Effects of non-modifiable risk factors of Alzheimer’s disease on intracortical myelin content

Three hundred eighty-seven cognitively normal older adults participated in the study. They were recruited from senior citizen’s associations, health-screening programs, and hospital outpatient services. All of them underwent neurological and neuropsychological assessment to discard the presence of dementia and/or objective cognitive impairment. Individuals with medical conditions that affect brain structure or function (e.g., cerebrovascular disease, epilepsy, head trauma, history of neurodevelopmental disease, alcohol abuse, hydrocephalus, and/or intracranial mass) were not included in the study. Participants met the following criteria: (i) normal global cognitive status in the Mini-Mental State Examination (scores ≥ 26); (ii) normal cognitive performance in the neuropsychological tests relative to appropriate reference values for age and education level; (iii) global score of 0 (no dementia) in the Clinical Dementia Rating; (iv) functional independence as assessed by the Spanish version of the Interview for Deterioration in Daily Living Activities [29]; (v) scores ≤ 5 (no depression) in the short form of the Geriatric Depression Scale [30]; and (vi) not be taking medications that affected cognition, sleep, renal, and/or hepatic function. All participants gave informed consent to the experimental protocol approved by the Ethical Committee for Clinical Research of the Junta de Andalucía according to the principles outlined in the Declaration of Helsinki.

From this sample, 50 individuals were APOE4 carriers (5 homozygous for the ɛ4 allele), 52 had a first-degree FH of late-onset sporadic AD, and 35 showed the co-occurrence of both AD risk factors (APOE4+FH; 1 homozygous for the ɛ4 allele). For each of these groups, we selected the same number of controls matched in age, sex, and years of education. A control subject was defined as a cognitively normal older adult, non-APOE4 carrier without FH of AD (neither first- nor second-degree relatives). The characteristics of the sample are detailed in Table 1.

All participants completed a neuropsychological assessment that included the following tests: Mini-Mental State Examination (MMSE), the Spanish version of the Memory Binding Test [31], the short form of the Boston Naming Test (BNT), Trail Making Test (forms A and B), and the Tower of London.

Images were acquired on a 3T Philips Ingenia MRI scanner using a 32-channel receive-only radio-frequency (RF) head coil and a transmit RF body coil (Philips, Best, Netherlands). The following MRI sequences were acquired in the same session: (i) 3D T1-weighted (T1w) magnetization prepared rapid gradient echo (MPRAGE) in the sagittal plane: repetition time (TR)/echo time (TE) = 2600 ms/4.7 ms, flip angle (FA) = 9°, acquisition matrix = 384 × 384, voxel resolution in acquisition = 0.65 mm³ isotropic, resulting in 282 slices without gap between adjacent slices; (ii) 3D T2w VISTA Turbo Spin Echo scan in the sagittal plane: TR/TE: 2500 ms/251 ms, FA = 90°, acquisition matrix = 384 × 384 mm, voxel resolution in acquisition = 0.65 mm³ isotropic, resulting in 282 slices without gap between adjacent slices; and (iii) T2w Fast Field Echo images using a blood-oxygen-level-dependent (BOLD) sensitive single-shot echo-planar imaging (EPI) sequence in the axial plane: TR/TE: 2000 ms/30 ms, FA = 80°, acquisition matrix = 80 × 80 mm, voxel resolution in acquisition = 3 mm³ isotropic, resulting in 35 slices acquired in posterior to anterior phase-encoding direction with 1 mm of gap between adjacent slices. To allow for optimal B1 shimming, a B1 calibration scan was applied before starting the EPI sequence. We acquired 250 EPI scans preceded by 4 dummy volumes to allow time for equilibrium in the spin excitation. Before starting the acquisition of the EPI sequence, participants were asked to remain still and keep their eyes closed without falling sleep. Pulse and respiratory rates were recorded using the scanner’s built-in pulse oximeter placed on the left-hand index finger and a pneumatic respiratory belt strapped around the upper abdomen, respectively. Brain images were visually examined after each MRI sequence; they were repeated if artifacts were identified. All participants underwent the same protocol in the same MRI scanner at the research MRI facility located at Pablo de Olavide University. The signal-to-noise ratio (SNR) was computed for each MRI sequence (i.e., T1w, T2w, and EPI) to confirm that this parameter did not differ among groups. Figure 1 in the Supplementary Material includes the SNR of each sequence for a representative subject of each group.

T1w scans were preprocessed using Freesurfer v6.0 (https://​surfer.​nmr.​mgh.​harvard.​edu/​). The Freesurfer’s pipeline included brain extraction, automated tissue segmentation, generation of WM and pial surfaces, correction of surface topology and inflation, co-registration, and projection of cortical surfaces to a sphere for the purpose of establishing a surface-based coordinate system [32]. Pial surface misplacements and erroneous WM segmentation were manually corrected on a slice-by-slice basis by one experienced technician. T2w images were registered to T1w images with bbregister using a trilinear interpolation method and a boundary-based cost function constrained to 6 degrees of freedom [33].

Individual T1w/T2w ratio volumes were sampled at the halfway between the WM and GM surfaces, resulting in midthickness surface maps of the T1w/T2w ratio. To mitigate contamination of cortical GM intensity by intensities of WM and cerebrospinal fluid (CSF), the tissue fraction effect was corrected in individual T1w/T2w ratio maps using the geometric transfer matrix-derived region-based voxel-wise method implemented in PETsurfer [34]. Finally, individual T1w/T2w ratio maps were projected onto the average cortical surface of each group, Z-transformed across vertices within each subject, and smoothed using non-linear spherical wavelet-based denoising schemes [35]. All processing steps were visually checked for quality assurance. The Z transformation applied to individual T1w/T2w ratio maps allowed us to determine how relatively different a region is compared to the mean cortical microstructure for each individual. This index was referred as the relative T1w/T2w (rT1w/T2w) ratio.

rs-fMRI data were preprocessed using AFNI functions (https://​afni.​nimh.​nih.​gov/​afni), version AFNI_20.3.01. For each participant, high-frequency spikes were eliminated (3dDespike), time-locked cardiac (measured by pulse oximeter) and respiratory motion artifacts on brain BOLD signals were minimized using RETROICOR [36], time differences in slice-acquisition were corrected (3dTshift), EPI scans were aligned using rigid body motion correction and the first volume as reference (3dVolreg), and aligned EPI scans were co-registered to their corresponding T1w volumes (align_epi_anat.py; cost function: lpc+ZZ).

Dynamics were removed provided that more than 5% of voxels exhibited signal intensities that deviated from the median absolute deviation of time series (3dToutcount) and/or when the Euclidean norm (enorm) threshold exceeded 0.3 mm in head motion. None of the participants showed more than 20% of artifactual dynamics after applying censoring. Simultaneous regression was further applied to minimize the impact of non-neuronal fluctuations on the rs-fMRI signal (3dTproject). Nuisance regressors included the following: (i) six head motion parameters (3 translational and 3 rotational) derived from the EPI scan alignment along with their first-order derivatives, (ii) time series of mean total WM/CSF signal intensity, and (iii) cardiac and respiratory fluctuations plus their derivatives to mitigate effects of extracerebral physiological artifacts on brain BOLD signals.

To estimate the sample size, we performed power analysis with the G*Power software (v3.1.9.6) (https://​www.​psychologie.​hhu.​de/​arbeitsgruppen/​allgemeine-psychologie-und-arbeitspsycholog​ie/​gpower.​html). Given the lack of evidence linking non-modifiable risk factors of AD to changes in T1w/T2w ratio maps and/or rs-FC patterns, an a priori (prospective) power analysis (fixed model, R2 deviation from zero) was performed to achieve statistical power of 80% with a significance level of 0.05 and an overall Cohen’s effect size (f²) ranging from 0.1 to 0.25. To detect an overall effect size of 0.1, we would require 113 and 133 participants for the additive (3 predictors) and interactive models (5 predictors), respectively. As different sample sizes were employed for assessing group differences in rT1w/T2w ratio, our statistical approach is not expected to identify overall effect sizes smaller than 0.11 and 0.16, respectively, in the case of additive models, and 0.13 and 0.20 in the case of interaction models.

We applied whole-cortex, vertex-wise analyses of covariance (ANCOVAs) to assess group differences in rT1w/T2w ratio levels, including age and sex as covariates of no interest. The results were corrected for multiple comparisons using a hierarchical statistical model that first controls the family-wise error rate at the cluster level by applying random field theory over smoothed statistical maps (p_vertex < 0.001, p_cluster < 0.05), and next controls the false discovery rate at the vertex level within each cluster (P < 0.05) over unsmoothed statistical maps [37]. Peaks of clusters that survived correction for multiple comparisons were employed to determine the anatomical location of significant changes using the Desikan-Killiany atlas [38].

To determine the effect size, we computed the Cohen’s f² for each additive and interactive model [39].

Table 1 summarizes demographic data and results of cognitive tests for each group. No significant group differences were found in any of these measures. The SNR of each MRI sequence was also statistically comparable between controls and AD risk groups.

Overall, cognitively normal older adults with non-modifiable AD risk factors showed significant changes in rT1w/T2w ratio levels across the cortex that differed as a function of the specific risk factor. These results are detailed in Table 2. Compared to controls, APOE4 carriers showed lower rT1w/T2w ratios in the left lingual gyrus (Fig. 1, left panel), whereas the FH group exhibited lower rT1w/T2w ratios in the right paracentral and posterior cingulate cortex along with higher rT1w/T2w ratios in different regions of the temporal lobe, cingulate cortex, and medial orbitofrontal cortex (Fig. 1, right panel). The group that merged individuals with either APOE4 or FH showed lower rT1w/T2w ratios in superior parietal regions bilaterally, left lingual gyrus, and somatosensory cortex together with higher rT1w/T2w ratios in superior temporal cortex bilaterally, left fusiform, and right supramarginal gyri (Fig. 2, left panel). However, size effects derived from this analysis were consistently lower than those obtained from analysis performed with each AD risk factor separately (see Table 2).

Significant differences in rT1w/T2w ratio levels were much more evident in the group showing the co-occurrence of the two AD risk factors. Thus, the APOE4+FH group showed lower rT1w/T2w ratios in the isthmus cingulate, pericalcarine, precentral, paracentral, and transverse temporal gyrus and higher rT1w/T2w ratios in medial and lateral aspects of the orbitofrontal, entorhinal, and insular cortex compared to controls (Fig. 2, right panel).

rT1w/T2w ratio levels also differed between APOE4+FH and APOE4 groups. Compared to APOE4, older adults with APOE4+FH showed lower rT1w/T2w ratios in precentral, paracentral, pericalcarine, and insular cortex along with higher rT1w/T2w ratios in medial orbitofrontal cortex bilaterally (Fig. 3). In contrast, no significant differences in relative T1w/T2w ratios were found between APOE4+FH and FH groups.

Significant group differences in vertex-wise correlation analysis between rs-FC patterns and cortical regions showing significant group differences in rT1w/T2w ratio maps were only evident when participants with APOE4+FH were compared with controls, as derived from group × rT1w/T2w ratio interactions in different rs-FC networks (Table 3). Post hoc analyses revealed that the lower the rT1w/T2w ratios in the left and right precentral gyrus, the lower the FC with the precuneus and paracentral gyrus of the right hemisphere, respectively, in the APOE4+FH group compared to controls (Fig. 4). Additionally, the lower the rT1w/T2w ratios in the left entorhinal and right medial orbitofrontal cortex, the higher the FC with different regions of the left frontal lobe in the APOE4+FH group compared to the control group (Fig. 4).

This study has some limitations that should be acknowledged. As the biological interpretation of cortical T1w/T2w ratio maps is still imprecise [63, 64], our findings should be interpreted cautiously. More research is clearly needed combining T1w/T2w ratio maps with other imaging modalities and post-mortem studies for a better understanding of microstructural changes in normal-appearing cortical GM. Moreover, the study sample was relatively small and these results should be considered as preliminary and replicated in further experiments. However, it should be noted that effect sizes were above the minimum established, even for the group comparison with the smallest number of participants (i.e., controls vs. APOE4+FH), which showed the largest effect sizes in rT1w/T2w ratio and rs-FC. Additionally, the EPI sequence employed in this study was not corrected for geometric distortions and, therefore, it may suffer from image distortion and signal losses in the most anterior regions of the frontal lobe [80]. Therefore, rs-FC results affecting frontal regions may be partially caused by susceptibility artifacts. Finally, the T1w/T2w ratio cancels RF receive field (B1−) artifacts but does not specifically correct for RF transmit field (B1+) errors, resulting in intensity inhomogeneities in T1w and T2w images that ultimately affect T1w/T2w ratio maps [81]. As B1+ intensity inhomogeneities were not corrected in the present study, we cannot rule out that B1+ errors are affecting our results. Future studies should acquire scans suitable for estimating the B1+ field to correct for transmit field inhomogeneities [82] in order to replicate these findings in datasets unambiguously unaffected by residual B1+ artifacts.