant with increased hippocampus tau burden in non-demented elders
2019 Therriaultet al. [86]	Multiple cohorts (281 HC, 75 AD, 133 MCI)	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Tau PET SUVR	Entorhinal cortex, hippocampus	The elevated risk of developing dementia conferred by APOE ε4 genotype involved mechanisms associated with both Aβ and tau aggregation
2019 Franzmeier et al. [87]	ADNI (49 HC, 40 MCI)	BIN1 rs744373	Univariate imaging—Univariate genetic	Candidate-based association	Global/stage- specific Tau PET SUVR	Brain Braak stage II–VI	BIN1 rs744373 SNP was associated with increased tau but not Aβ pathology, that is alterations in BIN1 may contribute to memory deficits via increased tau pathology
2020 Yan et al. [88]	ADNI (57 AD)	APOE	Multivariate imaging—Univariate genetic	Candidate-based association	Tau PET SUVR, GM volume	Temporal, parietal, posterior cingulate, entorhinal cortex, amygdala, parahippocampal gyrus, etc	Among elderly individuals with AD, sex modified the effects of the APOE ε4 allele on region-specific tau deposition and GM volume
2020 Neitzel et al. [89]	Multiple cohorts (n = 493)	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Baseline Tau PET SUVR, annual change rates	MTL (entorhinal cortex, parahippocampus)	There was an amyloid-independent association between APOE ε4 and elevated tau PET specifically in medial temporal regions
2021 Franzmeier et al. [90]	Multiple cohorts (n = 216)	BIN1 rs744373	Univariate imaging—Univariate genetic	Candidate-based association	ROI Tau PET SUVR, annual change rates	Whole brain	BIN1-associated AD risk was potentially driven by accelerated tau accumulation in the face of Aβ
2021 Neitzel et al. [91]	ADNI (347 HC, 48 AD, 156 MCI)	Klotho-VShet	Multivariate imaging—Univariate genetic	Candidate-based association	Global/ROI tau/Aβ PET SUVR	Whole brain; bilateral inferior temporal gyri	Findings proved a protective role of KL-VShet against amyloid-related tau pathology and tau-related memory impairments in elderly humans at risk of AD dementia
2021 Sun et al. [92]	ADNI (n = 158)	Summary statistics	Multivariate imaging—Multivariate genetic	PGS-based association	Global tau SUVR for Braak stage ROIs	Whole brain	The association between PGS and tau pathology was significant when APOE was excluded, even among females
Pathophysiological pathway: Neurodegeneration (sMRI)
2007 Lunetta et al. [93]	449 HC, 366 AD	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Cerebral atrophy, MTA, WMH, CVR	Cerebral atrophy, MTA, WMH	A substantial proportion of the additive genetic variation in MRI traits was explained by other genes, and MRI traits were heritable
2009 Potkin et al. [94]	ADNI (n = 381)	Genome-wide genotyping	Univariate imaging—Multivariate genetic	GWAS	GM voxels of hippocampal regions	The right and left hippocampal regions	The study identified candidate risk genes (EFNA5, CAND1, MAGI2, ARSB, and PRUNE2) for sporadic AD, involved in the regulation of protein degradation, apoptosis, neuronal loss and neurodevelopment
2010 Wolk et al. [95]	ADNI (91 AD)	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Cortical thickness, HV	Hippocampus, superior frontal gyrus,angular gyrus, MTL, precentral gyrus	The presence or absence of the APOE ε4 allele influenced the cognitive and anatomic phenotypic expression of AD in a dissociable manner
2010 Biffi et al. [96]	Multiple cohorts (215 HC, 168 AD, 357 MCI)	GWAS-validated and GWAS-promising novel AD loci	Univariate imaging—Multivariate genetic	Candidate-based association	HV, amygdala volume, WM lesion volume, parahippocampal, entorhinal, temporal pole cortex thickness	Hippocampal, parahippocampal gyrus, amygdala, entorhinal, temporal pole cortex	Loci associated with AD influenced neuroimaging correlates of this disease. And neuroimaging analysis identified 2 additional loci (BIN1 and CNTN5) of high interest for further study
2013 Meda et al. [97]	ADNI (156 HC, 140 AD, 281 MCI)	151 million SNPs within 212 KEGG pathways	Univariate imaging—Multivariate genetic	Candidate-based association	12-month regional structural atrophy rates	Hippocampus, entorhinal cortex	A total of 109 SNP-SNP interactions were associated with right hippocampus atrophy, and 125 were associated with right entorhinal cortex atrophy
2013 Jahanshad et al. [98]	366 HC	SPON1 gene	multivariate imaging—multivariate genetic	Candidate-based association	Heritable brain connections	Maps of the brain’s structural connectome	Rs2618516 was shown to affect brain structure in an elderly population with varying degrees of dementia
2014 Morgen et al. [99]	165 AD	PICALM, APOE	Univariate imaging—Multivariate genetic	Candidate-based association	GM volume	Prefrontal cortex	There was a synergistic adverse effect of homozygosity for the PICALM risk allele G in rs3851179 and APOE ε4 on prefrontal volume and performance on the Trail Making Test A, which is sensitive to processing speed and working memory function
2014 Hohman et al. [100]	ADNI (388 HC, 228 AD, 764 MCI)	Genome-wide genotyping	Univariate imaging—multivariate genetic	GWAS	Baseline ICV	Whole brain	One intergenic SNP rs4866650 and one SNP rs7849530 within the SPTLC1 gene modified the association between amyloid positivity and neurodegeneration
2015 Chauhan et al. [101]	8175– 11,550 HC	24 AD candidate loci (APOE, BIN1,HLA-DRB1,CR33,CR1,CLU, ABCA7, SORL1, etc.)	Multivariate imaging—Multivariate genetic	Meta- analysis	ICV, TBV, HV, WMH	Hippocampus	APOE rs2075650 was associated with smaller HV and CD33 rs3865444 with smaller ICV. There was associations of HLA-DRB1 with TBV and BIN1 with HV. A weighted AD genetic risk score was associated with smaller HV, even after excluding APOE locus
2015 Desikan et al. [102]	9386 HC, 6409 AD	Summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Longitudinal volume loss in MTL, entorhinal cortex, hippocampus	MTL, hippocampus, entorhinal cortex	Polygenic hazard scores predicted in vivo markers (volume loss in MTL, hippocampus, entorhinal cortex)
2016 Yang et al. [103]	ADNI (194 HC, 168 AD, 337 MCI)	PICALM, CLU	Univariate imaging—Multivariate genetic	Candidate-based association	HV, hippocampal shape	Hippocampus	Common LOAD risk loci in CLU and PICALM exhibited significant interaction effects on hippocampal morphology in both young healthy adults and elderly individuals
2016 Ramirez et al. [104]	50 HC, 98 MCI	the top 10 AD non-APOE genes	Univariate imaging—Multivariate genetic	Candidate-based association	Cortical thickness, hippocampal radial distance	Hippocampus	MS4A6A rs610932 and ABCA7 rs3764650 demonstrated significant associations with cortical and hippocampal atrophy
2016 Habes et al. [105]	n = 1472	APOE	Univariate imaging—Univariate genetic	Candidate-based association	AD-related GM volume	Lateral frontal, lateral temporal, medial frontal cortex, hippocampus	Measurable APOE-related brain atrophy did not occur in early adulthood and midlife and such atrophy may only occur more proximal to the onset of clinical symptoms of dementia
2016 Foley et al. [106]	n = 272	APOE, summary statistics	Multivariate imaging—Multivariat genetic	PGS-based association	HV	Hippocampus	A significant association was found between AD PGS and left HV, with higher risk associated with lower left HV, although excluding the APOE gene
2016 Harrison et al. [107]	n = 66	Summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Thickness in hippocampal subregions	Hippocampus, entorhinal cortex	Polygenic AD risk scores may be especially sensitive to structural change over time in regions affected early in AD, like the hippocampus and adjacent entorhinal cortex
2017 Wang et al. [108]	ADNI (281 HC, 48 AD, 483 MCI)	12 SNPs in HLA	Univariate imaging—Multivariate genetic	Candidate-based association	Structural volumes	Hippocampus, parahippocampus, posterior cingulate, middle temporal, etc	TNF-α SNPs at rs2534672, rs2395488, HFE rs1800562 and RAGE rs2070600 were correlated with various structures on MRI
2017 Wang et al. [109]	ADNI (281 HC, 48 AD, 483 MCI)	HLA-A2	Univariate imaging—Univariate genetic	Candidate-based association	Hippocampal/ parahippocampal/ amygdala/ middle temporal/ posterior cingulate volume, entorhinal cortex thickness	Hippocampus, parahippocampus, posterior cingulate, precuneus, middle temporal, entorhinal cortex, amygdala	HLA-A2 in Caucasians contributed to the risk of AD by modulating the alteration of HV and HLA-A gene variants appeared to play a role in altering AD-related brain structures on MRI
2017 Xiao et al. [110]	n = 231	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Activation in hippocampus ROI	Hippocampus	There was a cumulative deleterious effect of LOAD risk genes on hippocampal function even in healthy volunteers
2018 Axelrud et al. [111]	Multiple cohorts	Summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	HV	Left and right hippocampus	Genetic risk for AD may affect early-life cognition and HV
2018 Li et al. [112]	Multiple cohorts (n = 683)	Summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	GM volume	Precuneal cortex	An elevated AD PGR was associated with a smaller precuneal volume, and the effect remained after excluding the APOE genotype
2019 Lancaster et al. [113]	Multiple cohorts	AD SNPs within a microglia-mediated immunity network	Univariate imaging—Multivariate genetic	PGS-based association	HV	Hippocampus	The observations suggested that the relationship between AD and HV was partially explained by genes within an AD-linked microglia-mediated immunity network
2020 Lyall et al. [114]	UK Biobank (n = 8539)	APOE	Multivariate imaging—Univariate genetic	Candidate-based association	FA, MD, left/right HV, total GM, total WM and log WMHV	Left or right Hippocampus, total GM and WM	There was association between APOE ε4 and WMHV, but not TBV or WM integrity
2020 Cong et al. [115]	ADNI (41 HC, 26 AD, 67 MCI)	Genome- wide genotyping	Univariate imaging—Multivariate genetic	GWAS	14 MTL substructures	MTL	A novel association with right Brodmann area 36 volume was discovered in an ERC1 SNP rs2968869. And rs2968869 was associated with GM density and glucose metabolism in the right hippocampus and disease status
2020 De Marco et al. [116]	ADNI (317 HC, 562 MCI)	Summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	GM and WM volumes	Whole brain	PGS predicted volume in sensorimotor regions in ε3ε3 Aβ + participants. The link between polygenic hazard and neurocognitive variables varies depending on APOE ε4 allele status
2020 van der Meer et al. [117]	Multiple cohorts (n = 21,297)	Genome-wide genotyping	Univariate imaging—multivariate genetic	GWAS	Hippocampal and subfield volumes	Hippocampus	GWAS of whole HV identified eight whole-genome significant loci, including three novel loci, namely, TFDP2 SNP rs7630893, FAM175B rs2303611, and PARP11 rs1419859
2021 Foo et al. [118]	UK Biobank (n =  17,161)	Summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Volumes in hippocampal subregions	Multiple hippocampal regions	PGSAD had differential effects on the hippocampal subfield volumes
2021 Tank et al. [119]	UK Biobank (n = 32,790)	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Volumes of total GM, WM, WMH, whole brain, left/ right hippocampus	Left hippocampus	LOAD-PGR was associated with smaller HV and aspects of cognitive ability in healthy adults and could supplement APOE status in risk stratification of cognitive impairment/LOAD
Pathophysiological pathway: Neurodegeneration (FDG PET)
2010 Corneveaux et al. [120]	Multiple cohort (n = 1728)	KIBRA rs17070145	Univariate imaging—Univariate genetic	Candidate-based association	Glucose metabolism	Entorhinal cortex, hippocampus, middle temporal gyrus, posterior cingulate cortex, superior frontal gyrus, primary visual cortex	Non-carriers of the KIBRA rs17070145-T had increased risk of LOAD in an association study of 702 neuropathologically verified expired subjects and in a combined analysis of 1026 additional living and expired subjects
2014 Lehmann et al. [77]	52 AD	APOE	Multivariate imaging—Univariate genetic	Candidate-based association	PIB DVR, FDG SUVR	Lateral temporoparietal cortex, precuneus, posterior cingulate cortex, middle frontal gyrus, etc	APOE  ε4+ AD patients showed lower global amyloid burden and greater medial temporal hypometabolism compared with matched APOE  ε4- patients
2018 Miller et al. [121]	ADNI (n = 695)	EXOC3L4	Univariate imaging—Multivariate genetic	WGS	Global cortical glucose metabolism	Whole brain cortex	EXOC3L4 gene, was identified as significantly associated with global cortical glucose metabolism. Three loci that may affect splicing within EXOC3L4 helped to the association
2018 Kong et al. [122]	ADNI (37 HC, 59 AD, 126 MCI)	Genome-wide genotyping	Univariate imaging—Univariate genetic	GWAS	ROI glucose metabolic uptake	Left and right angular, temporal gyri, bilateral posterior cingulate	A genome-wide significant SNP rs12444565 in the RBFOX1, four suggestive loci (rs235141, rs79037, rs12526331 and rs12529764) were associated with 18F-FDG
2020 Seo et al. [123]	KBASE (336 HC, 84 AD, 136 MCI)	132 AD candidate genes	Multivariate imaging—Multivariate genetic	Candidate-based association	Aβ deposition, region cerebral glucose metabolism/ cortical thickness, HV	AD-signature cortical, hippocampus	Several novel loci for common variants were associated with AD pathology (PIWIL1, NME8 and PSEN2, PSEN1, CASS4). Cases carrying rare variants in LPL, FERMT2, NFAT5, DSG2, and ITPR1 displayed associations with the neuroimaging features
2021 Wang et al. [124]	ADNI (n = 586)	Genome- wide genotyping	Univariate imaging—Multivariate genetic	GWAS	Glucose metabolic uptake in ROIs	Left angular gyri, bilateral posterior cingulate gyrus, right /left middle/inferior temporal gyrus	Two genome-wide significant SNPs (rs4819351, rs13387360) in AGPAT3 and LOC101928196 served as protective sites to regulate the decline of glucose metabolism
2019 Li et al. [80]	ADNI (37 HC, 86 AD, 188 MCI)	Genome-wide genotyping	Univariate imaging—Multivariate genetic	GWAS	Glucose metabolic uptake in ROIs	Frontal, lateral parietal, lateral temporal regions, anterior/posterior cingulate regions	Indirect genetic effects on certain chemical compound or protein translocation were reflected in the PET scans and may be associated with AD
Pathophysiological pathway: Neurodegeneration (fMRI)
2000 Bookheimer et al. [125]	30 HC	APOE	Univariate imaging—Univariate genetic	Candidate-based association	Patterns of brain activation	Left hippocampal, parietal, prefrontal cortices	Both the magnitude and the extent of brain activation during memory-activation tasks in regions of the left hippocampal, parietal, and prefrontal regions, were greater among the carriers of the APOE ε4 allele than among the carriers of the APOE ɛ3 allele
2011 Erk et al. [126]	109 HC	CLU rs11136000	Univariate imaging—Univariate genetic	Candidate-based association	FC	Hippocampus, prefrontal cortex	Healthy carriers of the variant exhibited altered coupling between hippocampus and prefrontal cortex during memory processing
2011 Lancaster et al. [127]	43 HC	CLU rs11136000	Univariate imaging—Univariate genetic	Candidate-based association	Working memory values based on brain activity	Frontal, posterior cingulate cortex, hippocampus	Participants with the CLU risk genotype had higher activity than participants with the protective allele in frontal and posterior cingulate cortex and hippocampus
2014 Green et al. [128]	131 HC	APOE, CLU	Univariate imaging—Multivariate genetic	Candidate-based association	ROI BOLD signal change	Hippocampus, MTL	APOE ε4 and CLU-C had an additive effect on brain activity, that is, increased combined genetic risk was associated with decreased brain activity during executive attention, including in the MTL
2014 Guerini et al. [129]	n = 1680	SNAP-25 SNP	Univariate imaging—Univariate genetic	Candidate-based association	FMRI task accuracy	Cingulate cortex, frontal, temporoparietal cortices	FMRI analyses indicated that SNAP-25 genotypes correlated with a significantly decreased brain activity in the cingulate cortex and in the frontal (middle, superior gyri) and the temporo-parietal (angular gyrus) area
2014 Liu et al. [130]	Han Chinese (21 HC, 46 MCI)	TOMM40 rs157581	Univariate imaging—Univariate genetic	Candidate-based association	ALFF	Bilateral superior frontal gyrus, bilateral lingual gyrus, right calcarine sulcus, left cerebellar	TOMM40 rs157581 polymorphism may modulate regional spontaneous brain activity and relate to the progression of aMCI
2015 Lancaster et al. [131]	85 HC	CLU rs11136000	Multivariate imaging—Univariate genetic	Candidate-based association	Working memory task accuracy, GM density	Hippocampus, prefrontal, limbic areas	Young individuals with the CLU rs11136000-C had higher activation levels in prefrontal and limbic areas during a working memory task. And there were subtle reductions in GM in the right hippocampal formation in carriers of the risk variant
2015 Zhang et al. [132]	360 HC	BIN1 rs744373	Multivariate imaging—Univariate genetic	Candidate-based association	Working memory, GM volume, FC	Whole brain	Healthy homozygous carriers of the rs744373 risk allele exhibited worse high-load working memory performance, larger HV and lower FC between the bilateral hippocampus and right dorsolateral prefrontal cortex
2017 Sun et al. [133]	32 HC, 32 MCI	PICALM rs3851179	Univariate imaging—Univariate genetic	Candidate-based association	FC	DMN	The PICALM rs3851179 polymorphism significantly affected the DMN network in MCI
2017 Xiao et al. [110]	n = 231	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Activation in hippocampus ROI	Hippocampus	There was a cumulative deleterious effect of LOAD risk genes on hippocampal function even in healthy volunteers
2017 Su et al. [134]	131 HC, 87 MCI	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	FC in ROIs of DMN	Temporal cortex	The pMCIs exhibited tremendous decrements in DMN connections that were partially determined by the AD-related risk alleles
2018 Korthauer et al. [135]	76 HC	APOE	Multivariate imaging—Univariate genetic	Candidate-based association	Graph analysis of network efficiency	Whole brain functional-structural network	ε4 carriers had significantly lower global and local efficiency of the integrated resting-state structural connectome compared to non-carriers
2021 Franzmeier et al. [136]	Multiple cohort (n = 378)	BDNFVal66Met SNP	Univariate imaging—Univariate genetic	Candidate-based association	FC	DMN, DAN, SAL, CON	BDNFVal66Met was associated with a higher vulnerability of hippocampus-frontal connectivity to primary AD pathology
2019 Chandler et al. [137]	n = 75	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Whole-brain gmCBF	Frontal cortex	The results found a reduction in gmCBF in APOE ε4 carriers, a negative relationship between AD-PGS and gmCBF, and regional reductions in gmCBF in individuals with higher AD-PGS across the frontal cortex
2019 Axelrud et al. [138]	Multiple cohorts (n = 636)	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	FC among main nodes for 10 tau pathology networks	Precuneus, superior temporal gyrus	The PGS was associated with the connectivity between the right precuneus and the right superior temporal gyrus
2020 Chandler et al. [139]	n = 608	APOE, summary statistics	Univariate imaging—Multivariate genetic	PGS-based association	Bilateral hippocampus bold parameters	Hippocampus	AD-PGS, not APOE, selectively influenced activity within the HC in response to scenes, while other perceptual nodes remained intact
Pathophysiological pathway: Neurodegeneration (DTI)
2010 Smith et al. [140]	23 HC, 42 AD	APOE	Univariate imaging—Univariate genetic	Candidate-based association	FA	Inferior temporal lobe, amygdala/ hippocampal head region	Reduced FA was observed in the fronto-occipital and inferior temporal fasciculi (particularly posteriorly), the splenium of the corpus callosum, subcallosal white matter and the cingulum bundle
2005 Nierenberg et al. [141]	29 HC	APOE	Univariate imaging—Univariate genetic	Candidate-based association	FA, axD, radD	Parahippocampal region	The APOE ε4 carriers showed significantly lower fractional anisotropy and higher radial diffusivity in the parahippocampal WM 15 mm below the anterior commissure-posterior commissure plane than noncarriers
2014 Warstadt et al. [142]	n = 481	Genome-wide genotyping	multivariate imaging—multivariate genetic	GWAS	Diffusion tensor, FA	Corpus callosum, fornix, internal capsule, inferior fronto-occipital fasciculus	A follow-up analysis detected WM associations with rs5882 in the opposite direction
2015 Liang et al. [143]	126 HC	SORL1 rs2070045	Univariate imaging—Univariate genetic	Candidate-based association	FA, MD, axD, radD	Bilateral cingulum, cingulum hippocampal area	Sex moderated the effects of the SOR1 gene rs2070045 polymorphism on cognitive impairment and disruption of the cingulum hippocampal integrity in healthy elderly
2016 Foley et al. [106]	n = 197	APOE, summary statistics	Multivariate imaging—Multivariat genetic	PGS-based association	FA	Right cingulum	Fractional anisotropy of the right cingulum was inversely correlated with AD polygenic risk scores
2017 Cavedo et al. [144]	74 HC	APOE	Univariate imaging—Univariate genetic	Candidate-based association	FA, MD, radD, axD	Cingulum, corpus callosum, inferior fronto-occipital, inferior longitudinal fasciculi, internal, external capsule	These findings indicated a modulatory role of APOE ε4 on WM microstructure in elderly individuals at risk for AD suggesting early vulnerability and/or reduced resilience of WM tracts involved in AD
2018 Rutten-Jacobs et al. [145]	UK Biobank (n = 8448)	Genome-wide genotyping	Univariate imaging—Multivariate genetic	GWAS	FA, MD, WMHV	White matter hyperintensity	A novel genome-wide significant locus VCAN rs13164785 on chr5q14 was identified, which may work in the mechanisms underlying microstructural integrity of the WM measured as FA and MD
2019 Gu et al. [146]	GWAS Summary Statistics	PSEN1	Multivariate imaging—Univariate genetic	Meta- analysis	WM integrity, cerebral amyloid deposition and brain metabolism	Whole brain	PSEN1 mutation associated with WM changes and amyloid deposition occurred in AD. Increased MD was observed and showed significant increase with amyloid deposition
2020 Yan et al. [147]	ADNI (34 HC, 36 AD, 49 MCI)	34 GWAS AD risk SNPs	Univariate imaging—Multivariate genetic	Candidate-based association	Fibre anisotropy, fibre length and density	278 brain ROIs	Rs10498633 in SLC24A4 was found to be significantly associated with anisotropy, total number and length of fibres. APOE rs429358 showed nominal significance of association with the density of fibres between subcortical and cerebellum regions
2020 Horgusluoglu-Moloch et al. [148]	ADNI (34 HC, 15 AD, 56 MCI)	23 AD genes	Univariate imaging—Multivariate genetic	Candidate-based association	FA, MD, radD, axD, LIN, SPH, PLA, MOD	Hippocampus, cingulum, parahippocampal gyrus right, sagittal stratum, etc	A SNP rs2203712 in CELF1 was most significantly associated with several DTI-derived features in the hippocampus, the top ranked brain region

ALFF amplitude of low-frequency fluctuations, axD axial diffusivity, CVR rating of cerebrovascular disease, DAN dual attention network, DMN default mode network, DVR distribution volume ratios, FA fractional anisotropy, FC functional connectivity, FN frontoparietal network, HV hippocampal volume, ICV intracranial volume, gmCBF grey-matter cerebral blood flow, KBASE Korean brain aging study for early diagnosis and prediction of Alzheimer’s disease, KL-VS^het KL-VS heterozygosity, LIN linearity of the tensor, MD mean diffusivity, MOD mode of the tensor, MTA medial temporal atrophy, MTL medial temporal lobe, PLA planarity of the tensor, pMCI progressive MCI, radD radial diffusivity, SMC significant memory concern, SN salience network, SPH sphericity of the tensor, SUVR standard update value ratios, TBV total brain volume, WMH white matter hyperintensity