Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2025 Aug 22;46(12):e70328. doi: 10.1002/hbm.70328

Sex Differences in the Association of Cerebral Blood Flow and Glucose Metabolism in Normative Aging

Hamish A Deery 1,2, Chris Moran 3, Emma X Liang 1,2, Caroline Gurvich 4, Gary F Egan 2, Sharna D Jamadar 1,2,✉
PMCID: PMC12371555  PMID: 40844193

ABSTRACT

The coupling between cerebral blood flow (CBF) and glucose metabolism (CMRGLC) is critical for maintaining brain function. However, sex differences in this relationship remain poorly understood, despite the heightened risk of cognitive decline from metabolic and vascular alterations in older women. Here, we address this gap by examining CBF‐CMRGLC associations in 79 younger and older females and males using simultaneous MR/PET imaging and cognitive testing. Older adults exhibited weakened correlations between CBF and CMRGLC across functional networks. Sex moderated this decline, with older females showing significant negative CBF‐CMRGLC associations, a pattern absent in older males and younger females. Individuals with stronger CBF‐CMRGLC coupling performed better cognitively. Functional network parcellations (versus anatomical) better captured these sex‐ and age‐specific effects. Our results support the idea that brain function depends not only on absolute metabolic substrate availability but on their coordinated use across functional networks. We conclude that the reduced cognitive performance of older adults is attributable to a loss of synchronized vascular and metabolic dynamics in functional networks. Other factors moderate this association, including sex and cardiometabolic health. Across older females, there are strong, negative network CBF‐CMRGLC correlations, possibly reflecting a compensatory response in the face of attenuated rates of blood flow and glucose metabolism. The coupling of CBF and CMRGLC may serve as a biomarker for brain health and neurological conditions.

Keywords: cerebral blood flow, cognition, glucose, sex differences

Older age is associated with reduced correlation strength between cerebral blood flow and glucose metabolism within individuals. Across older females, there is a significant negative association between blood flow and glucose metabolism, a pattern not seen in older males nor younger adults. Lower correlations are associated with worse cognitive performance.

graphic file with name HBM-46-e70328-g002.jpg

1. Introduction

Metabolic and vascular dysfunction are primary drivers of age‐related cognitive decline and neurodegeneration (Zhang et al. 2021; Hy and Keller 2000; Huang et al. 2023). These dysfunctions are believed to arise from alterations in the brain's ability to match cerebral blood flow (CBF) and glucose metabolism (CMRGLC) to neuronal activity (Zhu et al. 2022). While coupling between neuronal activity, cerebral blood flow, and glucose metabolism has been known for decades (Roy and Sherrington 1890; Sokoloff 1977), recent evidence suggests that those relationships are more complex than previously assumed, varying by brain region, cognitive state, and individual (Drew 2022; Stiernman et al. 2021; Logothetis and Wandell 2004; Liu et al. 2023; Frahm et al. 1996). For example, during cognitive tasks, the default mode network exhibits decreased CBF but stable CMRGLC (Stiernman et al. 2021; Aiello et al. 2015), while attentional networks show tighter positive coupling (Villien et al. 2014; Hahn et al. 2018; Jamadar et al. 2019). Such heterogeneity raises the question of the extent to which these patterns reflect constraints of the underlying metabolic systems, individual‐specific physiology, systemic population‐level alterations (e.g., aging), or their interaction, and whether their dissociation signals early risk for cognitive decline.

Resting‐state studies further highlight the complexity of the association between cerebral blood flow and glucose metabolism. While moderate correlations exist between regional CBF and resting CMRGLC within individuals (Sokoloff 1978; Wang et al. 2021; Cha et al. 2013; Fox et al. 1988), their strength varies by the region's role in functional networks, with stronger associations in “hub” than non‐hub regions (Volpi et al. 2024; Manza et al. 2020). Critically, the relationship between regional CBF and CMRGLC across individuals—a potential marker of group‐wide cerebral metabolic‐vascular health (Zhu et al. 2022; Manza et al. 2020; Xu et al. 2023)—remains poorly understood. Some studies report positive correlations between blood flow and metabolism (Zhu et al. 2022; Cha et al. 2013), while others report no relationship between the two (Powers et al. 2011; Henriksen et al. 2018), possibly due to unaccounted moderators like age and metabolic health (Deery et al. 2024).

While systemic factors like age and insulin resistance may influence CBF–CMRGLC associations, the relationship between group‐level (across‐people) and individual‐level (within‐individuals) coupling remains largely unexplored. Discordance between these levels could signal physiological compensatory mechanisms or neural resilience, such as when a person maintains tight coupling despite belonging to a high‐risk group. Conversely, individuals who conform to group‐level association patterns but show weak within‐individual coupling may already exhibit localized dysfunction not evident through systemic or global markers. By simultaneously analyzing group‐ and individual‐level associations, we can assess whether they provide complementary predictive utility for brain health beyond either measure alone and potentially identify early signatures of vulnerability or resilience.

One important factor that influences blood flow–metabolism coupling is sex. Females exhibit higher baseline CBF than males across the lifespan (Martin et al. 2024) but also greater dementia risk, a disparity often attributed to post‐menopausal metabolic shifts (Baker et al. 2016; Robison et al. 2019). Although it is known that estrogen regulates glucose metabolism and vascular function (Robison et al. 2019), it is unknown whether the association between CBF and CMRGLC differs between the sexes and whether this coupling changes with age. Answering these questions could reveal why females face a higher risk of age‐related cognitive decline and may indicate whether hormonal interventions might preserve cerebral metabolic coupling and cognitive health in older women.

Here, we address these gaps by leveraging simultaneous MR/PET data to compare CBF‐CMRGLC associations within and across people, testing whether they have distinct predictive utility for cognitive performance. We evaluate sex and age differences in CBF‐CMRGLC associations, hypothesizing stronger associations in younger vs. older adults. We also expected stronger associations between CBF and CMRGLC in males vs. females, with females showing larger age‐related declines due to vascular and metabolic changes that occur in aging and the post‐menopausal hormonal changes in older females (Ali et al. 2023). We also assess whether these associations predict cognition. By focusing on functional networks (e.g., default mode, attentional networks) (Biswal et al. 1995; Calhoun et al. 2008), we link coupling variability to circuits critical for memory and executive function—cognitive domains vulnerable to aging (Harada et al. 2013; Glisky 2007). We hypothesize that higher CBF‐CMRGLC associations will be linked to better cognitive performance. Our approach extends prior work by contrasting systemic metabolic‐vascular factors across younger and older people from within‐individual associations, while testing sex‐specific vulnerabilities.

2. Materials and Methods

2.1. Ethical Considerations

The study protocol was reviewed and approved by the Monash University Human Research Ethics Committee. Administration of ionizing radiation was approved by the Monash Health Principal Medical Physicist, following the Australian Radiation Protection and Nuclear Safety Agency Code of Practice (2005). For adult participants, a 5 mSv annual radiation exposure limit applied. The effective dose in our study was 4.9 mSv. Participants provided informed consent to participate.

2.2. Participants

Local advertising was used to recruit 90 participants from the general community. An initial screening interview ensured that participants had the capacity to provide informed consent. Exclusion criteria were a diagnosis of diabetes, neurological or psychiatric illness, or dementia. Participants were also screened for claustrophobia, non‐MR compatible implants, and a PET scan in the past 12 months, and for women, current or suspected pregnancy. Participants received a $100 voucher for completing the study.

Eleven participants were excluded from further analyses due to excessive head motion (n = 2), incomplete PET scan or blood hemolysis (n = 3), incomplete ASL scans (N = 2) or consistently high or low regional ASL or CMRGLC values more than 2.5 standard deviations from the mean (N = 4). The final sample included 79 individuals, 17 younger males (mean 27.5; range 20–36 years), 20 younger females (28.4; 20–42), 22 older males (76.6; 66–84), and 20 older females (75.3; 66–89) (see Table 1).

TABLE 1.

Mean and standard deviation (SD) of demographics and cognitive test performance for the whole sample and the four groups based on sex and age category.

Whole sample (N = 79) Younger males (N = 17) Younger females (N = 20) Older males (N = 22) Older females (N = 20) Sex difference Age difference Sex × age effect
Mean SD Mean SD Mean SD Mean SD Mean SD p p‐FDR p p‐FDR p p‐FDR
Age (years) 53.5 24.9 27.5 5.4 28.4 7.0 76.6 5.4 75.3 6.8 NA NA NA NA NA NA
Years of education 16.9 3.2 17.9 2.7 17.9 2.7 16.9 2.8 15.1 3.8 0.197 0.364 0.009 0.018 0.538 0.948
Blood glucose (mmol/L) 5.0 0.5 4.9 0.3 4.7 0.4 5.3 0.5 5.0 0.6 0.032 0.095 0.002 0.005 0.584 0.948
Insulin (mU/L) 4.4 2.9 6.3 4.2 4.1 1.5 4.7 3.0 3.2 2.3 0.005 0.023 0.065 0.091 0.667 0.948
HOMA‐IR 1.0 0.7 1.4 1.0 0.9 0.3 1.2 0.8 0.8 0.6 0.005 0.023 0.119 0.139 0.711 0.948
HOMA‐IR2 0.6 0.4 0.8 0.5 0.5 0.2 0.6 0.4 0.4 0.3 0.005 0.023 0.345 0.345 0.924 0.948
Systolic BP (mmHg) 135.5 26.3 123.5 17.1 116.6 18.0 152.3 19.7 146.3 29.7 0.195 0.364 0.000 0.000 0.766 0.948
Diastolic BP (mmHg) 81.5 12.9 78.2 12.9 79.7 14.2 83.9 12.2 83.6 12.1 0.836 0.836 0.102 0.130 0.836 0.948
Heart rate (BPM) 78.1 15.8 80.4 14.7 84.1 19.8 71.9 12.7 77.0 13.4 0.208 0.364 0.028 0.045 0.348 0.948
Body mass index (BMI) 25.0 4.1 24.3 3.0 24.4 5.8 26.3 2.5 24.7 4.3 0.415 0.646 0.237 0.255 0.482 0.948
Cortical thickness (mm) 2.4 0.1 2.5 0.1 2.5 0.1 2.3 0.1 2.3 0.1 0.691 0.744 0.000 0.000 0.219 0.948
HVLT: delayed recall 8.3 2.6 9.2 3.0 9.3 2.0 6.5 2.3 8.7 2.0 0.034 0.095 0.002 0.005 0.057 0.399
Cat switch: mean RT (sec) 1.8 0.6 1.3 0.2 1.5 0.5 2.3 0.7 1.9 0.5 0.642 0.744 0.000 0.000 0.014 0.196
Digit Sub: correct count 44.4 22.4 65.3 9.0 62.7 14.2 28.5 11.4 26.7 16.9 0.473 0.662 0.000 0.000 0.888 0.948
Stop signal: RT (msec) 0.56 0.12 0.53 0.16 0.52 0.99 0.60 0.96 0.58 0.13 0.539 0.686 0.029 0.045 0.948 0.948
Open in a new tab

Note: p values are for the sex, age group, and sex and age group interaction effects from general linear models. Three participants didn't answer the years of education question; baseline blood samples not available for three participants for insulin and HOMA‐IR calculation.

2.3. Demographic and Cognitive Variables

Prior to the scan, participants completed a demographic questionnaire online. The measures included age, sex assigned at birth, gender identity, education, height, and weight. Female participants reported if they were pre‐, peri‐, or post‐menopausal. Participants also completed a battery of cognitive tests, comprised of domains validated in aging research (Raz and Rodrigue 2006) (see supplement for details): delayed recall from the Hopkins Verbal Learning Test (HVLT); reaction time in a task switching test to index cognitive control; reaction time in a stop signal task to measure response inhibition; and the number of correct responses in a digit substitution task to measure visuospatial performance.

2.4. MR/PET Data Acquisition

Participants underwent a 90‐min simultaneous MR/PET scan in a Siemens Biograph 3‐Tesla molecular MR scanner. They were requested to consume a high‐protein and low‐sugar diet for the 24 h prior to the scan, fast for 6 h, and to drink 2–6 glasses of water. Participants were cannulated in the vein in each forearm, and a 10 mL baseline blood sample was taken. At the beginning of the scan, half of the 260 MBq FDG tracer was administered via the left forearm as a bolus. The remaining 130 MBq of the FDG was infused at a rate of 36 mL/h over 50 min. This combined bolus and constant infusion protocol maximizes the signal‐to‐noise ratio over the period of the scan (Jamadar et al. 2020).

Participants were positioned supine in the scanner bore with their head in a 32‐channel radiofrequency coil. The scan sequence started with non‐functional MRI scans during the first 12 min, including a T1 3DMPRAGE (TA = 3.49 min, TR = 1640 ms, TE = 234 ms, flip angle = 8°, field of view = 256 × 256 mm2, voxel size = 1.0 × 1.0 × 1.0 mm3, 176 slices, sagittal acquisition) and T2 FLAIR (TA = 5.52 min, TR = 5000 ms, TE = 396 ms, field of view = 250 × 250 mm2, voxel size = 0.5 × 0.5 × 1 mm3, 160 slices). 13 min into the scan, list‐mode PET (voxel size = 2.3 × 2.3 × 5.0 mm3) sequences were started. A 40‐min resting‐state scan was acquired while participants watched a movie of a drone flying over the Hawaii Islands. At 53 min, a 5‐delay pseudo‐continuous arterial spin labelling (pcASL) scan was initiated (TR = 4220 ms; TE = 45.46 ms; FOV = 240 mm; slice thickness = 3 mm; voxel size 2.5 × 2.5 × 3.0 mm3). Post labelling delays were 0.5, 1.0, 1.5, 2.0, and 2.5 s, and the duration of the labelling pulse was 1.51 s.

Plasma radioactivity levels were measured during the scan from 5 mL blood samples taken from the right forearm every 10 min for a total of nine samples. The samples were spun in a centrifuge at 2000 rpm for 5 min. 1000‐μL of plasma was placed in a well counter for 4 min, and the count start time, total number of counts, and counts per minute were recorded.

2.5. Data Preparation and Preprocessing

2.5.1. Cortical Thickness

Cortical thickness was extracted using FreeSurfer (Fischl 2012) for the 100 regions of the Schaefer functional parcellation (Schaefer et al. 2018). To extract cortical thickness values for each region, we first used FreeSurfer's recon‐all pipeline to generate cortical surface reconstructions and vertex‐wise cortical thickness maps in each subject's native surface space. The Schaefer atlas (fsaverage version) was then mapped to each subject's native surface using the resampling tools in FreeSurfer. Finally, cortical thickness was averaged across all surface vertices belonging to each of the 100 regions, ensuring topologically accurate regional summaries while preserving anatomical accuracy.

2.5.2. Cerebral Blood Flow

We used the BASIL toolbox from the FMRIB's software library to quantify cerebral blood flow (Groves et al. 2009). A calibration map M0 of proton density weighted image was acquired for each participant. Single−subject, whole−brain CBF maps were calculated from perfusion weighted images with direct subtraction of label and control volumes. Image processing included motion, distortion, and partial volume correction, a macro vascular component, adaptive spatial regularization of perfusion, and T1 uncertainty. The arterial transit time was fixed at 1.3 s, T1/T1b at 1.3/1.66 s, and inversion efficiency at 0.85. The resulting native space CBF images were aligned to the anatomical T1 images and normalized to MNI152 space. Grey matter CBF for the Schaefer 100 regions was obtained for each participant.

2.5.3. PET Image Reconstruction and Pre‐Processing

The list‐mode PET data were binned into 344 3D sinogram frames of 16 s intervals. Attenuation was corrected via the pseudo‐CT method for hybrid MR/PET scanners (Burgos et al. 2014). Ordinary Poisson‐Ordered Subset Expectation Maximization algorithm (3 iterations, 21 subsets) with point spread function correction was used to reconstruct 3D volumes from the sinogram frames. The reconstructed DICOM slices were converted to NIFTI format with size 344 × 344 × 127 (size: 1.39 × 1.39 × 2.03 mm3) for each volume. All 3D volumes were temporally concatenated to form a single 4D NIFTI volume.

The PET volumes were motion corrected (Jenkinson et al. 2002) using the mean PET image to mask the 4D data. The data was then corrected for partial volume effects using PETsurfer tools within FreeSurfer (Fischl 2012), applying a 25% grey matter threshold (Greve et al. 2016). The PET images were also spatially smoothed with surface‐based smoothing (Greve et al. 2014) using a Gaussian kernel with a full width at half maximum of 8 mm. We used FreeSurfer to convert the smoothed PET data to native volume space, which ensures topologically constrained smoothing along the cortical sheet by preserving anatomical boundaries and minimizing cross‐sulcal blurring. Finally, ANTs was used to coregister the PET data to MNI space.

2.5.4. Cerebral Metabolic Rates of Glucose

Calculations of CMRGLC were undertaken using the FDG time activity curves for the Schaefer atlas regions. The FDG in the plasma samples was decay‐corrected for the time between sampling and counting as the input function to Patlak models (Karjalainen et al. 2020).

One younger female and one older female had relatively high CMRGLC values (see dots to the right of the Figure 1 panels). Physiological CMRGLC values in the healthy adult brain are around 5 mg/100 g/min, with regional typical ranges of 4 to 10 mg/100 g/min (Sokoloff 1977, 1978; Clarke and Sokoloff 1999; Deery et al. 2023). Although the one younger and older female with relatively high values fell within this range, we repeated analyses with them excluded to test for their effect on the group results (see Section 3).

FIGURE 1.

FIGURE 1

Open in a new tab

Mean rates and whole brain associations of cerebral blood flow and glucose metabolism across people. Regional mean cerebral blood flow (A) and glucose metabolism (B) for younger and older males and females. Significant sex and age group differences were found for CBF and CMRGLC (see Tables S1 and S2). Correlation of whole brain CBF and CMRGLC for the whole sample and sex and age groups with no covariates (C.i and D.i) and controlling for all demographics (C.ii and D.ii). In (C.ii and D.ii), the demographics were regressed out of CBF and CMRGLC across the whole sample, before the correlations of CBF and CMRGLC were calculated for each group.

2.5.5. Network‐Level Measures

In addition to regional measures, CBF and CMRGLC were calculated for eight functional networks. This eight‐network classification was derived from the 17‐network Schaefer parcellation (100‐region version), in which the 17 networks are nested within the broader seven‐network scheme. Notably, the Temporo‐Parietal (TP) network is explicitly labeled separately in this version, enabling an eight‐network structure. We included the TP network given the established age‐related alterations in glucose metabolism in temporal regions (Deery et al. 2023). For our analysis, regions within the eight networks and across both hemispheres were averaged to generate a single CBF and CMRGLC value per network. Because the cortical volume of regions within each network varies, regional values were weighted according to the proportion of each region's volume relative to the total volume of the corresponding network.

2.6. Statistical Analysis

All analyses were run in SPSS version 29.0.

2.6.1. Demographics and Cognition

Univariate general linear models (GLMs) were used to test for sex, age group, and sex × age group interaction effects for each demographic and cognitive measure, FDR‐corrected for multiple comparisons.

2.6.2. Rates of CBF and CMRGLC

Separate series of general linear models (GLMs) were run to test for sex, age, and sex and age group interaction effects in grey matter CBF and CMRGLC in the 100 regions. Each main effect and interaction in each series of GLMs was FDR‐corrected at p < 0.05.

2.6.3. CBF‐CMRGLC Correlations Across People

Whole brain and network level CBF and CMRGLC associations across participants were calculated in the whole sample and in the sex and age groups separately. Two series of correlations were undertaken. In the first, no covariates were included. For the second, the demographic variables were regressed on network CBF and CMRGLC separately across the whole sample. The residuals were then correlated across participants in the whole sample and for the sex and age groups separately. The correlations were tested for significance from zero using one‐sample t‐tests and between groups using z‐tests, FDR‐corrected for multiple comparisons for the network‐level analyses.

2.6.4. Within Individual CBF‐CMRGLC Correlations

For each participant, the correlation was calculated between their 100 regional CBF and CMRGLC values and between their eight network CBF and CMRGLC values. Whole sample and sex and age group–averaged correlations were tested for significance from zero using one‐sample t‐tests. GLMs testing for age group, sex, and age group × sex interaction effects were also undertaken. Each series of analysis was FDR‐corrected for multiple comparisons.

2.6.5. Within‐Individual and Across‐People CBF and CMRGLC Associations and Cognition

Multiple regression was used to test the association between network CBF‐CMRGLC correlations and cognitive performance. Principal Component Analysis (PCA) was used to reduce dimensionality in the cognitive data, with the resulting PC used as the dependent variable in the regression. The PCA was undertaken on normalized cognition scores, with the reaction time in the category switch and stop signal tasks multiplied by −1 so that higher scores reflect better performance in all cognitive tasks.

We included both within‐ and across‐people measures of CBF–CMRGLC association in the regression analysis to address the questions of their relative predictive utility for cognitive performance. Specifically, the regression model included a single within‐participant and single across‐people measure, regressed on the PCA score. For the within‐individual measure, we computed the correlation between CBF and CMRGLC across the eight functional networks. This within‐individual measure captures how tightly associated blood flow and glucose metabolism are across the functional brain networks for a given person, providing an individual‐level measure of coupling. For the across‐participant measure, we first averaged CBF and CMRGLC across the eight networks to obtain a single value per modality. We then regressed the average CMRGLC on the average CBF across all participants and extracted the predicted (fitted) values from this model for each person. These predicted values served as a single, combined index of how closely each person's CBF‐CMRGLC relationship aligned with the overall group‐level association.

3. Results

3.1. Sample Characteristics

All participants reported alignment between sex assigned at birth and gender identity. Sex and age group differences were found for all demographic variables, except for diastolic blood pressure and BMI. However, none of the sex and age group interaction effects were significant when corrected for multiple comparisons (p‐FDR < 0.05). All females in the older group reported being post‐menopausal. Males had higher insulin resistance (HOMA‐IR) than females. Older adults had fewer years of education than younger adults. Older adults also had lower heart rate and cortical thickness and higher fasting blood glucose and systolic blood pressure. Given the impact of these demographic variables on brain metabolism (Deery et al. 2024), we repeated the following analyses reported below controlling for cortical thickness, blood pressure, resting heart rate, insulin resistance, and BMI.

Older adults had lower HVLT delayed recall, fewer correct trials in the digit substitution task, and slower reaction time in the stop signal and category switch tasks compared to younger adults.

3.2. Rates of CBF and CMRGLC

We assessed whether known effects of age and sex on CBF and CMRGLC were obtained in our sample. Consistent with established findings from the literature, females had higher CBF than males at the whole brain and in 100 regions (p‐FDR < 0.05; Figure 1A,B; also see Tables S1–S3). Sex differences in CBF were retained in 98 regions after adjusting for the demographic variables, indicating that sex differences in cerebral blood flow are not attributable to differences in education level or the measured physiological factors (cortical thickness, HOMA‐IR, blood pressure, heart rate and BMI) between the sexes. As expected, younger adults had higher whole brain CBF than older adults and higher CBF in 97 regions (p‐FDR < 0.05).

None of the regional sex differences in CMRGLC survived correction for multiple comparisons (Table S2). Younger adults also had higher glucose metabolism in 84 regions (p‐FDR < 0.05). However, the age differences in CMRGLC were no longer significant after adjusting for the other physiological variables, with the effect of cortical thickness particularly strong.

3.3. CBF and CMRGLC Associations Across People

In our first set of CBF‐CMRGLC correlations, we analyzed age group and sex differences to understand whether there are different patterns in CBF‐CMRGLC associations across the groups.

3.3.1. Whole Brain

Scatterplots of the whole brain CBF and CMRGLC associations across participants are shown in Figure 1. There was a positive correlation between CBF and CMRGLC that approached significance in the whole sample (r = 0.20, p = 0.077; Figure 1C.i). When cortical thickness, blood pressure, resting heart rate, insulin resistance, and BMI were controlled, there was a significant negative correlation between CBF and CMRGLC (r = −0.26, p = 0.028; Figure 1C.ii).

In the sex and age group analyses, whole brain CBF‐CMRGLC correlations were not significant (Figure 1D.i). When controlling for demographic variables, the negative correlation between CBF and CMRGLC across older females was significant (r = −0.80, p < 0.001; Figure 1D.ii). The correlations among older females remained significant (r = −0.63, p = 0.004) when the participants with CMRGLC above 6.0 mg/100 g/min were excluded (see dots to the right of the Figure 1 panels). Given that the significant negative association of CMRGLC and CBF in older females was retained, and the fact that the subjects' CMRCLC values were within physiological range, we retained the participants in our analyses.

3.3.2. Network Level

With the exception of the visual network, there was a positive correlation across people in the whole sample between network CBF and CMRGLC (Figure 2A.i and Table S3). Note, however, that the correlations did not survive correction for multiple comparisons (p‐FDR > 0.05). The largest correlations were in the control (r = 0.19), limbic (0.21) and salience ventral attention (0.24) networks. When cortical thickness, blood pressure, resting heart rate, insulin resistance, and BMI were controlled, there was a non‐significant negative correlation across all participants between network CBF and CMRGLC (Figure 2A.ii). The largest correlations were in the default (r = −0.13), somatomotor (−0.13), visual (−0.14), and temporal parietal (−0.21) networks.

FIGURE 2.

FIGURE 2

Open in a new tab

Across‐people network cerebral blood flow and glucose metabolism associations. Association of network CBF and CMRGLC in the whole sample (A.i) and the sex and age subgroups (B.i to B.iv) with no covariates (and controlling for all demographics (A.ii and C.i to C.iv)). For (C), black asterisk (*) to the left of the bars indicates a significant correlation different from zero at p‐FDR < 0.05. Red asterisk (*) indicates a significant difference in correlation between older males and older females at p‐FDR < 0.05; and green asterisk (*) a significant difference between younger females and older females at p‐FDR < 0.05. The test of significant group differences in correlations are also summarized in Table S3. CON = control; DA = dorsal attention; DEF = default; LIM = limbic; SOM = somatomotor; SVA = salience ventral attention; TP = temporal parietal; VIS = visual.

Across both younger and older males, small‐to‐moderate but non‐significant negative correlations were found (Figure 2B.i,ii). Across younger females, small‐to‐moderate positive correlations were found between CBF and CMRGLC in all networks (Figure 2B.iii). The pattern seen in younger females was inverted for older females, who demonstrated moderate‐to‐high negative network CBF‐CMRGLC correlations (Figure 2B.iv). When other demographic variables were controlled, the negative correlations across older females were significant (p‐FDR < 0.05) for all networks except the limbic and control networks (Figure 2C.iv; Table S3). Significant differences were also found between the correlation of older males versus older females in all networks except the limbic and control networks; and between younger females and older females in all networks (p‐FDR < 0.05).

3.4. Within Individual CBF and CMRGLC Associations

In our second set of analyses, we examined within‐subject CBF‐CMRGLC associations to understand the coupling of blood flow and glucose metabolism within individuals and whether the strength and direction of the coupling on average differ between younger and older people and females and males.

The within‐subject CBF and CMRGLC correlations are shown in Figure 3 (also see Table S4). Across the 100 regions, small‐to‐moderate positive correlations were found. Those correlations were significantly different from zero for the whole sample and all sex and age groups (p‐FDR < 0.05). Across the eight networks, the correlations were also positive and significantly different from zero for younger females and males (p‐FDR < 0.05) but not older females and males. The age difference was also significant in the GLM across the eight networks (F = 9.1, p‐FDR < 0.05).

FIGURE 3.

FIGURE 3

Open in a new tab

Whole sample and sex and age group mean within‐individual correlations. Black asterisk indicates a correlation significantly different to zero, *p < 0.05 and **p < 0.01. The t‐test for correlations being different to zero and GLMs of group differences are provided in Table S4. The eight networks are the visual, somatomotor, dorsal attention, salience ventral attention, limbic, control, default, and temporal parietal networks.

3.5. CBF‐CMRGLC Associations and Cognition

Thirdly, we examined the relationship between CBF‐CMRGLC and cognition. One significant principal component (PC) was identified from the four cognitive measures with an eigenvalue greater than one, explaining 56.5% of the variance. The regression analysis of the association of the cognition PC with network CBF‐CMRGLC correlations was significant (F = 6.1, p = 0.004), explaining 14% of the variance. Higher within‐individual (beta = 0.26, p = 0.018) and across‐people (beta = 0.24, p = 0.028) network CBF‐CMRGLC correlations were associated with better cognitive performance (see Figure 4).

FIGURE 4.

FIGURE 4

Open in a new tab

Associations of cerebral blood flow and glucose metabolism correlations with cognition. Within‐individual CBF‐CMRGLC associations and cognition (i) and across‐people CBF‐CMRGLC associations and cognition (ii) from the regression analysis predicting cognition. Principle Component Analysis (PCA) of the cognitive data identified a single PC used that was used as the dependent variable in the regression analyses. Within‐individual CBF‐CMRGLC correlations and across‐people were used as the dependent variables. For the across‐people measure, the average network CMRGLC was regressed on the average CBF value and the predicted score for each participant saved for use in the regression analysis.

3.6. Anatomical Parcellation

To validate our results, we re‐ran all analyses using an anatomical brain parcellation. As it is currently unknown whether CBF and CMRGLC association differences relate to functional or anatomical subdivisions of the brain, we repeated the analyses using the Harvard Oxford anatomical atlas of 106 cortical and subcortical regions. Three salient differences were evident in the results of the anatomical parcellation compared to the functional parcellation (see Supporting Information). First, the within‐individual CBF‐CMRGLC correlations were 2–3 times stronger for the anatomical than the functional parcellation (Figure S2). Second, the CBF‐CMRGLC correlations across people showed similar patterns in the whole sample and sex and age groups using both parcellations, including moderate to high negative correlations among older females (Figure S1). Differences between older females and the other groups were also significant at p < 0.05 using the anatomical parcellation in the frontal, limbic, parietal, temporal, and occipital areas, although those differences did not survive FDR‐correction (p‐FDR = 0.085). Third, as noted above, both higher CBF‐CMRGLC correlations across people and within‐individuals significantly predict cognition in the functional parcellation. However, only the correlations across people (beta = 0.28, p = 0.010) significantly predicted cognition using the anatomical parcellation. The within‐individual correlations in the anatomical parcellation did not significantly predict cognition (beta = 0.21, p = 0.056).

4. Discussion

Using simultaneous MR/PET, we identified three novel patterns in neurovascular‐metabolic associations. We found a striking negative association between CBF and CMRGLC across older females that was absent in older males and younger females. The strength of the CBF and CMRGLC association within and across people in functional networks predicted cognitive performance and declined with age. We also found that functional (vs. anatomical) network parcellations were more sensitive to age and sex effects and cognitive performance, despite weaker absolute coupling in the functional parcellation. Our results suggest that coordinated CBF‐CMRGLC dynamics in functional networks are critical for maintaining cognition in aging and have sex‐specific vulnerabilities. Our results also support the idea that brain function depends not just on absolute metabolic substrate availability, but on its coordinated use across functional networks (Biswal et al. 1995; Calhoun et al. 2008).

4.1. Sex‐Specific Effects in Aging

Although it has been known for decades that increases in neural activity drive changes in local blood flow (Drew 2022), surprisingly little research has investigated sex differences in neurovascular coupling or the association of blood flow and glucose metabolism in the resting state (Burma et al. 2025). Our results contribute to the literature by revealing that sex moderates whole brain and network CBF‐CMRGLC associations across people in aging. The significant negative association between CBF and CMRGLC across older females was consistent with our hypothesis. We expected a weakening of the association due to vascular and metabolic changes that occur in aging and the post‐menopausal hormonal changes in older females (Ali et al. 2023).

We also found lower underlying rates of cerebral blood flow in older adults than younger adults, replicating well‐documented age‐related declines reported in the literature (Deery et al. 2023; Rodriguez et al. 1988; Parkes et al. 2004; Devous et al. 1986). However, our results extend previous findings by demonstrating that CBF and CMRGLC interrelationship changes with age and sex, and that their relationship impacts cognition. Together, these lower rates and negative associations between CBF and CMRGLC in older females may reflect the brain's drive to maintain metabolic homeostasis within a relatively tight physiological range (Deery et al. 2024). In other words, for older females to maintain the brain's supply of blood carrying oxygen and glucose, rates of blood flow and glucose metabolism may increase in compensation for a decrease in the other and vice versa. We also found that lower CBF‐CMRGLC correlations, as seen in older females, were associated with worse cognitive performance. These results suggest that the negative association of CBF and CMRGLC across older females might reflect a compensatory response in an effort to support cognition in the face of age‐related losses of blood flow, glucose metabolism, or both.

It has been suggested that the associations of CBF and CMRGLC may serve as a biomarker for brain health and neurological conditions (Zhu et al. 2022; Manza et al. 2020; Xu et al. 2023). However, to our knowledge, the relative utility of within‐individual and across‐people CBF‐CMRGLC correlations for aging and cognition has not previously been tested. Hence, our study is an important empirical step forward: it provides initial evidence that the strength of functional network CBF‐CMRGLC association—both within individuals and across people—is related to age, sex, and cognitive performance, albeit in different ways. Small, positive correlations within individuals were found between CBF and CMRGLC among younger and older males and females. However, the across‐people correlations were strongly negative in females. As noted above, the discordance between these patterns in older females may signal compensatory mechanisms or metabolic or neural resilience. If replicated and validated in larger clinical samples, this biomarker could distinguish between normal and pathological aging and be used for early detection before overt symptoms arise in conditions like dementia.

The reason that older males do not show a similar pattern of CBF and CMRGLC associations to older females is unclear. It is also possible that the small, positive correlation across older males is a relative “failure” of the negative coupling seen across healthy younger males and older females. Given that older males also have lower absolute rates of CBF and CMRGLC than older females, it is possible that a failure of the coupling occurs when CBF, CMRGLC, or both, drop below a critical threshold. Evidence for this idea comes from ischemia models in animals, in which decreases in CMRGLC occur with the decrease to 38% of baseline levels, below which an increase in CMRGLC is observed (Sako et al. 1985). The possibility of a threshold for the reversal of CBF‐CMRGLC associations in humans, including different threshold levels in women and men, could be tested in longitudinal studies that track the time course and trajectories of changes in the rates and coupling of CBF and CMRGLC.

The significant negative association between CBF and CMRGLC across older females but not older males may also reflect underlying biological differences, which are thought to be key contributors to different aging trajectories and disease risks in women and men, including the risk for neurodegeneration. For example, it is possible that older males compensate for a reduction in blood flow and glucose availability through different mechanisms. Reductions in CBF in older men (Moffat and Resnick 2007) have been attributed to the modulating role of testosterone in vascular function via angiogenesis and BBB permeability (Ahmadpour and Grange‐Messent 2021). Changes in estrogen and testosterone also contribute to the different courses of aging in females and males (see (Austad 2019) for review). Brain‐derived hormones, as well as proteins and peptides that act in the brain, such as growth hormone, insulin receptor substrate 2, and insulin‐like growth factor (IGF‐1), also contribute to differences in brain aging observed between the sexes (Austad 2019).

In this study, all females in the younger group were pre‐menopausal (we did not control for cycle phase or use of hormonal contraceptives). All females in the older group were post‐menopausal, and only three (15%) were on hormone replacement therapy (HRT). Therefore, it is tempting to conclude that the differences between young and older females are hormonal in origin. While we do not have data linking sex hormones to CBF in our sample, research suggests that sex hormones play a mediating role in the control of CBF. For example, CBF and carotid artery flow increase following acute estrogen administration and long‐term hormone replacement therapy in post‐menopausal women (Ohkura et al. 1996; Kaya et al. 2008; Ciccone et al. 2013). Future research could explore these linkages by measuring sex hormones together with cerebral blood flow and metabolism.

4.2. Moderating Factors and Clinical Implications

To further understand the effects of age and sex on metabolism, we controlled for other physiological variables, including cortical thickness, blood pressure, resting heart rate, insulin resistance, and BMI. With those variables as covariates, the number of brain regions showing age group differences in absolute rates of CBF was greatly reduced. Interestingly, these physiological variables did not strongly influence the relationship between sex and underlying rates of CBF. However, they did change the direction of correlations between CBF and CMRGLC across the whole sample from positive to negative and increased the strength of the negative correlation in older females. These findings emphasize that age‐related changes in cerebral metabolism are influenced by a constellation of physiological health markers, such as blood pressure, BMI, and insulin resistance. While some of the apparent age‐related decline in CBF may reflect poorer health, sex differences in its association with CMRGLC appear more resilient to such factors, suggesting distinct biological underpinnings. Furthermore, the modulation of CBF‐CMRGLC coupling by health markers points to the value of an integrative model of brain aging that recognizes the brain's dependence on whole‐body vascular and metabolic health (Liu et al. 2024; Bantubungi et al. 2022).

4.3. Functional Networks as Metabolic Organising Units

When we repeated the analyses using an anatomical (vs. functional) atlas, we observed both convergent and divergent findings. Notably, within‐individual network CBF‐CMRGLC correlations were 2–3 times stronger in the anatomical parcellation, suggesting that blood flow and glucose metabolism are more tightly coupled among spatially proximal regions than among distributed functional networks. However, this tighter anatomical coupling did not translate to the same level of cognitive relevance: while functional network‐based correlations were weaker, they were more predictive of age, sex, and cognitive performance. These dissociations highlight a critical distinction: whereas anatomical proximity may facilitate local neurovascular‐metabolic coupling, functional network alignment—despite weaker correlations—appears more sensitive to systemic factors (e.g., aging, sex and cardiometabolic differences) and cognitive outcomes.

Our findings of differences in functional and anatomical parcellations extend prior work in two key ways. First, they suggest that functional networks not only organize temporal dynamics to support cognitive function but also constrain the brain's metabolic architecture, with metabolic‐functional alignment reflecting cognitive health (Jamadar et al. 2025). Second, anatomical coupling, while stronger, may reflect “baseline” hemodynamic‐metabolic coordination, whereas functional coupling captures behaviorally and clinically relevant variability. We encourage future studies to adopt a dual‐parcellation approach, as our results indicate that anatomical and functional metrics index complementary aspects of neurovascular‐metabolic health. Functional parcellations may better identify early risk for cognitive decline, while anatomical measures could better index dysfunction in neurovascular coupling, potentially in advance of cognitive decline or progression of neurodegeneration.

We found small‐to‐moderate positive correlations between network CBF and CMRGLC across people and within individuals in our entire sample. The correlation coefficients were similar in strength to those reported in previous studies (Park et al. 2023; Baron et al. 1984; Gottstein et al. 1964). While the dynamics of resting‐state cerebral blood flow and glucose metabolism differ from active states, the significant correlations within and across people suggest that the principles of neurovascular coupling likely still apply at rest, including at the spatial level of large‐scale functional brain networks (Sokoloff 1977; Yellen 2018). They also support the idea that efficient neurovascular and neurometabolic function combine to support healthy brain function, providing a physiological basis for the large energy budget needed to support spontaneous brain activity at rest (Raichle 2015).

4.4. Limitations

The study has a number of limitations, in particular the cross‐sectional design, which prevents conclusions about the causal relationship between age and cerebral blood flow and metabolism changes. As noted above, the differences found between groups may reflect unmeasured physiological, hormonal, or health differences. While a key strength of this study was the simultaneous acquisition of perfusion and metabolic data, additional studies should be undertaken with larger samples. Lastly, in the absence of a middle‐aged cohort, we cannot infer when in the lifespan age‐related changes in CBF and CMRGLC and their association may occur. In particular, CBF alterations across the lifespan may follow a non‐linear pattern (Biagi et al. 2007). The absence of a middle‐aged adult cohort in the current study precluded non‐linear relationships from being tested but should be a focus of future studies.

5. Conclusion

In conclusion, our results extend the previously documented attenuation of cerebral blood flow and glucose metabolism in aging by demonstrating that their interrelationship changes with age and sex and impacts cognition. Our results support the idea that brain function depends on the coordinated deployment of metabolic substrates in functional networks. Older adults lose synchronized vascular and metabolic dynamics in large‐scale functional networks, which are necessary for cognitive processes. Other factors moderate this association, including sex and cardiometabolic health. Across older females, there is a strong, negative association of blood flow and glucose metabolism, possibly reflecting a compensatory response to optimize cognition and available metabolic substrates in the face of reduced cerebral blood flow, glucose metabolism, or both. In older males, there is an absence of association. A better understanding of the interplay of sex hormones, cerebrovascular, and cerebral metabolic differences in females and males can inform the development of interventions to optimize brain function and cognition across the adult lifespan.

Ethics Statement

The study protocol was approved by the Monash University Human Research Ethics Committee.

Consent

Participants provided informed consent to participate in the study.

Conflicts of Interest

Associate Editor is co‐author: Sharna D. Jamadar is handling editor of Human Brain Mapping and a co‐author of this article. To minimize bias, she was excluded from all editorial decision‐making related to the acceptance of this article for publication. Deputy Editor is co‐author: Gary F. Egan is a deputy editor of Human Brain Mapping and a co‐author of this article. To minimize bias, he was excluded from all editorial decision‐making related to the acceptance of this article for publication.

Supporting information

Data S1: Supporting Information.

HBM-46-e70328-s001.pdf (1.4MB, pdf)

Data S2: Supporting Information.

HBM-46-e70328-s002.xlsx (100.7KB, xlsx)

Acknowledgments

Open access publishing facilitated by Monash University, as part of the Wiley ‐ Monash University agreement via the Council of Australian University Librarians.

Deery, H. A. , Moran C., Liang E. X., Gurvich C., Egan G. F., and Jamadar S. D.. 2025. “Sex Differences in the Association of Cerebral Blood Flow and Glucose Metabolism in Normative Aging.” Human Brain Mapping 46, no. 12: e70328. 10.1002/hbm.70328.

Funding: Sharna D. Jamadar is supported by an Australian National Health and Medical Research Council (NHMRC) Fellowship (APP1174164).

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  1. Ahmadpour, D. , and Grange‐Messent V.. 2021. “Involvement of Testosterone Signaling in the Integrity of the Neurovascular Unit in the Male: Review of Evidence, Contradictions, and Hypothesis.” Neuroendocrinology 111, no. 5: 403–420. [DOI] [PubMed] [Google Scholar]
  2. Aiello, M. , Salvatore E., Cachia A., et al. 2015. “Relationship Between Simultaneously Acquired Resting‐State Regional Cerebral Glucose Metabolism and Functional MRI: A PET/MR Hybrid Scanner Study.” NeuroImage 113: 111–121. [DOI] [PubMed] [Google Scholar]
  3. Ali, N. , Sohail R., Jaffer S. R., et al. 2023. “The Role of Estrogen Therapy as a Protective Factor for Alzheimer's Disease and Dementia in Postmenopausal Women: A Comprehensive Review of the Literature.” Cureus 15, no. 8: e43053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Austad, S. N. 2019. “Sex Differences in Health and Aging: A Dialog Between the Brain and Gonad?” Geroscience 41, no. 3: 267–273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Baker, S. E. , Limberg J. K., Ranadive S. M., and Joyner M. J.. 2016. “Neurovascular Control of Blood Pressure Is Influenced by Aging, Sex, and Sex Hormones.” American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 311, no. 6: R1271–R1275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bantubungi, K. , Vieau D., Blum D., and Ferreira S. T.. 2022. “Uncovering Bidirectional Brain‐Body Interactions in Health and Disease.” Neuropharmacology 212: 109073. [DOI] [PubMed] [Google Scholar]
  7. Baron, J. C. , Rougemont D., Soussaline F., et al. 1984. “Local Interrelationships of Cerebral Oxygen Consumption and Glucose Utilization in Normal Subjects and in Ischemic Stroke Patients: A Positron Tomography Study.” Journal of Cerebral Blood Flow and Metabolism 4, no. 2: 140–149. [DOI] [PubMed] [Google Scholar]
  8. Biagi, L. , Abbruzzese A., Bianchi M. C., Alsop D. C., del Guerra A., and Tosetti M.. 2007. “Age Dependence of Cerebral Perfusion Assessed by Magnetic Resonance Continuous Arterial Spin Labeling.” Journal of Magnetic Resonance Imaging 25, no. 4: 696–702. [DOI] [PubMed] [Google Scholar]
  9. Biswal, B. , Zerrin Yetkin F., Haughton V. M., and Hyde J. S.. 1995. “Functional Connectivity in the Motor Cortex of Resting Human Brain Using Echo‐Planar MRI.” Magnetic Resonance in Medicine 34, no. 4: 537–541. [DOI] [PubMed] [Google Scholar]
  10. Burgos, N. , Cardoso M. J., Thielemans K., et al. 2014. “Attenuation Correction Synthesis for Hybrid PET‐MR Scanners: Application to Brain Studies.” IEEE Transactions on Medical Imaging 33, no. 12: 2332–2341. [DOI] [PubMed] [Google Scholar]
  11. Burma, J. S. , Bailey D. M., Johnson N. E., et al. 2025. “Physiological Influences on Neurovascular Coupling: A Systematic Review of Multimodal Imaging Approaches and Recommendations for Future Study Designs.” Experimental Physiology 110, no. 1: 23–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Calhoun, V. D. , Kiehl K. A., and Pearlson G. D.. 2008. “Modulation of Temporally Coherent Brain Networks Estimated Using ICA at Rest and During Cognitive Tasks.” Human Brain Mapping 29, no. 7: 828–838. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cha, Y. H. , Jog M. A., Kim Y.‐C., Chakrapani S., Kraman S. M., and Wang D. J. J.. 2013. “Regional Correlation Between Resting State FDG PET and pCASL Perfusion MRI.” Journal of Cerebral Blood Flow and Metabolism 33, no. 12: 1909–1914. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Ciccone, M. M. , Scicchitano P., Gesualdo M., et al. 2013. “Systemic Vascular Hemodynamic Changes due to 17‐Beta‐Estradiol Intranasal Administration.” Journal of Cardiovascular Pharmacology and Therapeutics 18, no. 4: 354–358. [DOI] [PubMed] [Google Scholar]
  15. Clarke, D. D. , and Sokoloff L.. 1999. “Substrates of Cerebral Metabolism.” In Basic Neurochemistry: Molecular, Cellular and Medical Aspects, edited by Siegel A. B., Albers G. J., and Siegel R. W.. Lippincott‐Raven. [Google Scholar]
  16. Deery, H. A. , di Paolo R., Moran C., Egan G. F., and Jamadar S. D.. 2023. “Lower Brain Glucose Metabolism in Normal Ageing Is Predominantly Frontal and Temporal: A Systematic Review and Pooled Effect Size and Activation Likelihood Estimates Meta‐Analyses.” Human Brain Mapping 44, no. 3: 1251–1277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Deery, H. A. , Liang E., di Paolo R., et al. 2024. “The Association of Regional Cerebral Blood Flow and Glucose Metabolism in Normative Ageing and Insulin Resistance.” Scientific Reports 14, no. 1: 14574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Devous, M. D., Sr. , Stokely E. M., Chehabi H. H., and Bonte F. J.. 1986. “Normal Distribution of Regional Cerebral Blood Flow Measured by Dynamic Single‐Photon Emission Tomography.” Journal of Cerebral Blood Flow and Metabolism 6, no. 1: 95–104. [DOI] [PubMed] [Google Scholar]
  19. Drew, P. J. 2022. “Neurovascular Coupling: Motive Unknown.” Trends in Neurosciences 45, no. 11: 809–819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Fischl, B. 2012. “FreeSurfer.” NeuroImage 62, no. 2: 774–781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fox, P. T. , Raichle M. E., Mintun M. A., and Dence C.. 1988. “Nonoxidative Glucose Consumption During Focal Physiologic Neural Activity.” Science 241, no. 4864: 462–464. [DOI] [PubMed] [Google Scholar]
  22. Frahm, J. , Krüger G., Merboldt K. D., and Kleinschmidt A.. 1996. “Dynamic Uncoupling and Recoupling of Perfusion and Oxidative Metabolism During Focal Brain Activation in Man.” Magnetic Resonance in Medicine 35, no. 2: 143–148. [DOI] [PubMed] [Google Scholar]
  23. Glisky, E. L. 2007. “Changes in Cognitive Function in Human Aging.” In Brain Aging: Models, Methods, and Mechanisms, edited by Riddle D. R.. CRC Press. [PubMed] [Google Scholar]
  24. Gottstein, U. , Bernsmeier A., and Sedlmeyer I.. 1964. “The Carbohydrate Metabolism of the Human Brain. Ii. Studies With Substrate Specific Enzymatic Methods in Patients With Diminished Cerebral Circulation due to Cerebrovascular Arteriosclerosis.” Klinische Wochenschrift 42: 310–313. [DOI] [PubMed] [Google Scholar]
  25. Greve, D. N. , Salat D. H., Bowen S. L., et al. 2016. “Different Partial Volume Correction Methods Lead to Different Conclusions: An (18)F‐FDG‐PET Study of Aging.” NeuroImage 132: 334–343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Greve, D. N. , Svarer C., Fisher P. M., et al. 2014. “Cortical Surface‐Based Analysis Reduces Bias and Variance in Kinetic Modeling of Brain PET Data.” NeuroImage 92: 225–236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Groves, A. R. , Chappell M. A., and Woolrich M. W.. 2009. “Combined Spatial and Non‐Spatial Prior for Inference on MRI Time‐Series.” NeuroImage 45, no. 3: 795–809. [DOI] [PubMed] [Google Scholar]
  28. Hahn, A. , Gryglewski G., Nics L., et al. 2018. “Task‐Relevant Brain Networks Identified With Simultaneous PET/MR Imaging of Metabolism and Connectivity.” Brain Structure & Function 223, no. 3: 1369–1378. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Harada, C. N. , Natelson Love M. C., and Triebel K. L.. 2013. “Normal Cognitive Aging.” Clinics in Geriatric Medicine 29, no. 4: 737–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Henriksen, O. M. , Vestergaard M. B., Lindberg U., et al. 2018. “Interindividual and Regional Relationship Between Cerebral Blood Flow and Glucose Metabolism in the Resting Brain.” Journal of Applied Physiology (Bethesda, MD: 1985) 125, no. 4: 1080–1089. [DOI] [PubMed] [Google Scholar]
  31. Huang, L. Y. , Ou Y. N., Yang Y. X., Wang Z. T., Tan L., and Yu J. T.. 2023. “Associations of Cardiovascular Risk Factors and Lifestyle Behaviors With Neurodegenerative Disease: A Mendelian Randomization Study.” Translational Psychiatry 13, no. 1: 267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Hy, L. X. , and Keller D. M.. 2000. “Prevalence of AD Among Whites: A Summary by Levels of Severity.” Neurology 55, no. 2: 198–204. [DOI] [PubMed] [Google Scholar]
  33. Jamadar, S. D. , Behler A., Deery H., and Breakspear M.. 2025. “The Metabolic Costs of Cognition.” Trends in Cognitive Sciences 29, no. 6: 541–555. [DOI] [PubMed] [Google Scholar]
  34. Jamadar, S. D. , Ward P. G. D., Close T. G., et al. 2020. “Simultaneous BOLD‐fMRI and Constant Infusion FDG‐PET Data of the Resting Human Brain.” Scientific Data 7, no. 1: 363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Jamadar, S. D. , Ward P. G. D., Li S., et al. 2019. “Simultaneous Task‐Based BOLD‐fMRI and [18‐F] FDG Functional PET for Measurement of Neuronal Metabolism in the Human Visual Cortex.” NeuroImage 189: 258–266. [DOI] [PubMed] [Google Scholar]
  36. Jenkinson, M. , Bannister P., Brady M., and Smith S.. 2002. “Improved Optimization for the Robust and Accurate Linear Registration and Motion Correction of Brain Images.” NeuroImage 17, no. 2: 825–841. [DOI] [PubMed] [Google Scholar]
  37. Karjalainen, T. , Tuisku J., Santavirta S., et al. 2020. “Magia: Robust Automated Image Processing and Kinetic Modeling Toolbox for PET Neuroinformatics.” Frontiers in Neuroinformatics 14: 3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Kaya, E. , Şahin F. K., Köken G., Köse M., and Cevrioglu A. S.. 2008. “Acute Effect of Intranasal Estrogen on Cerebral and Cerebellar Perfusion in Postmenopausal Women.” Maturitas 59, no. 1: 72–82. [DOI] [PubMed] [Google Scholar]
  39. Liu, S. , Yang X., Chen F., and Cai Z.. 2023. “Dysfunction of the Neurovascular Unit in Brain Aging.” Journal of Biomedical Research 37, no. 3: 153–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Liu, Y. , Tan Y., Zhang Z., Yi M., Zhu L., and Peng W.. 2024. “The Interaction Between Ageing and Alzheimer's Disease: Insights From the Hallmarks of Ageing.” Translational Neurodegeneration 13, no. 1: 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Logothetis, N. K. , and Wandell B. A.. 2004. “Interpreting the BOLD Signal.” Annual Review of Physiology 66: 735–769. [DOI] [PubMed] [Google Scholar]
  42. Manza, P. , Wiers C. E., Shokri‐Kojori E., et al. 2020. “Brain Network Segregation and Glucose Energy Utilization: Relevance for Age‐Related Differences in Cognitive Function.” Cerebral Cortex 30, no. 11: 5930–5942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Martin, J. , Reid N., Ward D. D., King S., Hubbard R. E., and Gordon E. H.. 2024. “Investigating Sex Differences in Risk and Protective Factors in the Progression of Mild Cognitive Impairment to Dementia: A Systematic Review.” Journal of Alzheimer's Disease 97, no. 1: 101–119. [DOI] [PubMed] [Google Scholar]
  44. Moffat, S. D. , and Resnick S. M.. 2007. “Long‐Term Measures of Free Testosterone Predict Regional Cerebral Blood Flow Patterns in Elderly Men.” Neurobiology of Aging 28, no. 6: 914–920. [DOI] [PubMed] [Google Scholar]
  45. Ohkura, T. , Teshima Y., Isse K., et al. 1996. “Estrogen Increases Cerebral and Cerebellar Blood Flows in Postmenopausal Women.” Menopause 2: 13–18. [Google Scholar]
  46. Park, J. C. , Lim H., Byun M. S., et al. 2023. “Sex Differences in the Progression of Glucose Metabolism Dysfunction in Alzheimer's Disease.” Experimental & Molecular Medicine 55, no. 5: 1023–1032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Parkes, L. M. , Rashid W., Chard D. T., and Tofts P. S.. 2004. “Normal Cerebral Perfusion Measurements Using Arterial Spin Labeling: Reproducibility, Stability, and Age and Gender Effects.” Magnetic Resonance in Medicine 51, no. 4: 736–743. [DOI] [PubMed] [Google Scholar]
  48. Powers, W. J. , Videen T. O., Markham J., Walter V., and Perlmutter J. S.. 2011. “Metabolic Control of Resting Hemispheric Cerebral Blood Flow Is Oxidative, Not Glycolytic.” Journal of Cerebral Blood Flow and Metabolism 31, no. 5: 1223–1228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Raichle, M. E. 2015. “The Brain's Default Mode Network.” Annual Review of Neuroscience 38: 433–447. [DOI] [PubMed] [Google Scholar]
  50. Raz, N. , and Rodrigue K. M.. 2006. “Differential Aging of the Brain: Patterns, Cognitive Correlates and Modifiers.” Neuroscience and Biobehavioral Reviews 30, no. 6: 730–748. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Robison, L. S. , Gannon O. J., Salinero A. E., and Zuloaga K. L.. 2019. “Contributions of Sex to Cerebrovascular Function and Pathology.” Brain Research 1710: 43–60. [DOI] [PubMed] [Google Scholar]
  52. Rodriguez, G. , Warkentin S., Risberg J., and Rosadini G.. 1988. “Sex Differences in Regional Cerebral Blood Flow.” Journal of Cerebral Blood Flow and Metabolism 8, no. 6: 783–789. [DOI] [PubMed] [Google Scholar]
  53. Roy, C. S. , and Sherrington C. S.. 1890. “On the Regulation of the Blood‐Supply of the Brain.” Journal of Physiology 11, no. 1–2: 85–158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sako, K. , Kobatake K., Yamamoto Y. L., and Diksic M.. 1985. “Correlation of Local Cerebral Blood Flow, Glucose Utilization, and Tissue pH Following a Middle Cerebral Artery Occlusion in the Rat.” Stroke 16, no. 5: 828–834. [DOI] [PubMed] [Google Scholar]
  55. Schaefer, A. , Kong R., Gordon E. M., et al. 2018. “Local‐Global Parcellation of the Human Cerebral Cortex From Intrinsic Functional Connectivity MRI.” Cerebral Cortex 28, no. 9: 3095–3114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Sokoloff, L. 1977. “Relation Between Physiological Function and Energy Metabolism in the Central Nervous System.” Journal of Neurochemistry 29, no. 1: 13–26. [DOI] [PubMed] [Google Scholar]
  57. Sokoloff, L. 1978. “Local Cerebral Energy Metabolism: Its Relationships to Local Functional Activity and Blood Flow.” Ciba Foundation Symposium 56: 171–197. [DOI] [PubMed] [Google Scholar]
  58. Stiernman, L. J. , Grill F., Hahn A., et al. 2021. “Dissociations Between Glucose Metabolism and Blood Oxygenation in the Human Default Mode Network Revealed by Simultaneous PET‐fMRI.” Proceedings of the National Academy of Sciences of the United States of America 118, no. 27: e2021913118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Villien, M. , Wey H. Y., Mandeville J. B., et al. 2014. “Dynamic Functional Imaging of Brain Glucose Utilization Using fPET‐FDG.” NeuroImage 100: 192–199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Volpi, T. , Silvestri E., Aiello M., et al. 2024. “The Brain's “Dark Energy” Puzzle: How Strongly Is Glucose Metabolism Linked to Resting‐State Brain Activity?” Journal of Cerebral Blood Flow and Metabolism 44, no. 8: 1433–1449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Wang, J. , Sun H., Cui B., et al. 2021. “The Relationship Among Glucose Metabolism, Cerebral Blood Flow, and Functional Activity: A Hybrid PET/fMRI Study.” Molecular Neurobiology 58, no. 6: 2862–2873. [DOI] [PubMed] [Google Scholar]
  62. Xu, K. , Niu N., Li X., et al. 2023. “The Characteristics of Glucose Metabolism and Functional Connectivity in Posterior Default Network During Nondemented Aging: Relationship With Executive Function Performance.” Cerebral Cortex 33, no. 6: 2901–2911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Yellen, G. 2018. “Fueling Thought: Management of Glycolysis and Oxidative Phosphorylation in Neuronal Metabolism.” Journal of Cell Biology 217, no. 7: 2235–2246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Zhang, X. X. , Tian Y., Wang Z. T., Ma Y. H., Tan L., and Yu J. T.. 2021. “The Epidemiology of Alzheimer's Disease Modifiable Risk Factors and Prevention.” Journal of Prevention of Alzheimer's Disease 8, no. 3: 313–321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Zhu, W. M. , Neuhaus A., Beard D. J., Sutherland B. A., and DeLuca G. C.. 2022. “Neurovascular Coupling Mechanisms in Health and Neurovascular Uncoupling in Alzheimer's Disease.” Brain 145, no. 7: 2276–2292. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1: Supporting Information.

HBM-46-e70328-s001.pdf (1.4MB, pdf)

Data S2: Supporting Information.

HBM-46-e70328-s002.xlsx (100.7KB, xlsx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES