Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database¹. The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer’s disease (AD). For up-to-date information, see www.adni-info.org.

Specific enrollment criteria for ADNI have been previously described in detail elsewhere (Petersen et al., 2010). Briefly, participants from ADNI were 55–90 years old, had ≥ 6 years of education or work-history equivalent, were fluent in English or Spanish, had a Geriatric Depression Scale < 6, had a Hachinski Ischemia Scale < 5, had adequate vision and hearing to perform neuropsychological tests, were in generally good health and without significant head trauma or neurologic disease, were stable on permitted medications, and had a reliable study partner. The current study included 53 participants from ADNI GO/ADNI 2 cohorts when ASL MRI was collected. Participants were included if they had ASL and neuropsychological data collected at their baseline visit and 12 months follow-up, did not have dementia at their baseline study visit, and underwent lumbar puncture for cerebrospinal fluid collection at baseline. See Supplementary Material for study sample flow chart.

Participants were characterized by ADNI as having normal cognition, significant memory complaint (SMC), early mild cognitive impairment (EMCI), or late MCI (LMCI) (Petersen et al., 2010). ADNI’s SMC diagnosis is based on having a significant memory concern but an unimpaired delayed recall score of Story A from Wechsler Memory Scale-Revised (WMS-R) Logical Memory II. ADNI’s MCI diagnoses were based on the following (1) subjective memory concern reported by the participant, study partner, or clinician; (2) abnormal memory function documented by scoring within education-adjusted ranges on delayed free recall of Story A from the WMS-R Logical Memory II subtest; (3) Mini-Mental State Examination (MMSE) score between 24 and 30; (4) global Clinical Dementia Rating (CDR) score of 0.5, with a Memory Box score of at least 0.5; and (5) general cognition and functional performance sufficiently preserved such that a diagnosis of dementia could not be made. All MCI participants who met these criteria were further classified as “early MCI” or “late MCI” by ADNI based on education-adjusted cut-offs on WMS-R Logical Memory II Story A. Finally, participants were classified as cognitively normal based on (1) no memory complaints, beyond what would be expected for age; (2) normal memory function documented by scoring above education-adjusted cutoffs on delayed free recall of Story A from WMS-R Logical Memory II (score of ≥ 9 for 16 or more years of education, score of ≥ 5 for 8–15 years of education, or score of ≥ 3 for 0–7 years of education); (3) MMSE score between 24 and 30; (4) global CDR of 0; and (5) no significant impairment in cognitive functioning or activities of daily living (Petersen et al., 2010).

This study was approved by the Institutional Review Board at the ADNI study sites. Treatment of human participants during this study was in full accordance with ethical standards set forth by the Helsinki Declaration. Informed written consent was obtained from all participants at each study site.

The IIV index is the variability across cognitive measures at a single time point. We calculated the index of dispersion, or IIV, using procedures previously described (Bangen et al., 2019). The IIV index consisted of standard summary measures from tests designed to assess multiple different cognitive abilities. Six neuropsychological measures were selected given their routine use in assessing early cognitive changes in AD, administration across all ADNI waves, and sampling of three different domains of cognition (i.e., language, processing speed/executive function, and episodic memory). These six measures were: (1) Animal Fluency, total score; (2) 30-item Boston Naming Test (BNT), total score; (3) Trail Making Test (TMT), Part A; time to completion, (4) TMT, Part B; time to completion, (5) Rey Auditory Verbal Learning Test (AVLT) 30-min delayed free recall; number of words recalled, and (6) AVLT recognition; number of words correctly recognized.

Individual raw scores for each measure were converted into age-, gender-, and education-adjusted z-scores with a mean of 0 and a standard deviation of 1 using regression coefficients derived from robust cognitively normal participants (n = 385) who had at least 1 year of follow up and remained cognitively normal throughout their participation in the ADNI study (Edmonds et al., 2015b). The TMT scores were multiplied by −1 so that higher z-scores represent better performance for all scores. The intraindividual standard deviation across these 6 z-scores was used to quantify IIV. A higher IIV index indicates greater variability across neuropsychological measures in a single session whereas a low IIV index indicates more consistency across measures. Mean level of cognitive performance was derived as the average of the 6 z-scores that were included in the IIV index. Change in IIV and mean neuropsychological performance were calculated as the difference between IIV and mean neuropsychological performance values from study baseline to 12 months follow-up.

Detailed information describing the imaging data acquisition and processing is available online at www.loni.usc.edu. Structural MRI and ASL scans were collected during the same session on a 3.0 Tesla Siemens scanner (Jack et al., 2015). FreeSurfer Version 5.1 was used to motion correct, skull strip, segment, and parcellate brain regions from structural scans² (Fischl et al., 2002, 2004).

Pulsed ASL scans were collected using QUIPS II with thin-slice TI1 periodic saturation with echo-planar imaging (Luh et al., 1999). The scan parameters are as follows: inversion time of arterial spins (TI1) = 700 ms, total transit time of spins (TI2) = 1,900 ms, tag thickness = 100 mm, tag to proximal slice gap = 25.4 mm, repetition time = 3,400 ms, echo time = 12 ms, field of view = 256 mm, matrix = 64 × 64, 24 4-mm thick axial slices [52 tag + control image pairs], time lag between slices = 22.5 ms.

ASL data processing involved automated motion correction, aligning each ASL frame to the first frame using a rigid body transformation, and least squares fitting using SPM 8³ as described previously (Sanchez et al., 2020; Thomas et al., 2020; Bangen et al., 2021). The difference of the mean-tagged and mean-untagged ASL images were calculated for the perfusion-weighted images. The images were intensity scaled to account for signal decay during acquisition and to generate intensities in meaningful physiological units. ASL images were aligned to structural T1 images using FSL after geometric distortion correction. A partial volume correction was performed that assumed that CBF in gray matter is 2.5 times greater than in white matter to mitigate effects of lower perfusion in white matter on CBF estimates. These images were normalized by the reference image (i.e., an estimate of blood water magnetization) to convert the signal into physical units (mL/100 g tissue/min). After, a global pass/fail rating was given based on visual inspection of signal uniformity, geometrical distortions, gray matter contrast, and presence of large artifacts for ADNI quality control purposes. A rating of “unusable” in any of these categories resulted in a global “fail” and that participant was excluded from this study.

To extract regional CBF estimates for each participant, FreeSurfer-derived anatomical regions of interest (ROIs) were applied to CBF maps. Our analyses examined the following six a priori ROIs: (1) hippocampus, (2) inferior parietal lobe (IPL), (3) inferior temporal gyrus (ITG), (4) medial orbitofrontal cortex (mOFC), (5) rostral middle frontal gyrus (rMFG), and (6) entorhinal cortex. These regions were selected given prior findings showing these regions are prone to early AD-related change (Dickerson et al., 2011) and consistent with our previous studies examining CBF in ADNI (e.g., Sanchez et al., 2020; Thomas et al., 2020). CBF values were residualized by precentral CBF which was used as the reference region as it is not thought to be impacted in early AD (allowing for adjustment of individual variability in CBF that is likely not due to AD pathologies) as well as to be concordant with previous ADNI ASL studies that used this approach (Mattsson et al., 2014; Yew et al., 2017). Mean CBF corrected for partial volume effects was extracted for each of the ROIs and the reference region for each hemisphere separately. Averaged bilateral CBF estimates for each ROI were used as the dependent variable in analyses to minimize the number of statistical comparisons. Bilateral CBF estimates were derived by averaging the mean CBF of each hemisphere. Change in CBF for each ROI was calculated as the difference between regional CBF values at study baseline and 12 months follow-up.

Cerebrospinal fluid biomarkers of amyloid-β (Aβ) and hyperphosphorylated tau (p-tau) were processed with Elecsys^® immunoassays. AD biomarker positivity was determined using a previously determined p-tau/Aβ ratio cut-off score (Schindler et al., 2018).

Analysis of variance (ANOVA) or Chi-squared (χ²) tests examined baseline differences in demographic and clinical characteristics between the AD biomarker-positive and biomarker-negative groups. Hierarchical linear regression, adjusting for age and gender, was used to examine the associations between the IIV index, CSF p-tau/Aβ positivity, and change in CBF from baseline to 12 months in the six a priori ROIs: hippocampus, IPL, ITG, mOFC, entorhinal, and rMFG. For all models, age and gender were entered on Step 1; p-tau/Aβ status and mean centered change in IIV were entered on step 2; and the interaction of p-tau/Aβ status and mean centered change in IIV was entered on step 3. In secondary models, change in mean level of cognitive performance was added as a covariate to see if results would remain. All analyses were performed using Statistical Package for the Social Sciences (SPSS) version 28 (SPSS IBM, NY, United States). An alpha = 0.05 was set for statistical significance.

Finally, to address potential inflation of type I error resulting from multiple comparisons, we applied the Benjamini–Hochberg procedure (Benjamini and Hochberg, 1995). We assessed results when the false discovery rate (FDR) was controlled at 0.10.

Participants’ baseline demographic and clinical data are presented in Table 1. Fifty-three older adults ranging in age from 55 to 85 (mean ± SD = 70.98 ± 6.3) comprised the present sample. There were 27 women (50.8%), the sample was 100% white, and the average number of years of formal education was approximately 17 (SD = 2.53). Compared to p-tau/Aβ– participants, the p-tau/Aβ+ group was significantly older and had poorer cognitive functioning (i.e., lower mean level of cognitive performance and poorer performance on each of the individual cognitive measures). There were no significant group differences in terms of education or gender (p-values > 0.05). The p-tau/Aβ+ group showed qualitatively higher IIV at baseline [mean (SD) IIV = 0.93 (0.45) vs. 0.77 (0.27), p = 0.12] and significantly higher IIV at 12-month follow-up [mean (SD) IIV = 0.97 (0.41) vs. 0.73 (0.23), p = 0.01] compared to p-tau/Aβ– individuals.

Hierarchical linear regression models showed that after adjusting for age and gender, as well as for the main effects of IIV change and p-tau/Aβ status, there was a significant interaction between IIV change and p-tau/Aβ status for change in entorhinal [β = −0.335, p = 0.047; Overall model: R² = 0.175, F (5,47) = 1.995, p = 0.097] and hippocampal CBF [β = −0.380, p = 0.020; Overall model: R² = 0.256, F (5,47) = 3.241, p = 0.014], but not for IPL, ITF, mOFC, or rMFG (p-values > 0.05). There were no main effects of p-tau/Aβ status or IIV change across the ROIs (see Table 2 and Figure 1). When FDR was applied, the significant finding for hippocampal CBF was maintained although the finding for entorhinal CBF were not.

When analyses were stratified by AD biomarker status, increases in IIV were associated with reductions in entorhinal [β = −0.509, p = 0.007; Overall model: R² = 0.369, F (3,17) = 3.307, p = 0.045] and hippocampal CBF [β = −0.447, p = 0.055; Overall model: R² = 0.208, F (3,17) = 1.488, p = 0.253] among individuals who were p-tau/Aβ+ (n = 21) (see Table 3). In contrast, there were no significant associations between change in IIV and CBF among those who were biomarker-negative (p-tau/Aβ–; n = 32) (see Table 4). The significant association between IIV and entorhinal CBF was maintained when the FDR was applied.

When secondary analyses also adjusted for change in mean neuropsychological performance, results remained qualitatively and statistically similar. That is, hierarchical linear regression models showed that after adjusting for age and gender, as well as for the main effects of IIV change and p-tau/Aβ status, there was a significant interaction between IIV change and p-tau/Aβ status for change in hippocampal CBF [β = −0.377, p = 0.020; Overall model: R² = 0.256, F (6,46) = 2.643, p = 0.028] and a nearly-significant interaction for change in entorhinal CBF [β = −0.326, p = 0.051; Overall model: R² = 0.188, F (6,46) = 1.773, p = 0.126]. The significant result for hippocampal CBF was not maintained when the FDR was applied.