Surface-based but not voxel-based morphometry reveals structural abnormalities in individuals with subjective cognitive decline

Abstract

Background

Subjective cognitive decline (SCD) refers to individuals who perceive a decrease in their cognitive functioning despite no detectable deficits on neuropsychological tests, and may represent the earliest clinical stage of Alzheimer's disease (AD).

Objective

The aim of this study was to assess grey matter (GM) volumes and cortical folding patterns in individuals with SCD compared with the broader AD spectrum.

Methods

A total of 107 participants were enrolled: 31 with AD, 23 with amnestic mild cognitive impairment (aMCI), 25 with SCD and 28 healthy participants (HS). All participants underwent neuropsychological assessment and 3T MRI to acquire T1-weighted images, which were processed using CAT-12 for voxel-based morphometry (VBM) and surface-based morphometry (SBM).

Results

VBM revealed expected atrophic patterns in AD and aMCI patients compared to SCD and HS, with no volumetric differences found between SCD individuals and HS. Notably, AD patients showed cortical surface abnormalities in all SBM measures. aMCI patients showed reduced cortical thickness and gyrification compared to SCD and HS. In contrast, SCD individuals showed reduced gyrification and sulcal depth compared to HS.

Conclusions

These findings suggest that cortical surface abnormalities may manifest early in the AD continuum, whereas volumetric changes occur later. The use of multiparametric MRI techniques may be beneficial in detecting subtle changes within the AD spectrum, highlighting the potential for early identification of structural changes associated with cognitive decline in SCD individuals. Further research is warranted to explore these relationships and their implications for early intervention strategies in AD.

Keywords

Alzheimer's disease brain structural changes magnetic resonance imaging mild cognitive impairment subjective cognitive decline surface-based morphometry voxel-based morphometry

Introduction

Subjective cognitive decline (SCD) defines individuals who perceive a decrease in their own cognitive function without a detectable deficit in neuropsychological tests.^1,2 Nevertheless, they are at higher risk of developing objective cognitive decline to the point of conversion to dementia, mainly Alzheimer's disease (AD).^3,4 A meta-analysis found a conversion rate of 24.4% from SCD to mild cognitive impairment (MCI) and 10.9% from SCD to dementia over 4 years. This conversion rate is significantly higher than that observed in healthy individuals (4.6%).⁵ A recent systematic review identified fifteen predictors of conversion to AD, including biomarkers of neurodegeneration (i.e., cerebrospinal fluid amyloid-β (Aβ) levels; hippocampal atrophy), demographic and clinical factors (i.e., age, instrumental activities of daily living impairment, depression and anxiety), and neuropsychological factors in the domain of executive function measures.⁶ Consistent with this, we have recently observed that SCD individuals perform worse than normal participants on working memory tests,⁷ despite scoring above the normative cut-off on standardized neuropsychological tests. Furthermore, their functional brain connectivity within executive networks is altered and correlates with performance on working memory tests.⁷

In addition, several studies have reported the presence of an AD-type profile of biomarkers (cerebrospinal fluid (CSF) Aβ and tau levels, plasma Aβ and PET Aβ levels, reduction of the entorhinal cortex and of the hippocampal volume, white matter (WM) microstructural changes, abnormal resting state connectivity) in a large proportion of SCD individuals,^8,9 suggesting that they are in an early transitional stage between normal aging and MCI.^5,10–12 However, individuals with SCD are a very heterogeneous population. A recent study showed that education, the presence of comorbidities and cardiovascular risk factors and affective disorders may modulate the onset and progression of the cognitive decline.¹³ In this context, brain changes can be considered as early predictors of progression in SCD.^12,14,15

Recent systematic reviews examined structural MRI studies based on volumetric assessments of regions of interest, voxel-based morphometry (VBM), and cortical thickness (CT) in SCD individuals compared to healthy controls.^16–18 The Authors reported inconsistent results regarding the brain changes observed in SCD individuals.^16,18 In particular, VBM studies in SCD participants showed grey matter (GM) volume reductions in the temporal lobes, mainly in the hippocampus and amygdala,^19–24 in the frontal lobes^22,25–27 and in the precuneus.^26,27 In contrast, other VBM studies failed to detect a volumetric difference between groups.^28–34 Some ROI studies have focused on the hippocampus,^35–39 amygdala^36,40 and entorhinal cortex^41,42 reported volumetric reductions in SCD individuals, while other studies found no group difference.^43–46 Similarly, some studies have shown significant reductions in cortical thickness in individuals with SCD, mainly in the entorhinal^24,47–49 and frontal cortices,⁴⁹ whereas others have reported no significant changes.^27,45,50 Against this inconsistent background, the application of different image analysis techniques to the same datasets could help improve our ability to detect subtle brain tissue abnormalities at early clinical stages.

In this context, a combination of unbiased volumetric and surface-based techniques may be useful. Among volumetric approaches, VBM is one of the most popular techniques for detecting anatomical brain tissue changes between groups. VBM can be used to infer the presence of reductions or, less frequently, increases of regional volumes in diseased brains.⁵¹ VBM involves spatial normalization of individual brain images to a common template, tissue segmentation, and voxel-wise statistical comparisons. To preserve local volumetric information, VBM applies modulation by scaling the segmented tissue maps using the Jacobian determinants from the normalization step. This accounts for expansion or contraction during warping, allowing for the analysis of absolute or relative brain volume differences across participants based on voxel intensity values. Differently, surface-based morphometry (SBM) is a neuroimaging technique used to analyze the structure of the brain's cortical surface. Unlike voxel-based methods, SBM models the cortex and meninges as a 3D mesh, enabling the quantification of measures such as cortical thickness, gyrification, fractal dimension, and sulcal depth. Cortical thickness (CT) is derived from the distance between the inner and outer boundaries of grey matter (GM) by using voxels as a quantitative parameter.⁵² Changes in cortical thickness provide a measure of microstructural alterations in brain tissue. Gyrification expresses the relationship between the length of the folded contour and its envelope contour.⁵³ Fractal dimension (FD) is a measure that captures the morphological detail of an object (i.e., cortical surface) across multiple spatial scales, providing a numerical representation of its self-similarity and overall structural complexity. The intricate geometry of the GM makes it well-suited for analysis using fractal geometry tools, as its structural patterns exhibit fractal properties. This complexity emerges from the combined contributions of different GM components, including the pial surface, the GM–WM interface, and the cortical ribbon.⁵⁴ Differences in FD between healthy participants and individuals with SCD might reveal subtle alterations in brain morphology that may go undetected using other imaging techniques. These subtle changes might serve as an early biomarker to identify at-risk populations. It is reasonable to hypothesize that individuals with SCD may exhibit a lower cortical FD, suggesting a simpler or less intricate brain tissue structure. Nonetheless, an alternative disease-driven evolution of FD is also possible: higher FD could be associated with compensatory changes in the brain's structural organization, reflecting increased neural resource allocation. These two effects might also occur simultaneously in different brain regions. Finally, the sulcal depth can be derived from SBM and it is a quantification of the amplitude of the folding pattern.⁵⁵

To investigate cerebral abnormalities along the AD continuum we aimed to compare SCD individuals with both healthy older adults and patients at different stages of AD. In addition, a structural multimodal approach was used to better quantify brain changes. In particular, we used VBM to assess GM volumetric changes and SBM to assess the changes in the cortical complexity in terms of thickness, gyrification, sulcal depth and fractal dimension.

Methods

Thirty-one patients with AD, 24 patients with amnestic MCI (aMCI), 25 individuals with SCD, and 28 healthy participants (HS) were recruited at the Memory Clinic of IRCCS Santa Lucia Foundation (Rome, Italy). All participants underwent neuropsychological assessment by using an extensive neuropsychological battery for screening purpose (data not shown), the Addenbrooke's Cognitive Examination-Revised battery (ACE-R) and 3T-MRI. The diagnosis of AD was formulated according to the latest research criteria,⁵⁶ and with the exception of participants with SCD and healthy controls, all patients (AD, and aMCI) had a neurobiological diagnosis of AD based on CSF biomarkers (according with Core 1 Research criteria). Clinically, all aMCI patients fell into aMCI single-domain category,⁵⁷ and, by definition, did not have to meet the diagnostic criteria for major cognitive disorders.⁵⁸ Their Clinical Dementia Rating (CDR) score⁵⁹ could not exceed 0.5, while their Mini-Mental State Examination (MMSE) score^60,61 had to fall within the normality cut-off (>23.8). As detailed below, medial temporal lobe atrophy was assessed in all participants by using the medial temporal lobe scale (MTA).⁶² SCD individuals were referred by GPs to the Memory Clinic for a screening evaluation because of their own subjective feeling of experiencing cognitive disorders. Following the Jessen proposal,^1,2 SCD individuals met the following inclusion criteria: the presence of subjective memory complaints in daily life; no evidence of cognitive deficits in memory or in other cognitive domains on formal neuropsychological testing; the absence of any other clinical condition that could explain their symptoms. The inclusion criteria for the HS group were as follow: no evidence of subjective cognitive complaints in daily life; no evidence of memory or cognitive deficits on formal neuropsychological assessment; no presence of significant medial temporal lobe atrophy. All recruited subjects with a Hachinski score⁶³ greater than 4 were excluded. Major systemic, psychiatric, and other neurological diseases were also carefully screened and excluded in all participants. Finally, subjects had to be right-handed, as assessed by the Edinburgh Handedness Inventory.⁶⁴

The study was approved by the Ethics Committee of the Santa Lucia Foundation and written informed consent was obtained from all participants and or their legal guardians before study commenced. All procedures performed in this study were in accordance with the Helsinki Declaration of 1964 and its subsequent amendments or equivalent ethical standards.

Neuropsychological assessment

Addenbrooke's cognitive examination-revised battery

After recruitment, all participants, underwent the Italian version of the Addenbrooke's Cognitive Examination-Revised (ACE-R).⁶⁵ It assesses attention, memory, verbal fluency, language, and visuospatial cognitive functions and provides a score ranging from 0 to100. The Italian cut-off score for normality is > 66.92.

In addition, the ACE-R allows the derivation of the MMSE score. Individual raw MMSE scores and the sum of ACE-R scores were adjusted for age and years of formal education according to the Italian normative data.^61,66

MRI acquisition and processing

All participants underwent a 3T MRI brain scan (MAGNETOM Prisma MRI scanner, Siemens Healthcare, Erlangen, Germany) equipped with a 64-channel head-and-neck coil. All participants underwent the following whole brain acquisitions: (a) T2-weighted turbo spin-echo (TSE) [matrix 448 × 448, FOV 220 × × 220 mm^2, 29 transversal slices, slice thickness 4mm, repetition time (TR) = 3490 ms, echo time (TE) = 95 ms, acquisition time 2 min 42 s]; (b) fast fluid-attenuated inversion recovery (FLAIR) [matrix 240 × 256, FOV 240 × 256 mm^2, 176 sagittal slices, slice thickness 1 mm, TR = 8000 ms, TE = 314 ms, inversion time (TI) = 2350 ms, acquisition time 7 min 54 s]; (c) T1-weighted multi-echo MPRAGE sequence (MEMPRAGE) [matrix 256 × 256, FOV 256 × 256 mm^2, 176 sagittal slices, slice thickness 1 mm, TR = 2500 ms, TE = 1.67/3.48/5.29/7.10 ms, TI = 1080 ms, flip angle 8°, acquisition time 7 min 27 s].

The T1-weighted images were pre-processed by using voxel-based and surface-based morphometry pipelines included in CAT-12 (Computational Anatomy Toolbox 12), a toolbox of SPM 12 (https://www.fil.ion.ucl.ac.uk/spm/software/spm12/).

Briefly, images were denoised using a Spatial Adaptive Non-Local Means filter.⁶⁷ Images underwent bias-correction and affine-registration followed by standard SPM "unified segmentation",⁶⁸ which includes segmentation into probability maps of WM, GM, and CSF, normalization to the MNI template and modulation to preserve tissue volume after spatial normalization. SPM-based tissue segmentation was used here as starting point for successive refinements, which included skull striping and further bias correction based on Local Intensity Correction,⁶⁹ and including also segmentation of WM hyperintensities. Final segmentations into tissue classes were obtained using adaptive maximum a-posteriori segmentation (AMAP)⁷⁰ with partial volume estimations,⁷¹ and registered to the ICBM 2009c Nonlinear Asymmetric space using the SPM Geodesic Shooting approach.⁷² The resulting GM and WM maps were smoothed using a 8-mm FWHM Gaussian kernel to enhance the signal-to-noise ratio, reduce the impact of misregistration, and improve normality of noise distribution.

The SBM pipeline was performed after the completion of the previous VBM steps. Projection-based thickness estimation was used to calculate the cortical thickness (CT), the central surface, the gyrification, the fractal dimension and the sulcal depth,⁵⁵ including partial volume correction, sulcal blurring and asymmetry corrections. CAT-12 allows for the repair of topological defects, using a method based on spherical harmonics.⁷³ To allow analysis between subjects, CAT12 re-parameterizes the cortex into a common spherical coordinate system using an algorithm that reduces the area distortion.⁷⁴ An adapted two-dimensional diffeomorphic DARTEL algorithm was then used to register the spherically transformed cortical surface. Finally, a smoothing with a Gaussian kernel of 15 mm (FWHM) was applied to each dataset.

Medial temporal lobe atrophy

The MTA scale⁶² was used on T1-weighted images to assess the presence of hippocampal atrophy in each participant. The score ranges from 0 to 4, with scores >1.5⁷⁵ indicating significant atrophy. For each participant the scores obtained in the right and left hemispheres were averaged to obtain a single measure of atrophy.

Statistical analyses

Demographic, clinical, and neuropsychological characteristics

Group differences in demographic, clinical and cognitive variables were analyzed by using a series of one-way ANOVAs, and Tukey test was used for post hoc analyses.

Sex distribution between groups was tested using the Chi-square.

All statistical analyses were performed by SPSS-20.0 (https://www.ibm.com/it-it/analytics/spss-statistics-software).

MRI statistical analyses

The MRI data analyses were performed in the framework of the General Linear Model by using CAT-12 (Computational Anatomy Toolbox in SPM12) (https://www.fil.ion.ucl.ac.uk/spm/software/spm12/). To assess between-group differences in each MRI metric a series of full factorial models were used for voxel-wise comparisons of GM volumes in VBM analysis, for cortical thickness, gyrification, for sulcal depth and for fractal dimension. Age, sex, intracranial volumes (ICV, obtained as the sum of GM, WM, and CSF to account for individual differences in ICV, thereby compensating for the affine component of the modulation) and MMSE scores (to control for difference in cognitive decline and avoiding the risk to inflate brain differences) were used as covariates of no interest for VBM analysis. Only the demographical variables were use as covariate for SBM analyses. All results were accepted if they survived at correction for multiple comparisons (p < 0.05 FWE at cluster level and p < 0.001 uncorrected level to define cluster).

Results

Demographic characteristics

There was a significant difference between groups in mean age (F_3,104 = 4.10, p = 0.009) as shown in Table 1, as AD patients were older than the HE group as shown by Tukey's post-hoc (p = 0.013). There was also a significant difference in years of formal education (F_3,104 = 4.89, p = 0.003). Post-hoc analysis showed that AD patients were less educated than aMCI and HS groups (p = 0.034 and p < 0.003, respectively). There were differences in sex distribution between SCD individuals and both aMCI patients (Chi-square = 7.58, dg = 1, p = 0.006) and HS (Chi-square = 6.34, dg = 1, p = 0.011); no other group differences were observed (AD versus aMCI: Chi-square = 2.85, dg = 1, p = 0.091; AD versus SCD: Chi-square = 1.63, dg = 1, p = 0.202; AD versus HS: Chi-square = 1.95, dg = 1, p = 0.162; aMCI versus HS: Chi-square = 0.12, dg = 1, p = 0.730).

Table 1.

Demographic and clinical characteristics.

	AD	aMCI	SCD	HS
Sample size	N = 31	N = 24	N = 25	N = 28
Age (mean ± SD)	73.3 ± 7.4^a	72.7 ± 9.7	69.1 ± 6.7	67.1 ± 7.3
Years of formal education (mean ± SD)	10.9 ± 4.5^a,c	14.0 ± 4.4	12.9 ± 3.4	14.8 ± 3.5
Sex F/M	20/11	10/14^d	20/5^b	13/25
MTA (mean ± SD)	2.4 (0.9)^a,e	1.8 (0.9)^d,f	0.7 (0.7)	0.9 (0.9)

p < 0.05 AD versus HS.

p < 0.05 SCD versus HS.

p < 0.05 AD versus aMCI.

p < 0.05 aMCI versus SCD.

p < 0.05 AD versus SCD.

p < 0.05 aMCI versus HS.

AD: Alzheimer's disease; aMCI: amnestic mild cognitive impairment; HS: healthy subjects; SCD: subjective cognitive decline; F: female; M: male; MTA: medial temporal lobe scale.

There was also a significant difference in the MTA scores (F_3,104 = 21.8, p < 0.001). Post-hoc analyses showed that AD patients had a higher MTA score than aMCI patients and HS (p < 0.001 in both cases), and aMCI patients had a higher score than SCD individuals and HS (p < 0.001 and p = 0.005, respectively).

Neuropsychological assessment

There were significant differences in the general cognition between the groups as shown by both MMSE (F_3,104 = 31.4, p < 0.001) and ACE-R total scores (F_3,104 = 47.9, p < 0.001), with AD and aMCI patients scoring lower than other groups (p < 0.001 in each comparison). Looking at the different cognitive domains of the ACE-R we observed group differences in Attention/Orientation (F_3,104 = 21.7, p < 0.001), Memory (F_3,104 = 49.8, p < 0.001), Executive functions (F_3,104 = 22.3, p < 0.001) Language (F_3,104 = 17.3, p < 0.001) and Visuo-spatial abilities (F_3,104 = 9.9, p < 0.001). These differences were due to the fact that both AD and aMCI patients performed worse than the other groups (p < 0.001 for each comparison). No other differences were found. Please see Table 2 for further details.

Table 2.

Neuropsychological results.

	AD	aMCI	SCD	HS
Sample size	N = 31	N = 24	N = 25	N = 28
General cognitive efficacy
MMSE Mean ± SD	23.9 ± 3.3^a,c,e	26.7 ± 2.3^d,f	29.1 ± 1.4	29.5 ± 0.5
ACE-T Total score Mean ± SD	65.8 ± 11.2^a,c,e	77.8 ± 7.6^d,f	88.5 ± 5.8	88.9 ± 3.4
ACE-R domains
Orientation/Attention score Mean ± SD	14.4 ± 2.8^a,c,e	16.7 ± 1.9	17.9 ± 0.3	17.9 ± 0.2
Memory score Mean ± SD	10.8 ± 5.2^a,c,e	17.5 ± 4.7^d,f	22.6 ± 3.4	24.2 ± 2.5
Executive Functions score Mean ± SD	6.4 ± 2.7^a,c,e	9.3 ± 1.9^f	10.9 ± 2.7	11.6 ± 1.8
Language score Mean ± SD	22.4 ± 3.0^a,c,e	24.4 ± 1.6	25.6 ± 0.7	25.8 ± 0.4
Visuo-spatial abilities score Mean ± SD	11.3 ± 3.8^a,c,e	14.0 ± 3.2	15.0 ± 1.7	15.1 ± 1.1

p < 0.05 AD versus HS.

p < 0.05 SCD versus HS.

p < 0.05 AD versus aMCI.

p < 0.05 aMCI versus SCD.

p < 0.05 AD versus SCD.

p < 0.05 aMCI versus HS.

ACE-R: Addenbrooke's Cognitive Examination-Revised; AD: Alzheimer's disease; aMCI: amnestic mild cognitive impairment; HS: healthy subject; MMSE: Mini-Mental State Examination.

MRI results

Voxel based-morphometry results

As reported in Figure 1 and Table 3, patients with AD and aMCI showed the expected pattern of brain atrophy compared to the other groups, mainly affecting temporal and frontal lobe structures. Conversely, no significant reduction in brain volume was observed in SCD participants compared to HS.

Table 3.

Voxel-based morphometry results.

				MNI coordinates			Z	p-level FWE-cluster level
Group	Brain areas	Side	Size	x	y	z	Z	p-level FWE-cluster level
AD < aMCI	Parahippocampal gyrus	R	1736	27	3	−27	3.96	0.005
	Parahippocampal			20	0	−24	3.92
	gyrus/Amygdala
	Hippocampus			33	−16	−15	3.90
AD < HS	Hippocampus	R	67892	33	−16	−15	6.94	<0.001
	Amygdala			21	−8	−16	6.73
	Parahippocampal gyrus			−27	−18	−26	6.71
	Hippocampus	R	1432	39	9	33	5.85	0.011
	Hippocampus			27	0	50	4.40
	Sup. Frontal gyrus			22	12	52	4.11
	Frontal pole	R	1484	32	52	−3	5.26	0.009
	Frontal pole			34	50	4	4.66
				36	42	18	4.18
	Angular gyrus	L	1389	−54	−58	24	5.03	0.012
	Post central gyrus			−40	−34	42	4.47
	Supramarginal gyrus			−50	−51	42	4.22
AD < SCD	Hippocampus	R	19453	21	−36	2	6.74	<0.001
	Hippocampus			33	−21	−18	6.00
	Middle temporal gyrus			60	−22	−10	5.88
	Hippocampus	L	22137	−18	−38	0	6.09	<0.001
	Hippocampus			−26	−33	−8	5.68
	Inferior temporal gyrus			−51	−52	−14	5.60
	Thalamus	R	15043	8	−21	40	5.57	<0.001
	Precuneus	L		−2	−64	30	5.57
	Cingulate gyrus			−10	12	36	5.28
aMCI < HS	Fusiform gyrus	L	3724	−34	−9	−34	4.45	<0.001
	Parahippocampal gyrus			−18	−16	−24	4.26
	Parahippocampal gyrus			−27	−16	−27	4.25
	Amygdala	L	2490	18	−8	−14	4.17	0.001
	Hippocampus			34	−16	−18	3.84
	Frontal orbital cortex			21	12	−15	3.86
	Inferior temporal gyrus	R	2015	48	−34	−21	4.47	0.002
	Fusiform gyrus			38	−34	−20	4.33
	Middle Temporal gyrus			60	−12	−9	3.70
	Middle Temporal gyrus	L	911	−57	−12	−24	4.00
	Middle Temporal gyrus			−51	−2	−28	3.75
	Cerebellar lobule IX	B	931	−4	−54	−45	3.90	0.045
	Cerebellar lobule IX			−6	−58	−56	3.88
	Cerebellar lobule IX			6	−64	−56	3.75
aMCI < SCD	Angular gyrus	R	2560	54	−56	21	4.28	0.036
	Temporo-occipital gyrus			57	−52	12	4.08
	Temporo-occipital gyrus			58	−46	3	3.64
	Thalamus	R		20	−34	2	4.09
	Inf. Temporal gyrus			50	−33	−21	4.02
	Fusiform gyrus			36	−34	−21	3.82
	Crus I/ Cerebellar			30	−36	−32	3.82
	lobules VI

Surface based-morphometry results

Cortical thickness

As shown in Figure 2 and Table 4 AD patients showed significantly reduced CT compared to all other groups, with a wide anatomical distribution. Specifically, compared to HS, AD patients showed thinner thickness in the left frontal pole and orbital frontal cortex, in the left superior and middle frontal gyri, in the right precentral and inferior frontal gyri, in the right temporal pole and lingual gyrus. AD patients compared to SCD showed reduced CT in the left frontal areas (orbital cortex, frontal pole, middle frontal gyrus) in the left cingulate gyrus and amygdala, in the right frontal pole, in the superior and middle frontal gyri, the precuneus and fusiform cortex. Patients with AD compared to those with aMCI showed reduced CT in the frontal pole and temporal cortices bilaterally. Patients with aMCI compared to HS showed reduced CT in the left planum temporale and in the superior temporal gyrus, in the right supramarginal gyrus, temporal pole and inferior temporal gyrus. Finally, aMCI patients had thinner CT in the left planum temporale and insular cortex and in the right superior temporal gyrus and cerebellar lobule IX compared with SDC individuals. No other differences were found.

Table 4.

Surface-based morphometry: thickness results.

Group	Brain areas	Side	Size	MNI coordinates			Z	p-level FWE-cluster level
Group	Brain areas	Side	Size	x	y	z	Z	p-level FWE-cluster level
AD < HS	Frontal orbital cortex	L	27782	−33	28	−18	7.13	<0.001
	Frontal pole			−27	39	−17	5.99
	Planum temporale			−47	−31	4	5.69
	Cingulate gyrus	L	7007	−1	−47	8	5.38
	Superior frontal gyrus	L	3211	−19	38	49	4.74	0.001
	Superior frontal gyrus			−24	26	52	4.06
	Frontal pole	L	676	−30	54	25	4.59	0.016
	Middle frontal gyrus	L	548	−37	29	32	3.80	0.034
	Temporal pole	R	19687	26	21	−29	6.69	<0.001
	Inferior frontal gyrus			47	29	−16	6.10
	Temporal pole			55	18	−22	5.93
	Inferior frontal gyrus	R	11174	52	21	26	5.35	<0.001
	Frontal pole			27	37	49	4.95
	Middle frontal gyrus			38	33	50	4.71
	Lingual gyrus	R	4546	29	−50	2	5.18	<0.001
	Post. cingulate gyrus			11	−43	3	5.10
	Temporoccipital gyrus	R	4546	47	−39	13	5.17	<0.001
	Middle temporal gyrus	R		56	−32	−6	4.22
	Frontal pole	R	809	9	68	−5	4.16	0.009
	Fusiform gyrus	R	648	35	−5	−36	3.65	0.021
	Precentral gyrus	R	926	60	3	30	3.65	0.004
	Precentral gyrus			48	−2	30	3.56
AD < SCD	Frontal orbital cortex	L	26157	−46	33	−7	5.93	<0.001
AD < SCD	Frontal orbital cortex			−35	29	−17	5.56
	Cingulate gyrus	L	4701	−12	−47	11	4.72	<0.001
	Post. Cingulate gyrus			1	−46	12	4.57
	Temporal pole		909	−33	−2	−20	4.16	0.004
	Middle frontal gyrus	L	588	−29	33	43	3.82	0.027
	Middle frontal gyrus			−35	26	41	3.72
	Temporal pole	L	789	−58	10	−15	3.45	0.008
	Middle temporal gyrus			−61	−3	−11	3.34
	Frontal pole	R	18332	49	41	−17	6.13	<0.001
	Orbitofrontal cortex			26	23	−29	5.48
	Planum polare			48	2	−3	5.42
	Precuneus	R	6674	33	−49	9	5.00	<0.001
	Sup.Temporal gyrus			58	−32	9	4.56
	Middle temporal gyrus			46	−33	−0	4.54
	Middle frontal gyrus	R	4382	45	20	42	4.79	<0.001
	Frontal pole			32	40	39	4.64
	Lingual gyrus	R	3789	25	−48	−0	4.34	<0.001
	Middle frontal gyrus	R	875	31	1	42	3.85	0.006
	Middle frontal gyrus			32	22	56	3.44
	Amygdala	R	1153	31	0	−23	4.35	0.001
	Superior frontal gyrus	R	588	17	33	48	4.42	0.030
AD < aMCI	Frontal pole	L	3483	−46	43	−7	4.80	<0.001
AD < aMCI	Frontal pole			−36	42	−5	3.50
	Frontal pole	R	3864	49	46	−15	4.76	<0.001
	Inf. Frontal gyrus			52	19	−4	3.43
aMCI < HS	Planum Temporale	L	1396	−34	−39	15	4.14	<0.001
	Precuneus			−21	−55	18	3.54
	Sup. temporal gyrus	L	1430	−53	−16	−1	4.05	<0.001
	Sup. temporal gyrus			−50	−28	−2	3.72
	Middle temporal gyrus	L	533	−54	−18	−14	3.19	0.037
	Supramarginal gyrus	R	1354	45	−37	11	3.94	<0.001
	Temporal pole	R	949	26	20	−32	4.21	0.004
	Orbitofrontal cortex			18	31	−21	3.33
	In. Temporal gyrus	R	1021	56	−8	−33	4.02	0.003
	Middle temporal gyrus			61	−4	−21	3.28
aMCI < SCD	Planum Temporale	L	660	−38	−39	9	3.89	0.018
	Planum Temporale			−39	−41	11	3.49
	Superior temporal	R	567	43	−28	3	3.54	0.034
	gyrus
	Cerebellar lobule IX	R	650	11	−56	−35

Gyrification

As reported in Table 5 and Figure 3, compared to HS, AD patients showed significantly reduced gyrification in the left insular cortex and in the right inferior frontal gyrus; when compared to SCD individuals they showed reduced gyrification in the bilateral insular cortex. Patients with aMCI compared to HS showed reduced gyrification in the left frontal operculum and in the right middle frontal gyrus; Individuals with SCD compared to HS showed reduced gyrification in the left precentral gyrus and in the right temporal and frontal poles.

Table 5.

Surface-based morphometry: gyrification results.

Group	Brain areas	Side	Size	MNI coordinates			Z	p-level FWE-cluster level
Group	Brain areas	Side	Size	x	y	z	Z	p-level FWE-cluster level
AD < HS	Insular cortex	L	1958	−36	18	−3	4.77	<0.001
AD < HS	Inferior frontal gyrus	R	1659	44	29	−17	5.20	<0.001
AD < SCD	Insular cortex	L	590	−32	9	8	4.13	0.001
	Insular cortex	R	2686	38	10	1	3.05	<0.001
aMCI < HS	Frontal operculum	L	282	−34	24	8	3.46	0.048
	Temporal pole	R	659	49	15	−15	3.97	0.001
SCD < HS	Precentral gyrus	L	1655	−56	4	0	3.44	0.007
	Sup. temporal gyrus			−59	2	−11	2.84
	Planum temporale			−59	−15	5	2.34
	Frontal pole	L	1217	−32	43	11	2.68	0.050
	Insular cortex			−37	21	1	2.56
	Frontal pole	R	2063	42	41	1	3.49	0.001

Sulcal depth

Compared to HS, AD patients showed shorter sulcal depth in the left precentral gyrus, in the right frontal pole and superior temporal gyrus; compared with aMCI patients, they showed reduced sulcal depth in the left precentral gyrus, frontal operculum and in the right planum temporale; compared with SCD individuals, they showed reduced sulcal depth in the left temporal pole, superior temporal gyrus and in the right middle temporal gyrus. Patients with aMCI compared with SCD individuals showed shorter sulcal depth in the left middle temporal gyrus. Finally, SCD individuals compared with HS showed shorter sulcal depth in the right temporal and frontal poles (see Table 6 and Figure 4).

Table 6.

Surface-based morphometry: sulcal depth results.

Group	Brain areas	Side	Size	MNI coordinates			Z	p-level FWE-cluster level
Group	Brain areas	Side	Size	x	y	z	Z	p-level FWE-cluster level
AD < HS	Precentral gyrus	L	3360	−54	2	9	3.26	0.010
	Opercular cortex			−54	−12	8	3.21
	Frontal pole	R	32002	5	72	−2	65.6	<0.001
	Sup.Temporal gyrus	R	12986	64	12	−25	45.7	<0.001
	Planum temporale	R	5402	66	−12	9	20.3	<0.001
AD < aMCI	Frontal operculum	L	3593	−50	−7	8	3.58	0.006
	Sup. Temporal gyrus			−55	−8	−1	2.14
	Precentral gyrus	L	3400	−7	−21	52	3.41	0.010
	Temporal pole	R	3023	27	13	−30	3.25	0.011
AD < SCD	Superior temporal gyrus	L	10126	−53	−12	−2	3.59	<0.001
	Mid. Temporal gyrus	R	4670	57	−8	−20	5.00	<0.001
	Frontal pole	R	5867	40	38	18	4.59.	<0.001
aMCI < SCD	Middle temporal gyrus	L	3751	−47	−22	−11	3.77	0.005
aMCI < HS	Frontal pole	R	52282	5	73	−1	61.7	<0.001
aMCI < HS	Precentral gyrus	R	5799	66	9	18	25.1	<0.001
SCD < HS	Frontal pole	R	31853	5	73	−1	64.4	<0.001
SCD < HS	Anterior Cingulate Cortex			8	15	29	44.4
	Temporal pole	R	12019	64	12	−25	43.8	<0.001
	Inf. Temporal gyrus			50	−1	−38	31.6
	Temporal pole	R	4890	67	9	18	24.6	<0.001

Fractal dimension

When considering fractal dimension, AD patients showed reduced complexity compared with HS in the left frontal pole, inferior frontal gyrus, in the right frontal operculum frontal pole and post central gyrus; when compared with SCD individuals, they showed reduced fractal dimension in the left pre central gyrus and planum temporale, and in the right frontal operculum and temporal pole; finally, when compared with aMCI patients, they showed reduced fractal dimension in the left post central gyrus (See Table 7 and Figure 5). No other between groups differences were found.

Table 7.

Surface-based morphometry: fractal dimension results.

Group	Brain areas	Side	Size	MNI coordinates			Z	p-level FWE-cluster level
Group	Brain areas	Side	Size	x	y	z	Z	p-level FWE-cluster level
AD < HS	Frontal pole	L	1482	−36	36	9	3.43	<0.001
	Planum temporale	L	2903	−45	−36	6	3.13	0.012
	Frontal operculum	R	625	38	19	6	4.52	0.007
	Frontal pole	R	541	40	42	−6	4.60	0.015
AD < SCD	Pre central gyrus	L	3499	−43	−8	34	3.53	0.003
	Inf. Frontal gyrus			−21	10	28	2.67
	Frontal operculum	R	514	39	21	8	4.42	0.018
	Temporal pole	R	506	49	13	−14	4.17	0.019
AD < aMCI	Post central gyrus	L	2948	−60	−5	23	3.30	0.010
AD < aMCI	Frontal pole	R	396	34	46	13	4.49	0.047

Discussion

In the present study, we aimed to investigate structural brain changes in the AD continuum (including individuals with SCD) by using a multimodal approach combining both volumetric and surface-based techniques.

AD patients were older and less educated than HS, whereas all other groups showed no differences in age or education. In terms of sex distribution, we observed significant differences in the proportion of female/male in the SCD individuals compared to the aMCI and HS groups. In fact, the SCD group included more females than the other groups. This finding confirmed a previous study⁷ in which we reported that the higher proportion of females observed in the SCD group was in line with that observed in AD patients.⁷⁶ This “sex-effect” may be due to several factors including hormonal, genetic, and lifestyle variables.⁷⁷ In addition, it is possible that women generally pay more attention to early cognitive changes, as shown by others.⁷⁸

As expected, patients with AD and aMCI showed greater hippocampal atrophy than individuals with SCD and HS (see MTA scores). Conversely, no significant difference in hippocampal atrophy was observed between SCD and HS, confirming the absence of early macroscopic brain changes in SCD.

From a neuropsychological point of view, we observed an expected pattern of cognitive performance in all groups: AD patients performed worse than all other groups on all cognitive tests, while patients with aMCI reported lower scores than SCD and HS within the memory domain. aMCI patients also performed worse than HS also in executive functions. Although aMCI is typically conceptualized as predominantly involving deficits in episodic memory, several studies have previously demonstrated an impairment in executive functions⁷⁹ across multiple subdomains, including inhibitory control,⁷⁹ working memory and task switching.^79,80

When considering MRI findings, AD and aMCI patients showed the expected patterns of temporal and frontal lobe GM atrophy compared to both, SCD individuals and HS.

In addition, thinner thickness was found in the frontal and temporal cortices of AD and aMCI patients, again compared to SCD individuals and HS. It is remarkable that aMCI patients compared with SCD individuals showed reduced volumes and thinner thickness also in Crus-I, Lobule VI and Lobule IX of the cerebellum. These regions are part of the posterior cerebellum known to be linked in a reciprocal feed-forward and feed-back manner with several associative brain areas (i.e., prefrontal, parietal, temporal, parahippocampal and cingulate regions) involved in high order cognitive functions.⁸¹ Several previous structural and functional MRI studies have shown a cerebellar contribution to memory deficits observed in patients with AD at different clinical stages.^82–85 We argue here that the memory impairment shown by MCI patients compared with SCD individuals might be, at least partially, explained by cerebellar involvement. Conversely, no differences in GM volume or in cortical thickness were observed between SCD and HS.

On the one hand, this study confirms previous results on the progressive accumulation of GM atrophy in associative brain areas of AD patients at different stages.^18,86,87 Conversely, it shows that, despite a high reproducibility in discriminating between AD and MCI patients, GM volumetrics and cortical thickness fail to detect changes in the earliest clinical stages of the disease (i.e., SCD). Consistently, some previous studies, using VBM or CT analysis, failed to detect significant changes in SCD individuals.^28–34 However, we cannot ignore that other studies have reported atrophic changes between SCD individuals and healthy subjects mainly in the limbic,^19–24 frontal^22,25–27 and parietal lobes.^26,27 There are several possible explanations for these inconsistencies between studies, including the study design (clinical versus community-based) and the patient genetic status. SCD individuals recruited from memory clinics have, on average, more GM atrophy than those recruited from the community.^16,17,31 In addition, SCD individuals who are APOE4 carriers show more brain atrophy than non-carriers when compared to HS.¹⁴

The main finding of the current study was the identification of structural brain changes in individuals with SCD when using other surface-based techniques, such as gyrification and sulcal depth. Specifically, SCD individuals showed reduced gyrification in the left precentral gyrus and in the right temporal and frontal poles compared to HS. In addition, SCD individuals showed a shorter sulcal depth than HS in the right frontal and temporal poles.

Reductions in gyrification and sulcal depth were observed with a progressive pattern when moving to more advanced stages of AD. AD compared to aMCI patients, SCD individuals and HS, showed reduced gyrification in the insular cortex, and in the inferior and middle frontal gyri. aMCI patients compared with HS, showed reduced gyrification in the frontal operculum and temporal pole.

With respect to changes in sulcal depth, AD patients compared to all groups, and aMCI patients compared to SCD individuals and HS, showed shorter sulcal depth in frontal and temporal regions. Finally, regarding fractal dimension, significant differences were found only in frontal areas when comparing AD patients with all other groups (aMCI, SCD and HS). We argue that fractal dimension may only reflect gross changes in brain structure.

Remarkably, all of the structural brain changes identified in the current work fall within areas typically affected by AD pathology. Nevertheless, in this study VBM and SBM techniques did not show the same ability to detect structural changes in individuals with SCD, who represent the possible earliest stage of AD. As reported above heterogeneous results are often reported in the literature and attributed to different clinical characteristics of the recruited SCD individuals. In contrast to the heterogeneity of findings in previous studies in SCD, we observed here different levels of sensitivity when using different neuroimaging techniques in the same SCD sample. Gyrification and sulcal depth measures were able to capture significant differences in SCD individuals as compared to HS. The degree of gyrification, or cortical folding, depends in part on processes of cortical development and organization of cortico-cortical connections during embryogenesis and childhood. In this framework, increased gyrification during brain development has been hypothesized to be one of the most efficient mechanisms for promoting cortico-cortical connections.⁸⁸ The reduced gyrification and consequent reduction in sulcal depth that we observed in the fronto-temporal areas of SCD individuals may reflect early pathological changes dominated by loss of connections. Consistent with this, we have previously shown, by using a resting-state fMRI in an independent group of SCD individuals a significant reduction in functional brain connectivity in a network involving the same frontal areas.⁷ Moreover, we observed in SCD compared to healthy controls tissue loss in the left precentral gyrus, a motor area that is known to play a role also in cognitive functions such as language comprehension and working memory.⁸⁹ We hypothesize that tissue loss in this brain area may impact on working memory efficiency and its ability to hold information available to execute complex ongoing cognitive processes. This might account for the subjective cognitive impairment experienced by SCD individuals. Taken altogether, these findings support the hypothesis that a very early fronto-temporal disconnection may occur in SCD individuals, which may account for their subjective cognitive symptoms.⁷

Furthermore, gyrification is also thought to be modulated by lifelong experience.⁸⁸ Indeed, aging itself is known to have an effect on the morphometric characteristics of the brain.⁸⁸ Previous studies have shown that patients with very mild and mild AD have a lower gyrification index than normal subjects.^90,91 In addition, decreased gyrification was observed in patients with very mild and mild AD, resembling their clinical progression,⁹⁰ thus suggesting gyrification as a sensitive biomarker of disease progression.

In the current study comparing different structural techniques, we were unable to detect early changes in SCD individuals when looking at regional GM volumetrics and cortical thickness. We argue that these techniques are unable to fully capture the complexity of human brain morphology, such as the architecture and shape of surface structures (i.e., gyri and sulci).⁹² In support of this interpretation, a negative correlation between measures of cortical thickness and cortical complexity has previously been suggested.^88,93 Based on all our findings, we can propose a model for structural changes across the AD continuum, with early changes characterized by loss of circumvolutions integrity (depth of sulci and gyrification), followed by loss of cortical thickness, and regional GM atrophy.

It is possible that our observation of greater sensitivity of cortical complexity metrics compared to volumetric ones is at least partially related to the experimental procedures. Indeed, our sample was investigated using multiecho MPRAGE (MEMPRAGE), which is characterized by reduced distortion and a small additional T₂* weighting compared to MPRAGE, both factors affecting positively the reliability of cortical segmentation, with a better delineation of the cortical/pial surface, specifically associated to T2* weighting.⁹⁴

A peculiar finding of this study was the hemispheric lateralization of changes, especially in the early stages of SCD, with a prominent involvement of the right hemisphere. Lateralization of changes in SCD individuals was prominent in the temporal and frontal poles with respect to both indices (gyrification and sulcal depth), showing a better integrity on the left side. These observations are consistent with previous structural imaging studies from other groups.^{14,18,19,23,39} The consistency across studies is not only related to the left lateralization but also to specific anatomical regions, thus making these findings more robust.^19,23,39,94

An explanation for lateralization patterns of brain abnormalities in individuals with SCD may be found in developmental brain specialization and different levels of local vulnerability.⁹⁵ Handedness, sex and interactions between genetic and environmental factors such as cognitive reserve mechanisms have been suggested as major contributors to brain asymmetry.^96,97 Previous studies have shown that gains or losses in cognitive function are associated with changes in brain asymmetry.⁹⁵ These changes are likely to be altered in the presence of a neurodegenerative disease, such as AD. Furthermore, measures of cortical complexity are likely to capture early pathological changes occurring in AD brains, including synaptic loss in neuronal circuits rich in cortico-cortical connections.⁹⁸ Consistent with our findings, individuals with SCD are most likely to show early changes in those cortical regions with the highest level of cortico-cortical connectivity, thus resulting in patterns of brain asymmetry.

In addition, there is emerging evidence that WM abnormalities may contribute to the cognitive changes subjectively perceived by SCD individuals.⁹⁹ In this particular context, a multimodal MRI approach including diffusion MRI metrics has the potential to enhance the sensitivity in detecting early-stage AD pathology, as it has been previously demonstrated in patients at the MCI stage.^100–103

Limitations

A limitation of the present work is that the SCD population was not characterized using AD-related biomarkers. This increases the heterogeneity of the population by including subjects who will not convert to AD. However, the current clinical practice does not include any screening of SCD individuals based on CSF or by Amyloid-PET imaging biomarkers. The current gold standard implies clinical and neuropsychological follow-up of SCD individuals. According to Jessen et al. (2020),² SCD individuals can reverse their subjective impairment, remain stable or progress to dementia. All our SCD participants entered a longitudinal clinical and neuropsychological follow-up. In future studies, the availability of standardized blood-based biomarkers to allow a non-invasive neurobiological characterization of SCD individuals will help reduce sample variability and make results more robust. Another limitation of the present study is the relatively small-sample size however, our participants were clinically well-characterized in order to reduce potential heterogeneity to control the selection biases.

Conclusions

In conclusion, the present study, although in a relatively small-sample shows that GM volumetric changes and surface abnormalities do not progress in parallel across the AD continuum. Specifically, VBM is able to detect the presence of GM atrophy when the clinical deficits are evident at the stages of aMCI and AD. When cognitive symptoms are subjective, VBM fails to detect volumetric changes. Conversely, in SCD individuals, SBM measures are able to detect early changes in cortical architecture.

We suggest that there is a need for further investigation in larger populations, given the importance of an early diagnosis of AD for future clinical trials. In addition, a multiparametric approach, including both volumetric and surface-based techniques, should be considered in longitudinal studies aimed at capturing structural changes across the whole AD continuum, which currently includes SCD as the earliest stage of the disease that is increasingly referred to memory clinics. In the future to come, thanks to continuous improvements in data analyses based on AI algorithms, this comprehensive approach has the potential to be transferred to clinical settings for diagnostic purposes.

Footnotes

Acknowledgements

During the preparation of this work the Authors used:

- DeepL translate () in order to improve the readability of the manuscript.

After using this tool, the Authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

ORCID iDs

Laura Serra

Sabrina Bonarota

Carlotta Di Domenico

Carlo Caltagirone

Giulia Caruso

Giovanni Giulietti

Federico Giove

Ethical considerations

The study was approved by the Ethics Committee of the Santa Lucia Foundation. All procedures performed in this study were in accordance with the Helsinki Declaration of 1964 and its subsequent amendments or equivalent ethical standards.

Consent to participate

Written informed consent was obtained from all participants and or their legal guardians before study commenced.

Consent for publication

Not applicable

Author contribution(s)

Laura Serra: Conceptualization; Formal analysis; Methodology; Project administration; Supervision; Writing – original draft.

Sabrina Bonarota: Data curation; Formal analysis; Investigation; Writing – original draft.

Carlotta Di Domenico: Data curation; Investigation.

Giulia Caruso: Data curation; Investigation.

Giovanni Giulietti: Software.

Marta Rodini: Data curation.

Marco Laudadio: Data curation.

Francesco Di Lorenzo: Data curation.

Federico Giove: Resources; Supervision.

Carlo Caltagirone: Resources; Supervision.

Marco Bozzali: Conceptualization; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funded by the European Union - Next Generation EU - NRRP M6C2 - Investment 2.1 Enhancement and strengthening of biomedical research in the NHS under the grant PNRR-MAD-2022-12376889 CUP J83C22002120007.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Laura Serra and Marco Bozzali are Editorial Board Members of this journal but was not involved in the peer-review process of this article nor had access to any information regarding its peer-review.

The remaining authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The authors certify that all data are available after request received by Santa Lucia Foundation IRCCS, Rome, Italy.

References

Jessen

Amariglio

van Boxtel

, et al. A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer's disease. Alzheimers Dement 2014; 10: 844–852.

Jessen

Amariglio

Buckley

, et al. The characterisation of subjective cognitive decline. Lancet Neurol 2020; 19: 271–278.

Kaup

Nettiksimmons

LeBlanc

, et al. Memory complaints and risk of cognitive impairment after nearly 2 decades among older women. Neurology 2015; 85: 1852–1858.

Kuhn

Perrotin

Tomadesso

, et al. Subjective cognitive decline: opposite links to neurodegeneration across the Alzheimer's continuum. Brain Commun 2021; 3: fcab199.

Mitchell

Beaumont

Ferguson

, et al. Risk of dementia and mild cognitive impairment in older people with subjective memory complaints: meta-analysis. Acta Psychiatr Scand 2014; 130: 439–451.

Gao

Huang

, et al. Predictors of progression from subjective cognitive decline to objective cognitive impairment: a systematic review and meta-analysis of longitudinal studies. Int J Nurs Stud 2024; 149: 104629.

Serra

Bonarota

Di Domenico

, et al. Preclinical brain network abnormalities in patients with subjective cognitive decline. J Alzheimers Dis 2023; 95: 1119–1131.

Pavisic

Keuss

, et al. Subjective cognitive complaints at age 70: associations with amyloid and mental health. J Neurol Neurosurg Psychiatry 2021; 92: 1215–1221.

Lista

Molinuevo

Cavedo

, et al. Evolving evidence for the value of neuroimaging methods and biological markers in subjects categorized with subjective cognitive decline. J Alzheimers Dis 2015; 48: S171–S191.

10.

Fernández-Blázquez

Ávila-Villanueva

Maestú

, et al. Specific features of subjective cognitive decline predict faster conversion to mild cognitive impairment. J Alzheimers Dis 2016; 52: 271–281.

11.

Tsutsumimoto

Makizako

Doi

, et al. Subjective memory complaints are associated with incident dementia in cognitively intact older people, but not in those with cognitive impairment: a 24-month prospective cohort study. Am J Geriatr Psychiatry 2017; 25: 607–616.

12.

Xing

Han

. Subjective cognitive decline and related cognitive deficits. Front Neurol 2020; 11: 247.

13.

Ribaldi

Rolandi

Vaccaro

, et al. The clinical heterogeneity of subjective cognitive decline: a data-driven approach on a population-based sample. Age Ageing 2022; 51: afac209.

14.

Striepens

Scheef

Wind

, et al. Volume loss of the medial temporal lobe structures in subjective memory impairment. Dement Geriatr Cogn Disord 2010; 29: 75–81.

15.

Peter

Scheef

Abdulkadir

, et al. Gray matter atrophy pattern in elderly with subjective memory impairment. Alzheimers Dement 2014; 10: 99–108.

16.

Pini

Wennberg

. Structural imaging outcomes in subjective cognitive decline: community vs. clinical-based samples. Exp Gerontol 2021; 145: 111216.

17.

Wang

Huang

, et al. Neuroimaging advances regarding subjective cognitive decline in preclinical Alzheimer's disease. Mol Neurodegener 2020; 15: 55.

18.

Arrondo

Elía-Zudaire

Martí-Andrés

, et al. Grey matter changes on brain MRI in subjective cognitive decline: a systematic review. Alzheimers Res Ther 2022; 14: 98.

19.

Hafkemeijer

Altmann-Schneider

Oleksik

, et al. Increased functional connectivity and brain atrophy in elderly with subjective memory complaints. Brain Connect 2013; 3: 353–362.

20.

Liang

Zhao

Wei

, et al. Structural and functional hippocampal changes in subjective cognitive decline from the community. Front Aging Neurosci 2020; 17: 12.

21.

Perrotin

de Flores

Lamberton

, et al. Hippocampal subfield volumetry and 3D surface mapping in subjective cognitive decline. J Alzheimers Dis 2015; 48: S141–S150.

22.

Saykin

Wishart

Rabin

, et al. Older adults with cognitive complaints show brain atrophy similar to that of amnestic MCI. Neurology 2006; 67: 834–842.

23.

Scheef

Spottke

Daerr

, et al. Glucose metabolism, gray matter structure, and memory decline in subjective memory impairment. Neurology 2012; 79: 1332–1339.

24.

Cedres

Diaz-Galvan

Diaz-Flores

, et al. The interplay between gray matter and white matter neurodegeneration in subjective cognitive decline. Aging 2021; 13: 19963–19977.

25.

Chételat

Villemagne

Bourgeat

, et al. Relationship between atrophy and beta-amyloid deposition in Alzheimer disease. Ann Neurol 2010; 67: 317–324.

26.

Choi

Yoon

B-N

Choi

, et al. Reduced gray matter volume in subjective cognitive decline: a voxel-based morphometric study. Dement Neurocognitive Disord 2015; 14: 143.

27.

Hong

Yoon

Shim

, et al. Gray and white matter degenerations in subjective memory impairment: comparisons with normal controls and mild cognitive impairment. J Korean Med Sci 2015; 30: 1652.

28.

Dong

Yang

, et al. Dynamic network connectivity predicts subjective cognitive decline: the Sino-Longitudinal Cognitive impairment and dementia study. Brain Imaging Behav 2020; 14: 2692–2707.

29.

Erk

Spottke

Meisen

, et al. Evidence of neuronal compensation during episodic memory in subjective memory impairment. Arch Gen Psychiatry 2011; 68: 845–852.

30.

Parker

Smart

Scarapicchia

, et al. Identification of earlier biomarkers for Alzheimer’s disease: a multimodal neuroimaging study of individuals with subjective cognitive decline. J Alzheimers Dis 2020; 77: 1067–1076.

31.

Perrotin

La Joie

de La Sayette

, et al. Subjective cognitive decline in cognitively normal elders from the community or from a memory clinic: differential affective and imaging correlates. Alzheimers Dement 2017; 13: 550–560.

32.

Sun

Dai

, et al. Subjective cognitive decline: mapping functional and structural brain changes - a combined resting-state functional and structural MR imaging study. Radiology 2016; 281: 185–192.

33.

Xue

Sun

, et al. Disrupted patterns of rich-club and diverse-club organizations in subjective cognitive decline and amnestic mild cognitive impairment. Front Neurosci 2020; 14: 575652.

34.

Topiwala

Suri

Allan

, et al. Subjective cognitive complaints given in questionnaire: relationship with brain structure, cognitive performance and self-reported depressive symptoms in a 25-year retrospective cohort study. Am J Geriatr Psychiatry 2021; 29: 217–226.

35.

Cantero

Iglesias

Van Leemput

, et al. Regional hippocampal atrophy and higher levels of plasma amyloid-beta are associated with subjective memory complaints in nondemented elderly subjects. J Gerontol Ser A Biol Sci Med Sci 2016; 71: 1210–1215.

36.

Yue

Wang

, et al. Asymmetry of hippocampus and amygdala defect in subjective cognitive decline among the community dwelling Chinese. Front Psychiatry 2018; 9: 226.

37.

Zhao

Luo

Zhao

, et al. Automated brain MRI volumetry differentiates early stages of Alzheimer’s disease from normal aging. J Geriatr Psychiatry Neurol 2019; 32: 354–364.

38.

Zajac

Koo

B-B

Tripodis

, et al. Hippocampal resting-state functional connectivity patterns are more closely associated with severity of subjective memory decline than whole hippocampal and subfield volumes. Cereb Cortex Commun 2020; 1: tgaa019.

39.

Cong

Risacher

West

, et al. Volumetric comparison of hippocampal subfields extracted from 4-minute accelerated vs. 8-minute high-resolution T2-weighted 3T MRI scans. Brain Imaging Behav 2018; 12: 1583–1595.

40.

Schultz

Koscik

, et al. Subjective memory complaints, cortical thinning, and cognitive dysfunction in middle-age adults at risk of AD. Alzheimers Dement (Amst) 2015; 1: 33–40.

41.

Jessen

Feyen

Freymann

, et al. Volume reduction of the entorhinal cortex in subjective memory impairment. Neurobiol Aging 2006; 27: 1751–1756.

42.

Ryu

Lim

, et al. Hippocampal and entorhinal structures in subjective memory impairment: a combined MRI volumetric and DTI study. Int Psychogeriatrics 2017; 29: 785–792.

43.

Caillaud

Hudon

Boller

, et al. Evidence of a relation between hippocampal volume, white matter hyperintensities, and cognition in subjective cognitive decline and mild cognitive impairment. J Gerontol Ser B Psychol Sci Soc Sci 2020; 75: 1382–1392.

44.

Fan

Lai

Chen

, et al. Diminution of context association memory structure in subjects with subjective cognitive decline. Hum Brain Mapp 2018; 39: 2549–2562.

45.

Marcotte

Potvin

Collins

, et al. Brain atrophy and patch-based grading in individuals from the CIMA-Q study: a progressive continuum from subjective cognitive decline to AD. Sci Rep 2019; 9: 13532.

46.

Risacher

WuDunn

Tallman

, et al. Visual contrast sensitivity is associated with the presence of cerebral amyloid and tau deposition. Brain Commun 2020; 2: fcaa019.

47.

Eliassen

Reinvang

Selnes

, et al. Biomarkers in subtypes of mild cognitive impairment and subjective cognitive decline. Brain Behav 2017; 7: e00776.

48.

Meiberth

Scheef

Wolfsgruber

, et al. Cortical thinning in individuals with subjective memory impairment. J Alzheimers Dis 2015; 45: 139–146.

49.

Lauriola

Esposito

Delli Pizzi

, et al. Sleep changes without medial temporal lobe or brain cortical changes in community-dwelling individuals with subjective cognitive decline. Alzheimers Dement 2017; 13: 783–791.

50.

Niemantsverdriet

Ribbens

Bastin

, et al. A retrospective Belgian multi-center MRI biomarker study in Alzheimer’s disease (REMEMBER). J Alzheimers Dis 2018; 63: 1509–1522.

51.

Ashburner

Friston

. Voxel-based morphometry-the methods. Neuroimage 2000; 11: 805–821.

52.

Pettigrew

Soldan

Zhu

, et al. Cortical thickness in relation to clinical symptom onset in preclinical AD. Neuroimage Clin 2016; 12: 116–122.

53.

Zilles

Armstrong

Schleicher

, et al. The human pattern of gyrification in the cerebral cortex. Anat Embryol (Berl) 1988; 179: 173–179.

54.

Meregalli

Alberti

Madan

, et al. Cortical complexity estimation using fractal dimension: a systematic review of the literature on clinical and nonclinical samples. Eur J Neurosci 2022; 55: 1547–1583.

55.

Dahnke

Yotter

Gaser

. Cortical thickness and central surface estimation. Neuroimage 2013; 65: 336–348.

56.

Jack

Jr Andrews

Beach

, et al. Revised criteria for diagnosis and staging of Alzheimer's disease: alzheimer's Association Workgroup. Alzheimers Dement 2024; 20: 5143–5169.

57.

Albert

DeKosky

Dickson

, et al. The diagnosis of mild cognitive impairment due to Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement 2011; 7: 270–279.

58.

American Psychiatric Association (APA). Diagnostic and statistical manual of mental disorders. 5th edn. Milano: American Psychiatric Association, 2013.

59.

Hughes

Berg

Danziger

, et al. A new clinical scale for the staging of dementia. Br J Psychiatry 1982; 140: 566–572.

60.

Folstein

McHugh

. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–198.

61.

Measso

Cavarzeran

Zappalà

, et al. The Mini-Mental State Examination: normative study of an Italian random sample. Dev Neuropsychol 1993; 9: 77–85.

62.

Scheltens

Launer

Barkhof

, et al. Visual assessment of medial temporal lobe atrophy on magnetic resonance imaging: interobserver reliability. J Neurol 1995; 242: 557–560.

63.

Hachinski

Iliff

Zilhka

, et al. Cerebral blood flow in dementia. Arch Neurol 1975; 32: 632–637.

64.

Büsch

Hagemann

Bender

. The dimensionality of the Edinburgh Handedness Inventory: an analysis with models of the item response theory. Laterality 2010; 15: 610–628.

65.

Mioshi

Dawson

Mitchell

, et al. The Addenbrooke's Cognitive Examination Revised (ACE-R): a brief cognitive test battery for dementia screening. Int J Geriatr Psychiatry 2006; 21: 1078–1085.

66.

Pigliautile

Chiesi

Rossetti

, et al. Normative data for the ACE-R in an Italian population sample. Neurol Sci 2015; 36: 2185–2190.

67.

Manjón

Coupé

Martí-Bonmatí

, et al. Adaptive non-local means denoising of MR images with spatially varying noise levels. J Magn Reson Imaging 2010; 31: 192–203.

68.

Ashburner

Friston

. Unified segmentation. Neuroimage 2005; 26: 839–851.

69.

Gaser

Dahnke

Thompson

, et al. CAT: a computational anatomy toolbox for the analysis of structural MRI data. Gigascience 2024; 13: giae049.

70.

Rajapakse

Giedd

Rapoport

. Statistical approach to segmentation of single channel cerebral MR images. IEEE Trans Med Imaging 1997; 16: 176–186.

71.

Tohka

Zijdenbos

Evans

. Fast and robust parameter estimation for statistical partial volume models in brain MRI. Neuroimage 2004; 23: 84–97.

72.

Ashburner

Friston

. Diffeomorphic registration using geodesic shooting and Gauss-Newton optimisation. Neuroimage 2011; 55: 954–967.

73.

Yotter

Dahnke

Thompson

, et al. Topological correction of brain surface meshes using spherical harmonics. Hum Brain Mapp 2011; 32: 1109–1124.

74.

Yotter

Nenadic

Ziegler

, et al. Local cortical surface complexity maps from spherical harmonic reconstructions. Neuroimage 2011; 56: 961–973.

75.

Pereira

Cavallin

Spulber

, et al. Influence of age, disease onset and ApoE4 on visual medial temporal lobe atrophy cut-offs. Intern Med 2014; 2: 317–330.

76.

Riedel

Thompson

Brinton

. Age, APOE and sex: triad of risk of Alzheimer's disease. J Steroid Biochem Mol Biol 2016; 160: 134–147.

77.

Arenaza-Urquijo

Boyle

Casaletto

, et al. Reserve, resilience and protective factors professional interest area, sex and gender professional interest area and the ADDRESS! Special interest group. Sex and gender differences in cognitive resilience to aging and Alzheimer's disease. Alzheimers Dement 2024; 20: 5695–5719.

78.

Sánchez-Benavides

Grau-Rivera

Cacciaglia

. Distinct cognitive and brain morphological features in healthy subjects unaware of informant-reported cognitive decline. J Alzheimers Dis 2018; 65: 181–191.

79.

Johns

Phillips

Belleville

. The profile of executive functioning in amnestic mild cognitive impairment: disproportionate deficits in inhibitory control. J Int Neuropsychol Soc 2012; 18: 541–555.

80.

Zheng

Dong

Sun

, et al. The overall impairment of core executive function components in patients with amnestic mild cognitive impairment: a cross-sectional study. BMC Neurol 2012; 12: 138.

81.

Schmahmann

. The cerebellum and cognition. Neurosci Lett 2019; 688: 62–75.

82.

Olivito

Serra

Marra

, et al. Cerebellar dentate nucleus functional connectivity with cerebral cortex in Alzheimer's disease and memory: a seed-based approach. Neurobiol Aging 2020; 89: 32–40.

83.

Toniolo

Serra

Olivito

, et al. Cerebellar white matter disruption in Alzheimer's disease patients: a diffusion tensor imaging study. J Alzheimers Dis 2020; 74: 615–624.

84.

Serra

D'Amelio

Di Domenico

, et al. In vivo mapping of brainstem nuclei functional connectivity disruption in Alzheimer's disease. Neurobiol Aging 2018; 72: 72–82.

85.

Toniolo

Serra

Olivito

, et al. Patterns of cerebellar gray matter atrophy across Alzheimer's disease progression. Front Cell Neurosci 2018; 12: 430.

86.

Liu

Paajanen

Zhang

, et al. Analysis of regional MRI volumes and thicknesses as predictors of conversion from mild cognitive impairment to Alzheimer's disease. Neurobiol Aging 2010; 31: 1375–1385.

87.

Huang

Zheng

Yang

, et al. Voxel-based morphometry and a deep learning model for the diagnosis of early Alzheimer's disease based on cerebral gray matter changes. Cereb Cortex 2023; 33: 754–763.

88.

Hogstrom

Westlye

Walhovd

. The structure of the cerebral cortex across adult life: age-related patterns of surface area, thickness, and gyrification. Cereb Cortex 2013; 11: 2521–2530.

89.

Emch

von Bastian

Koch

. Neural correlates of verbal working memory: an fMRI meta-analysis. Front Hum Neurosci 2019; 13: 180.

90.

Dickerson

Feczko

Augustinack

, et al. Differential effects of aging and Alzheimer's disease on medial temporal lobe cortical thickness and surface area. Neurobiol Aging 2009; 30: 432–440.

91.

King

Brown

Hwang

, et al. Fractal dimension analysis of the cortical ribbon in mild Alzheimer's disease. Neuroimage 2010; 53: 471–479.

92.

Frisoni

Fox

Jack

Jr , et al. The clinical use of structural MRI in Alzheimer disease. Nat Rev Neurol 2010; 6: 67–77.

93.

Lee

Yoon

, et al. Fractal dimension in human cortical surface: multiple regression analysis with cortical thickness, sulcal depth, and folding area. Hum Brain Mapp 2006; 27: 994–1003.

94.

van der Kouwe

AJW

Benner

Salat

, et al. Brain morphometry with multiecho MPRAGE. Neuroimage 2008; 40: 559–569.

95.

Kim

Lee

Kim

, et al. Cortical asymmetries in normal, mild cognitive impairment, and Alzheimer's disease. Neurobiol Aging 2012; 33: 1959–1966.

96.

Hirnstein

Leask

Rose

, et al. Disentangling the relationship between hemispheric asymmetry and cognitive performance. Brain Cogn 2010; 73: 119–127.

97.

Nadeau

. Hemispheric asymmetry: what, why, and at what cost? J Int Neuropsychol Soc 2010; 16: 593–595.

98.

Masliah

Mallory

Hansen

, et al. Patterns of aberrant sprouting in Alzheimer's disease. Neuron 1991; 6: 729–739.

99.

Archer

Moore

Pamidimukkala

, et al. The relationship between white matter microstructure and self-perceived cognitive decline. Neuroimage Clin 2021; 32: 102794.

100.

Uchida

Onda

Nishimaki

, et al. Longitudinal changes in brain diffusion characteristics associated with cognition and vascular risk factors: the ARIC-NCS Study. Neurology 2025; 105: e213867.

101.

Uchida

Nishimaki

Soldan

, et al. Change points for dynamic biomarkers in the Alzheimer's disease pathological cascade: a 30-year cohort study. Alzheimers Dement 2025; 8: e70544.

102.

Uchida

Nishimaki

Soldan

, et al. Acceleration of brain atrophy and progression from normal cognition to mild cognitive impairment. JAMA Netw Open 2024; 7: e2441505.

103.

Uchida

Hou

Gomez-Isaza

, et al. Quantification of perforant path fibers for early detection of Alzheimer's disease. Alzheimers Dement 2025; 21: e70142.