APOE Genotype Modifies the Association of Fusiform Gyrus Cerebral Metabolic Rate of Oxygen Consumption and Object Naming Performance

Abstract

Background:

The apolipoprotein E (APOE) ɛ4 allele confers risk for age and Alzheimer’s disease related cognitive decline but the mechanistic link remains poorly understood. Blood oxygenation level dependent (BOLD) response in the fusiform gyrus (FG) during object naming appears greater among APOE ɛ4 carriers even in the face of equivalent cognitive performance, suggesting neural compensation. However, BOLD is susceptible to known age and APOE-related vascular changes that could confound its interpretation.

Objective:

To address this limitation, we used calibrated fMRI during an object naming task and a hypercapnic challenge to obtain a more direct measure of neural function – percent change cerebral metabolic rate of oxygen consumption (%ΔCMRO₂).

Methods:

Participants were 45 older adults without dementia (28 ɛ4–, 17 ɛ4+) between the ages of 65 and 85. We examined APOE-related differences in %ΔCMRO₂ in the FG during object naming and the extent to which APOE modified associations between FG %ΔCMRO₂ and object naming accuracy. Exploratory analyses also tested the hypothesis that %ΔCMRO₂ is less susceptible to vascular compromise than are measures of %ΔCBF and %ΔBOLD.

Results:

We observed a modifying role of APOE on associations between FG %ΔCMRO₂ and cognition, with ɛ4 carriers (but not non-carriers) demonstrating a positive association between right FG %ΔCMRO₂ and object naming accuracy.

Conclusion:

Results suggest that the relationship between neural function and cognition is altered among older adult APOE ɛ4 carriers prior to the onset of dementia, implicating CMRO₂ response as a potential mechanism to support cognition in APOE-related AD risk.

Keywords

Aging Alzheimer’s disease calibrated fMRI cerebral blood flow cognition cognitive decline metabolism

INTRODUCTION

The ɛ4 allele of the apolipoprotein E (APOE) gene confers increased risk for cognitive decline and Alzheimer’s disease (AD) dementia [1 –4], but precisely how APOE ɛ4 exerts its negative impacts on the brain is still poorly understood. Elucidating the early alterations in neural function that accompany APOE ɛ4 possession could improve mechanistic understanding of the negative effects of APOE ɛ4 on cognition and inform interventions aimed at preventing or delaying cognitive decline and AD progression. Although the assessment of cognitive change in AD risk has largely focused on episodic memory, mounting evidence suggests that semantic knowledge and word-finding ability may be amongst the earliest cognitive abilities impacted in the preclinical phase of this disease [5 –8]. In addition to impacts in preclinical AD, evidence suggests semantic and word finding alterations in cognitively normal APOE ɛ4 carriers, relative to non-carriers [9]. In fact, even when cognitively normal ɛ4 carriers show equivalent object naming performance (relative to non-carriers), they demonstrate greater BOLD activation in the fusiform gyrus (FG), a focal point for convergence and integration of visual semantic information subserving object naming [10]. These findings suggest that neural activation in the fusiform gyrus during object naming may be altered prior to objective cognitive decline and could represent a very early biomarker of AD-risk.

Neural activation is most commonly assessed with functional magnetic resonance imaging (fMRI) techniques such as the blood oxygen level dependent (BOLD) response. However, despite being the most widely used technique, BOLD fMRI does not directly measure neural activity, but reflects local changes in deoxyhemoglobin content, which in turn exhibits a complex dependence on changes in cerebral blood flow (CBF), cerebral blood volume (CBV), and the cerebral metabolic rate of oxygen consumption (CMRO₂) [11 –13]. Please see Fig. 1 for a visual depiction of these relationships. Therefore, factors that affect CBF or the coupling between CBF and CMRO₂— such as cerebrovascular compromise— may alter the BOLD response even when neural activity is unchanged [12 , 15]. This limitation of the BOLD response is particularly relevant to fMRI studies of genetic risk for AD, as APOE is thought to play a role in cerebrovascular integrity, with ɛ4 carriers demonstrating differences in resting CBF, relative to non-carriers [16, 17]. Therefore, apparent APOE ɛ4-related CBF alterations could produce BOLD measurements that do not accurately reflect neural activity. Results from a study by Fleischer and colleagues largely support this notion, demonstrating that cognitively intact middle-aged individuals at APOE ɛ4-related (and familial) risk for AD show elevated resting medial temporal lobe (MTL) CBF, which significantly influenced apparent differences in BOLD activations during an associative encoding task [18]. Bangen and colleagues extended this prior work, comparing resting CBF, %ΔCBF, and %ΔBOLD during a memory encoding task among a group of older adults without dementia [19]. They reported similar increases in resting MTL CBF among ɛ4 carriers, relative to non-carriers, but found no apparent group differences in MTL %ΔBOLD or %ΔCBF. These findings suggest that alterations in resting CBF may not fully account for APOE ɛ4-related differences in %ΔBOLD and that ɛ4 carriers may experience alteration in other neurovascular or neurometabolic variables. Notably, CMRO₂ is thought to be a more direct measure of neural function, as neurons necessarily expend energy to accomplish their work [20], and to be less susceptible to vascular compromise than are measures of BOLD and CBF [13, 21]. Therefore, calculation of CMRO₂, using calibrated fMRI [22], may provide more meaningful information regarding potential APOE ɛ4-related alterations in neural activity.

Fig. 1

BOLD Signal Path-adapted from Liu et al. (2013) [35]. Increases in neural activity give rise to increases in CBF and CMRO2. Cerebral blood volume (CBV) typically increases with CBF. Increases in CBF tend to decrease the amount of deoxyhemoglobin (dHB) content in the microvasculature, while increases in CMRO₂ and CBV tend to increase dHb. For most stimuli, the effect of the CBF increases dominates, leading to an overall decrease in dHb and an increase in the BOLD signal. + denotes a positive relationship and - denotes a negative relationship.

To date, only a handful of studies have used calibrated fMRI to explore the CMRO₂ response during higher cognitive functions, such as memory and executive functioning [23 –25] and none, to our knowledge, have applied this methodology to those at genetic risk for AD. Moreover, most CMRO₂ investigations have focused on activations in the MTL during memory encoding tasks, leaving CMRO₂ activation in the FG during language tasks, such as object naming, largely uncharacterized. Additionally, prior studies have not explored associations between CMRO₂ activation and objective cognitive performance, which could provide critical information as to whether APOE ɛ4-related alterations in neural function might represent compensatory mechanisms invoked to maintain cognitive performance. For example, positive associations between %ΔCMRO₂ and cognitive performance would support the notion of successful compensation whereas negative correlations would support impending neuronal disruption [26]. Therefore, there is a need to extend studies of CMRO₂ response in genetic risk for AD outside the MTL during non-memory cognitive tasks (i.e., object naming), and in relation to objective cognitive performance. To bridge this gap in the literature, the current study used calibrated fMRI to assess potential group differences (APOE ɛ4 + /-) in FG %ΔCMRO₂ during an object naming task and examine the extent to which APOE genotype modifies associations between FG %ΔCMRO₂ and objective naming accuracy among a sample of older adults without dementia. We hypothesized that APOE ɛ4 carriers would demonstrate greater FG %ΔCMRO₂ (relative to non-carriers) and that increased FG %ΔCMRO₂ among ɛ4 carriers (but not non-carriers) would be associated with better naming performance. To better characterize our findings, exploratory analyses also investigated the role of FG %ΔCBF and FG %ΔBOLD in these same models and tested the hypothesis that %ΔCMRO₂ is less susceptible to vascular compromise than are measures of %ΔCBF and %ΔBOLD.

MATERIALS AND METHODS

Participants

See Table 1 for participant demographic and cognitive characteristics. Participants were 45 primarily Caucasian (89%), community-dwelling older adult volunteers enrolled in ongoing research studies at the VA San Diego Healthcare System (VASDHS) and the University of California, San Diego (UCSD) between the ages of 65 and 85 (mean age = 73), 28 of whom were APOE ɛ4– (ɛ3/ ɛ3: n = 27, ɛ2/ ɛ3: n = 1) and 17 of whom were APOE ɛ4+ (ɛ3/ɛ4: n = 16, ɛ4/ɛ4: n = 1). Potential participants were excluded if they had a previous dementia diagnosis or if overall performance on the Dementia Rating Scale (DRS) was more than 1 standard deviation below age-appropriate norms (see Supplementary Table 1 for specific cognitive tests, domains, and normative data). In addition, participants were excluded if they had a history of severe head injury, uncontrolled hypertension, or a DSM-IV Axis I diagnosis of learning disability, attention deficit disorder, mood disorder, substance abuse, or had contraindications to MRI scanning. All participants provided written informed consent prior to enrollment and data were collected in accordance with all ethical standards as stipulated by the UC San Diego and VA San Diego Healthcare System institutional review board-approved procedures.

Table 1

Participant demographic and assessment characteristics

ɛ4– (n = 28)			ɛ4+ (n = 17)
Variable	Mean	SD	Mean	SD	t or χ ²	df	p
Age (y)	71.89	5.47	73.70	6.55	1.00	43	0.3224
Sex (male/female)	5/23	–	4/13	–	0.21	1	0.6447
Education (y)	15.64	2.34	15.70	2.39	0.08	43	0.9313
FSRP Stroke Risk %	8.41	6.58	9.94	6.86	0.74	42	0.4631
Left FG %ΔCMRO₂	12.50	2.3	13.10	2.5	0.73	37	0.4651
Right FG %ΔCMRO₂	41.36	26.6	46.06	39.1	0.45	38	0.6544
Left FG %ΔCBF	25.22	16.7	31.94	22.5	1.06	37	0.2954
Right FG %ΔCBF	55.02	29.8	59.3	31.8	0.40	36	0.6869
Left FG %ΔBOLD	1.02	0.28	1.04	0.36	0.17	41	0.8597
Right FG %ΔBOLD	1.17	0.37	1.20	0.47	0.23	39	0.8175
Left FG M	12.2	11.2	7.68	5.78	1.36	35	0.1812
Right FG M	5.42	15.8	7.01	12.4	0.31	36	0.7550
Left FG n	1.72	0.59	1.84	0.46	0.58	29	0.5647
Right FG n	1.49	0.69	1.56	0.69	0.29	38	0.7697
Cognitive Status (normal/MCI)	26/2	–	14/3	–	1.18	1	0.2770
DRS Total Score	141.2	2.33	141.1	2.16	0.06	42	0.9483
Object Naming Percent Correct	92.38	6.51	90.82	8.32	0.69	42	0.4903
BNT/MiNT Percent Correct	94.93	5.11	94.44	6.23	0.28	43	0.7757

FSRP, Framingham Stroke Risk Profile; %ΔCMRO₂, percent change cerebral metabolic rate of oxygen consumption; FG, fusiform gyrus; DRS, Mattis Dementia Rating Scale; df, degrees of freedom; all scores represent raw scores unless otherwise stated.

Apolipoprotein E genotyping

Genotyping was performed by the ADCS Biomarker Core at UCSD using real time PCR Restriction Fragment Length Polymorphism analysis. Genomic DNA was collected from participants using buccal swab and extracted using Qiamp DNA blood mini kit (Qiagen) followed by PCR amplification [27]. Those with at least one ɛ4 allele (i.e., ɛ2/ɛ4, ɛ3/ɛ4, ɛ4/ɛ4) were classified as APOE ɛ4 carriers (ɛ4+) and those without an ɛ4 allele (i.e., ɛ2/ɛ2, ɛ2/ɛ3, ɛ3/ɛ3) were classified as non-carriers (ɛ4–). Given that participants were grouped by APOE status, they were not randomly assigned to groups. However, investigators were blind to group allocation when collecting outcome measures (i.e., MRI, cognition).

Cognitive testing

Participants underwent a comprehensive neuropsychological assessment that includes measures of global cognitive functioning, attention, language, learning and memory, visuospatial functioning, executive functioning, simple motor functioning, and practical measures of independent living skills (see Supplementary Table 1 in the supplement for specific cognitive tests, domains, and normative data). Participants were diagnosed with mild cognitive impairment (MCI) if performance on more than one measure within a cognitive domain was more than one standard deviation below age-appropriate norms, consistent with the empirically-derived criteria for diagnosis of MCI developed by Jak and colleagues [28]. The primary cognitive measure of interest was an object naming task that was performed inside and outside the MRI scanner (see Functional Task). Post hoc analyses also explored associations with standardized clinical (versus research-based) measures of confrontation naming (i.e., Boston Naming Test [BNT], Multilingual Naming Test [MiNT]). Due to mid-study changes in the neuropsychological battery, 33 participants completed the BNT and 12 completed the MiNT. Of note, there were no significant group differences (e.g., APOE ɛ4 status, DRS, CMRO₂, BOLD, CBF) between those who completed the BNT and those who completed the MiNT. In order to facilitate the exploration of standardized clinical confrontation naming performance across the entire sample, we created a clinical naming score that was defined as percent correct naming performance on either of these two measures.

Functional task

During calibrated fMRI, participants performed a picture naming task of visually presented living and non-living objects (animals, tools, and vehicles) balanced for word frequency. This naming task was developed in house and is known to reliably activate the FG bilaterally in healthy older adults and those with dementia [10 , 30]. Each run contained 6 active blocks with 10 trials presented (2.8 s each) per block. Each 60-trial run was 364 seconds in length, i.e., 28 s HEAD + (28 s ON, 28 s OFF) * 6 and acquired 104 functional images (364 s/3.5 s TR) for each slice. Participants completed 4 runs for a total of 240 trials or a total run time of 24:16 min. The order of runs was counterbalanced for each participant. A 28-s fixation block preceded and followed each active block to allow the hemodynamic response to return to baseline for block comparisons, during which participants were instructed not to think any words to themselves, to rest quietly, and to look at a fixation cross displayed in the center of the screen. Stimuli were presented using E-Prime software via an LCD projector on a screen at the end of the scanner bed with stimuli viewed through a mirror mounted on the head coil. A limitation of the blocked design is that overt naming cannot be done due to task-related motion artifact. Thus, responses were covert (i.e., silently naming the object in their mind) and behavioral performance on the same naming task was subsequently measured outside the scanner.

Hypercapnic task

A hypercapnic challenge scan was performed prior to the functional task described above. Subjects breathed through a mouthpiece connected to a low-deadspace non-rebreathing valve. A three-way valve on the inspired limb of the circuit was used to switch between room air (0% CO₂) and CO₂ enriched gas (5% CO₂, 21% O₂, balance N₂). Inspired gas was dispensed into a small Douglas bag connected to the inspired limb of the breathing circuit. The hypercapnic task consisted of 2 min of room air, followed by 3 min of 5% CO₂, and 2 min of room air for 2 runs.

Calibrated fMRI acquisition

Participants were scanned on a 3T Discovery MR750 system (GE Healthcare, Waukesha, WI) with an 8-channel receive-only head coil at the UCSD Center for Functional MRI. Both hypercapnic and functional scans were carried out using a pseudocontinuous ASL (PCASL) sequence with dual echo 2D single-shot spiral readout (TR = 3.5 s, TE1 = 3.2 ms, TE2 = 30 ms, labeling duration = 1.5 s, post label delay = 1.5 s, FOV = 22 cm, acquisition matrix = 64×64, slice thickness = 6 mm, 9 contiguous slices) to simultaneously acquire functional CBF and BOLD data. In order to ensure optimal labeling efficiency for the duration of the object naming and hypercapnia scans, a set of 3-min prescan calibrations were performed using the optimized PCASL (OptPCASL) technique [31] whereby the compensatory RF phase term and in-plane gradient amplitudes for the PCASL pulse train were determined. Nine prescribed slices were carefully aligned with the fusiform gyrus and extended superiorly to above the lateral inferior frontal gyrus. Cardiac pulse and respiratory effort data were monitored using a pulse oximeter and a respiratory effort transducer, respectively. The pulse oximeter was placed on the subject’s right index finger. The respiratory effort belt was placed around the subject’s abdomen. Physiological data were sampled at 40 samples per second using a multi-channel data acquisition board. In addition to the physiological data, scanner TTL pulse data were recorded at 1 kHz. The TTL pulse data were used to synchronize the physiological data to the acquired images. Cardiac, respiratory, and TTL data were used to calculate the physiological noise regressors in P.

Following the tasks, a 36-s spiral scan was acquired with a long TR (4000 ms) and short TE (3.4 ms) to estimate the equilibrium magnetization of cerebral spinal fluid, which was used to convert the perfusion signal into calibrated CBF units (ml blood/100 g tissue/min). A 32-s minimum contrast image was also acquired to remove B1 transmit and receive coil inhomogeneities from the task data. Finally, a field map was acquired, which was incorporated during the spiral image reconstruction to minimize signal bunching and dropouts from the acquired data [32].

Data analysis

Data processing was performed using neuroimaging routines from AFNI (afni.nimh.nih.gov), FSL (Oxford, United Kingdom) and FreeSurfer (surfer.nmr.mgh.harvard.edu), and called from locally created MATLAB scripts. Anatomical volumes were aligned to the functional data at the subject level and all functional images were motion corrected. Data from the two echoes were analyzed separately, with the first and second echo data used to analyze CBF and BOLD signals, respectively. Functional ASL images were computed from the running subtraction of the control and tag series of the first echo, while the BOLD images were derived from the running average (average of each image with the mean of its two nearest neighbors) of the second echo [33] time series. ASL and BOLD runs were averaged to form one time series per voxel for each type of scan and the data were analyzed with a general linear model (GLM) framework with physiological noise and low frequency terms as covariates of no interest. More specifically, noise correction was based on a general linear model of the perfusion signal, which is estimated using a filtered subtraction of the tag and control images, assuming the relative weighting of the physiological noise differs between the tag and control images. Please see Restom et al. [34] for more details. To estimate functional CBF and BOLD response, %ΔCBF and %ΔBOLD response was calculated by normalizing each time course to its baseline for the naming and hypercapnia tasks. Left and right FG ROIs were defined using the Talairach atlas and each ROI was treated as a search region. Small volume correction was determined with Monte Carlo simulations (via AFNI 3dClustSim) to guard against false-positive results, with an a priori individual voxel probability of p < 0.05 and a posteriori ROI-wise probability of p < 0.05. Furthermore, to be included in the analyses, at least one activated CBF voxel was required to be contiguous with an active BOLD voxel to form a significant cluster. To estimate CMRO₂ changes with functional activation, a calibration factor (M) was applied to the CBF and BOLD measurements from the functional naming task [22]. More specifically, the fractional BOLD signal change (ΔS/S₀) is related to the underlying changes in CBF and CMRO₂ through the following equation: b = (ΔS/S₀) = M(1 –f ^α–^βm^β), where f = CBF/CBF₀ and m = CMRO₂/CMRO₂₀ represent the physiological quantities normalized by their respective baseline values [35]. The unitless parameter M defines the maximum possible BOLD signal change for a brain region and can be written as M = A(CBV₀) (TE) ([dHB]₀)^β, where A is a multiplicative factor that depends on magnetic field strength, CBV₀ is the baseline blood volume, TE denotes echo time, and [dHB]₀ is the baseline concentration of deoxyhemoglobin. The additional parameters are determined empirically, but are well approximated as α ≈ 0.4 and β ≈ 1.5 [12, 35]. Of note, sensitivity analysis showed that findings held when assuming different values of these parameters (i.e., α ≈ 0.14 and β ≈ 0.91) [35]. Coupling ratio n is defined as n = (f – 1)/(m – 1) and reflects the strength of the coupling between the CBF and CMRO2 responses [22, 35]. Prior to group analyses, brain imaging data points falling above or below 3 standard deviations from the mean were removed to reduce outlier effects, resulting in the removal of approximately 6% of the %ΔCMRO₂ data, 3% of the %ΔCBF data, and 1% of the %ΔBOLD data.

Statistical analyses

Students t-tests and χ²-tests were used to compare groups (APOE ɛ4 +/–) on demographic (age, sex, education) clinical (stroke risk percent), cognitive (cognitive status, DRS, object naming, BNT/MiNT), and brain variables (%ΔCMRO_2, %ΔCBF, %ΔBOLD, M, and n in the LFG and RFG). Results were considered significant at p < 0.003 (Bonferroni corrected for 13 comparisons). Primary analyses employed multiple robust linear regression models in R (lmrob) to explore the association between APOE and left and right FG %ΔCMRO₂ (i.e., APOE → left FG %ΔCMRO₂; APOE → right FG %ΔCMRO₂) and the two-way interaction of APOE and left and right FG %ΔCMRO₂ on object naming performance (i.e., APOE×left FG %ΔCMRO₂→object naming; APOE×right FG %ΔCMRO₂→object naming). These models statistically adjusted for the effects of age, sex, education, cognitive status (normal cognition, MCI), and estimated 10-year risk for future stroke, as measured by the Framingham Stroke Risk Profile (FSRP) [36] and results were considered significant at p < 0.0125 (Bonferroni corrected for 4 comparisons). To further characterize the findings, post hoc analyses also investigated the role of left and right FG %ΔCBF and %ΔBOLD within this same model framework (APOE×left FG %ΔCBF→object naming; APOE×right FG %ΔCBF→object naming; APOE×left FG %ΔBOLD→object naming; APOE×right FG %ΔBOLD→object naming), statistically adjusting for the effects of age, sex, education, cognitive status, and stroke risk (p < 0.0125, Bonferroni corrected for 4 comparisons). Post hoc analyses also explored the associations of these same variables with respect to performance on a clinical measure of confrontation naming (i.e., APOE×left FG %ΔCMRO₂→BNT/MiNT; APOE×right FG %ΔCMRO₂→BNT/MiNT), statistically adjusting for the same variables described above, and which of the two naming measures was administered (i.e., BNT versus MiNT; see Cognitive Testing). Results were considered significant at p < 0.025 (Bonferroni corrected for 2 comparisons). Exploratory analyses also examined the interaction of APOE and FSRP stroke risk percent on average bilateral FG %ΔCMRO₂, %ΔCBF, and %ΔBOLD, statistically adjusting for the effects of age, sex, and cognitive status (i.e., p < 0.016, Bonferroni corrected for 3 comparisons). Multicollinearity between independent variables was examined by application of the multicollinearity index VIF (all VIF < 2), linearity was examined through inspection of residuals versus fits plots, normality was examined through inspection of Q-Q plots, non-heteroskedasticity was examined through inspection of scale location plots, and influential cases/outliers were examined through inspection of residuals versus leverage plots.

Data and code availability statements

Data that was newly acquired in this study is not able to be made openly available to the public, as we did not gain approval from the local IRB to disclose this information. No new code was developed for the execution of this study.

RESULTS

Group differences in demographic, cognitive, and brain variables

APOE groups (ɛ4+, ɛ4–) did not differ significantly on age, sex, years of education, or stroke risk, nor did they differ significantly on any of the brain imaging or cognitive variables (see Table 1).

APOE and FG %ΔCMRO₂

There were no significant associations between APOE status and %ΔCMRO₂ in the left or right FG. Similarly, post hoc analyses did not reveal associations between APOE and %ΔCBF or %ΔBOLD in the left or right FG.

Interaction of APOE and FG %ΔCMRO₂ on object naming

There was a significant interaction of APOE and %ΔCMRO₂ in the right FG on object naming performance (see Table 2), whereby higher %ΔCMRO₂ during object naming was associated with better object naming performance among ɛ4 carriers but not among non-carriers (see Fig. 2). Post hoc analysis revealed that results held when including left and right FG %ΔCMRO₂ in the same model. Results were not significant in the left FG. Post hoc analyses with clinical naming measures, revealed similar findings, whereby higher right FG %ΔCMRO₂ during object naming was associated with better BNT/MiNT performance among ɛ4 carriers but not among non-carriers (see Supplementary Figure 1). Post hoc analyses did not reveal significant interactions between APOE and %ΔCBF or %ΔBOLD in the right or left FG on object naming performance.

Table 2

Interaction of APOE and FG %ΔCMRO₂ on object naming performance

DV: Object Naming Percent Correct	β	se	t-statistic	p
Right FG
APOE	– 0.316	0.240	1.318	0.198
Right FG %ΔCMRO₂	– 0.016	0.093	0.177	0.860
APOE * Right FG %ΔCMRO₂	0.815	0.149	5.47	<0.001*
Left FG
APOE	0.056	0.266	0.211	0.834
Left FG %ΔCMRO₂	0.041	0.173	0.238	0.814
APOE * Left FG %ΔCMRO₂	0.133	0.264	0.502	0.619

Only variables of interest are included in table; FG = fusiform gyrus; APOE, apolipoprotein E gene; %ΔCMRO₂, percent change cerebral metabolic rate of oxygen consumption; β, standardized regression coefficient; se, standard error; *significance at p < 0.025.

Fig. 2

Interaction of APOE and right FG %ΔCMRO₂ on object naming performance. Greater %ΔCMRO₂ in the right FG during an object naming task was associated with better object naming performance among APOE ɛ4 carriers but not among non-carriers. APOE, apolipoprotein E gene; %ΔCMRO₂, percent change cerebral metabolic rate of oxygen consumption; RFG, right fusiform gyrus; p, p-value; ns, non-significant simple slope.

Vascular risk, APOE, and FG %ΔCMRO₂, %ΔCBF, and %ΔBOLD

Results revealed a significant main effect of vascular risk on FG %ΔBOLD, such that FSRP stroke risk percent was negatively associated with FG %ΔBOLD among ɛ4 carriers and non-carriers (see Fig. 3 and Table 3). An interaction of vascular risk and APOE on %ΔBOLD was not observed. An interaction of vascular risk and APOE on %ΔCBF was observed, such that FSRP stroke risk percent was positively associated with bilateral FG %ΔCBF among ɛ4 carriers, but not among non-carriers. However, this did not survive multiple comparison correction. There was neither a main effect of vascular risk, nor an interaction of vascular risk and APOE, on FG % CMRO₂.

Fig. 3

Association of vascular risk and calibrated fMRI variables. FSRP stroke risk percent was not associated with BFG %ΔCMRO₂ among ɛ4 carriers or non-carriers. However, FSRP stroke risk percent was negatively associated with BFG %ΔBOLD among ɛ4 carriers and non-carriers and positively associated with BFG %ΔCBF among ɛ4 carriers, but not non-carriers. FSRP, Framingham Stroke Risk Profile; %ΔCMRO₂, percent change cerebral metabolic rate of oxygen consumption; %ΔCBF, percent change cerebral blood flow; %ΔBOLD, percent change blood oxygen level dependent; BFG, bilateral fusiform gyrus; p, p-value; ns, non-significant simple slope.

Table 3

Association of vascular risk and calibrated fMRI variables

Variable	β	se	t-statistic	p
DV: BFG % Δ CMRO₂
APOE	0.43	0.328	1.318	0.198
FSRP Stroke Risk %	0.18	0.234	0.798	0.431
APOE * FSRP Stroke Risk %	0.465	0.267	1.740	0.093
DV: BFG % Δ CBF
APOE	– 0.151	0.197	1.697	0.098
FSRP Stroke Risk %	0.169	0.072	2.409	0.021+
APOE * FSRP Stroke Risk %	0.407	0.269	2.135	0.039+
DV: BFG % Δ BOLD
APOE	– 0.223	0.391	0.342	0.734
FSRP Stroke Risk %	– 0.321	0.137	2.549	0.015*
APOE * FSRP Stroke Risk %	– 0.164	0.275	– 0.247	0.806

Only variables of interest are included in table; DV, dependent variable; APOE, apolipoprotein E gene; FSRP, Framingham Stroke Risk %ΔCMRO₂, percent change cerebral metabolic rate of oxygen consumption; %ΔCBF, percent change cerebral blood flow; %ΔBOLD, percent change blood oxygen level dependent; BFG, bilateral fusiform gyrus; β, standardized regression coefficient; se, standard error; +marginal significance at p < 0.05; *statistical significance at p < 0.0166.

DISCUSSION

This is the first study to examine %ΔCMRO₂ in the FG during object naming in older APOE ɛ4 carriers without dementia to elucidate mechanisms of cognitive function associated with AD risk. This study was motivated in part by findings from a prior fMRI BOLD study [10], which suggested that neural activation in the FG during object naming may be altered among ɛ4 carriers prior to objective changes in cognition, and thus could serve as a possible biomarker of incipient cognitive decline. Due to potential vascular confounds of the BOLD response, we attempted to confirm these prior findings by measuring FG %ΔCMRO₂ during an object naming task, as this measure is thought to be a more direct assessment of neural function that is less susceptible to vascular compromise than the BOLD signal. Although we did not find evidence of APOE-related differences in left or right FG %ΔCMRO₂, we observed a modifying role of APOE on the relationship between right FG %ΔCMRO₂ and cognition, such that ɛ4 carriers demonstrated a positive association between right FG %ΔCMRO₂ and object naming performance, whereas this association was not evident among non-carriers. It should be noted that this rather large effect occurred in the context of equivalent object naming performance (and FG %ΔCMRO₂). Results suggest that the relationship between neural activation and cognitive performance is altered among ɛ4 carriers prior to the onset of dementia and implicate CMRO₂ response as a potential compensatory mechanism to support cognition in APOE-related AD risk.

We propose that our novel finding of APOE-related differences in the relationship between right FG %ΔCMRO₂ and language function may be understood in the context of neurometabolic compensation. Older adults experience age-related losses in the number of available neural processing units (e.g., neurons or synapses) [37], decreases in computational efficiency, and a reduction in the specificity of neural activation [38, 39], including in response to naming tasks [29]. Moreover, brain aging and APOE ɛ4 possession appear to have a synergistic negative effect on brain function [40]. Therefore, the brains of older adult ɛ4 carriers may require a compensatory increase in oxygen consumption rate per unit tissue mass in order to maintain a similar level of cognitive functioning, when compared to their non-carrying counterparts. Against this backdrop, we speculate that the observed positive relationship between neurometabolic activation and cognition among ɛ4 carriers (higher %ΔCMRO₂ being associated with better object naming performance) could reflect effective attempts to maintain adequate brain oxygenation in the face of accelerated brain aging and/or neuropathological damage related to the APOE ɛ4 allele, whereas the lack of evidence for this relationship among non-carriers could reflect the absence of neurometabolic compensatory processes in this group. These findings seem to complement results previously reported by our group suggesting the presence of cerebrovascular (i.e., resting CBF) compensatory mechanisms in older adult APOE ɛ4 carriers [41 –43] and add to this literature by providing evidence of potential neurometabolic (i.e., CMRO₂) compensatory processes being invoked in this same AD risk group.

With regard to the potential laterality of findings (i.e., significant findings in right FG and not left FG), it is conceivable that APOE and/or AD related compensatory processes that support language function are first evident in the right hemisphere of the brain. Generally supporting this notion, over-recruitment of right MTL activity has been reported during encoding and consolidation in adults at genetic risk for AD [44, 45] and in individuals with MCI [46], though little is known about the brain’s response during language tasks. It is also possible that laterality of the findings is reflective of insufficient statistical power to detect findings in the left hemisphere. Future studies should seek to directly explore laterality in this context.

Although the study design was motivated by data suggesting that measures of %ΔCMRO₂ are likely less sensitive to vascular compromise than are measures of %ΔCBF and %ΔBOLD, this is the first study (to our knowledge) to test this hypothesis in a sample of older adults. We found that neither ɛ4 carriers nor non-carriers demonstrated significant associations between vascular risk and %ΔCMRO₂ during object naming, whereas ɛ4 carriers (but not non-carriers) showed an association between vascular risk and %ΔCBF, and both ɛ4 carriers and non-carriers showed an association between vascular risk and %ΔBOLD. Taken together these data support the notion that %ΔCMRO₂ is indeed less sensitive to vascular risk than are measures of %ΔCBF and %ΔBOLD and provides evidence supporting its use as a measure of neural function in genetic risk for AD that is less confounded by vascular compromise. Future studies should utilize larger samples of APOE ɛ4 carriers and more robust markers of cerebrovascular risk to replicate and validate these findings.

As mentioned previously, we did not observe APOE-related differences in %ΔCMRO₂, as was hypothesized based on findings from an object naming BOLD study by Wierenga et al. (2010) [10]. Given that %ΔCMRO₂ was less sensitive to vascular risk in our current sample, it is possible that APOE-related differences in %ΔBOLD in this prior study were more reflective of underlying APOE-related vascular alteration. However, post hoc analyses in the current sample did not reveal APOE-related differences in %ΔBOLD or %ΔCBF either. Alternatively, it is possible that APOE-related differences in the associations between neural activation and cognition are apparent prior to differences in each measure independently, and thus group differences in neural activation were not yet evident in our sample, which was approximately 5 years younger on average when compared to the previous study. Notably, our group has consistently observed APOE-related differences in associations between CBF and cognition in the face of equivalent CBF values across cognitively normal ɛ4 carriers and non-carriers [41, 42]. Of note, more well-powered calibrated fMRI studies are needed to characterize the range of %ΔCMRO₂ values expected in response to tasks of higher cognitive functions. For example, based on the several studies that have explored these relationships, average %ΔCMRO₂ values seem to hover around 25%. However, the current study revealed an average %ΔCMRO₂ of 35% when averaged across hemispheres and APOE status, which is slightly higher than would be expected based on previous reports. This could suggest that changes in CMRO₂ vary by age, APOE ɛ4 status, and/or AD progression. Current theoretical models of how %ΔCMRO₂ changes in relation to these variables are lacking. As such, future studies should seek to characterize longitudinal changes in %ΔCMRO₂ as well as other calibrated fMRI variables (e.g., %ΔCBF, M, n) among heterogeneous clinical samples to better understand how changes over time in particular clinical groups may contribute to the development of cognitive impairment and/or AD.

It is important to discuss the potential strengths and limitations of the calibrated-BOLD approach to measure CMRO₂. This approach [22] measures both BOLD and CBF responses to breathing CO₂ in order to exploit the fact that the ASL signal depends only on CBF changes, while the BOLD signal depends on CBF and CMRO₂ changes and that mild hypercapnia is thought to produce a large change in CBF without changing CMRO₂ [47]. As described earlier, the fractional BOLD signal change (ΔS/S₀) is related to the underlying changes in CBF and CMRO₂ through the following equation: b = (ΔS/S₀) = M(1 – (CBF/CBF₀)^{α - β} (CMRO₂/CMRO₂₀)^β) (see Fig. 1). However, it should be noted that there is still some controversy regarding this latter assumption [48]. Although there may be additional limitations related to other assumptions within the Davis model, estimates of fractional CMRO₂ change in a calibrated-BOLD experiment appear to be relatively robust to uncertainties (e.g., values of α and β in the Davis equation), primarily because the same model is used to analyze both the hypercapnia experiment and the activation experiment, which is thought to lead to self-correcting [22, 49]. Another reason for the apparent robustness of the Davis model is that the physiological variables that have strong effects on BOLD response magnitude are essentially lumped into the parameter M. This includes physiological parameters (e.g., baseline venous blood volume, hematocrit), aspects of the image acquisition (e.g., echo time, field strength), and even the way that a region of interest is selected for averaging [50]. For this reason, the direct estimation of M in calculation of CMRO₂ response in the current study represents a significant strength over other methods that assume a value of M reported in the literature and apply it to data collected in a different way [51, 52]. Another limitation with respect to calibrated fMRI is that this technique does not provide a measure of baseline CMRO₂, which could improve our understanding of functional CMRO₂ response. To date, the literature regarding changes in baseline CMRO₂ with advancing age is somewhat mixed, with most studies showing age-related decreases [53 –55] and some showing increases [56]. More general limitations of the current study include a relatively small sample size and a cross-sectional design, which restricted our ability to draw causal conclusions or infer directionality. As a first step in validating this approach in the current sample, we utilized an ROI approach and future studies should seek to explore these relationships using a voxel-wise whole-brain approach.

Our sample was also characterized by relatively high levels of education and was mainly (89%) Caucasian, and although these demographic factors were not associated with APOE genotype, our findings cannot be generalized to other racial/ethnic groups or to those with lower educational attainment. Of note, our sample size was relatively small, particularly with respect to APOE ɛ4 carriers (n = 17) and those with MCI (n = 5), which decreased our statistical power. Larger sample sizes are needed to have sufficient power to determine whether findings differ by cognitive status. Future studies should seek to replicate these findings in larger MCI and cognitively normal samples. Despite these limitations, it should be noted that studies of CMRO₂ in higher cognitive functions are very limited. In fact, this is the first study to explore %ΔCMRO₂ in the context of genetic risk for AD and to directly test the hypothesis that %ΔCMRO₂ is less sensitive to vascular compromise than are measures of %ΔCBF or %ΔBOLD. Additional strengths include our inclusion of a well-controlled and well-characterized sample of older adults without dementia, which included the use of several cognitive test performances to characterize cognitive status.

In conclusion, we report novel findings from calibrated fMRI suggesting that the relationship between neural function (FG %ΔCMRO₂) and cognition (naming performance) is altered among older adult APOE ɛ4 carriers prior to the onset of dementia and implicates CMRO₂ response as a potential mechanism to support cognition in APOE-related AD risk. Results also suggest that %ΔCMRO₂ is less sensitive to vascular risk than are measures of %ΔCBF and %ΔBOLD, offering support for its use among APOE ɛ4 carriers and other populations for whom vascular compromise is suspected, though future research is needed to confirm this. Future studies should utilize longitudinal designs with larger more diverse samples to clarify the temporal sequence of changes among these different brain variables. Future investigations should also explore additional calibrated fMRI variables (e.g., oxygen extraction fraction, baseline CMRO₂) and AD biomarkers (e.g., CSF tau and amyloid) to improve understanding of the dynamic mechanisms underlying APOE-related changes in neural activity.

Footnotes

ACKNOWLEDGMENTS

We would like to acknowledge the work of Dr. Rick Buxton, who’s work in this area was critically influential in guiding the conceptualization and methods of this study.

FUNDING

Research reported in this publication was supported by VA CSR&D Merit Award [5I01CX000565 C.E.W.], the National Institute on Aging of the National Institutes of Health [K23AG049906 Z.Z.Z.], and the VA San Diego Healthcare System CESAMH and VA CSR&D CDA Award [IK2CX002335 C.C.H.W.].

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the findings of this study are available on request from the corresponding author.

The supplementary material is available in the electronic version of this article: .

References

Bretsky

, Guralnik

, Launer

, Albert

, Seeman

, MacArthur

Studies of Successful Aging

(2003) The role of APOE-epsilon4 in longitudinal cognitive decline: MacArthur Studies of Successful Aging. Neurology 60,1077–1081.

Caselli

, Dueck

, Osborne

, Sabbagh

, Connor

, Ahern

, Baxter

, Rapcsak

, Shi

, Woodruff

, Locke

DEC

, Snyder

, Alexander

, Rademakers

, Reiman

(2009) Longitudinal modeling of age-related memory decline and the APOE epsilon4 effect. N Engl J Med 361,255–263.

Schiepers

OJG

, Harris

, Gow

, Pattie

, Brett

, Starr

, Deary

(2012) APOE E4 status predicts age-related cognitive decline in the ninth decade: Longitudinal follow-up of the Lothian Birth Cohort 1921. Mol Psychiatry 17,315–324.

Tsuang

, Leverenz

, Lopez

, Hamilton

, Bennett

, Schneider

, Buchman

, Larson

, Crane

, Kaye

, Kramer

, Woltjer

, Trojanowski

, Weintraub

, Chen-Plotkin

, Irwin

, Rick

, Schellenberg

, Watson

, Kukull

, Nelson

, Jicha

, Neltner

, Galasko

, Masliah

, Quinn

, Chung

, Yearout

, Mata

, Wan

, Edwards

, Montine

, Zabetian

(2013) APOE ɛ4 increases risk for dementia in pure synucleinopathies. JAMA Neurol 70,223–228.

Blackwell

, Sahakian

, Vesey

, Semple

, Robbins

, Hodges JR (2004) Detecting dementia: Novel neuropsychological markers of preclinical Alzheimer’s disease. Dement Geriatr Cogn Disord 17,42–48.

Clark

, Gatz

, Zheng

, Chen

Y-L

, McCleary

, Mack

(2009) Longitudinal verbal fluency in normal aging, preclinical, and prevalent Alzheimer’s disease. Am J Alzheimers Dis Dementiasr 24,461–468.

Mickes

, Wixted

, Fennema-Notestine

, Galasko

, Bondi

, Thal

, Salmon

(2007) Progressive impairment on neuropsychological tasks in a longitudinal study of preclinical Alzheimer’s disease. Neuropsychology 21,696–705.

Papp

, Mormino

, Amariglio

, Munro

, Dagley

, Schultz

, Johnson

, Sperling

, Rentz

(2016) Biomarker validation of a decline in semantic processing in preclinical Alzheimer’s disease. Neuropsychology 30,624–630.

Miller

, Rogers

, Siddarth

, Small

(2005) Object naming and semantic fluency among individuals with genetic risk for Alzheimer’s disease. Int J Geriatr Psychiatry 20,128–136.

10.

Wierenga

, Stricker

, McCauley

, Simmons

, Jak

, Chang

Y-L

, Delano-Wood

, Bangen

, Salmon

, Bondi

(2010) Increased functional brain response during word retrieval in cognitively intact older adults at genetic risk for Alzheimer’s disease. Neuroimage 51,1222–1233.

11.

Buxton

(2009), Introduction to Functional Magnetic Resonance Imaging: Principles and Techniques, Cambridge University Press.

12.

Buxton

, Uludağ

, Dubowitz

, Liu

(2004) Modeling the hemodynamic response to brain activation.S. NeuroImage 23(Suppl 1), 220–233.

13.

Iannetti

, Wise

(2007) BOLD functional MRI in disease and pharmacological studies: Room for improvement? Magn Reson Imaging 25,978–988.

14.

Stefanovic

, Warnking

, Rylander

, Pike

(2006) The effect of global cerebral vasodilation on focal activation hemodynamics. Neuroimage 30,726–734.

15.

Wierenga

, Bondi

(2007) Use of functional magnetic resonance imaging in the early identification of Alzheimer’s disease. Neuropsychol Rev 17,127–143.

16.

Hays

, Zlatar

, Wierenga

(2016) The utility of cerebral blood flow as a biomarker of preclinical Alzheimer’s disease. Cell Mol Neurobiol 36,167–179.

17.

Wierenga

, Hays

, Zlatar

(2014) Cerebral blood flow measured by arterial spin labeling MRI as a preclinical marker of Alzheimer’s disease. J Alzheimers Dis 42(Suppl 4), S411–419.

18.

Fleisher

, Podraza

, Bangen

, Taylor

, Sherzai

, Sidhar

, Liu

, Dale

, Buxton

(2009) Cerebral perfusion and oxygenation differences in Alzheimer’s disease risk. Neurobiol Aging 30,1737–1748.

19.

Bangen

, Restom

, Liu

, Wierenga

, Jak

, Salmon

, Bondi

(2012) Assessment of Alzheimer’s disease risk with functional magnetic resonance imaging: An arterial spin labeling study. J Alzheimers Dis 31(Suppl 3), S59–74.

20.

Hyder

(2004) Neuroimaging with calibrated fMRI. Stroke 35(11 Suppl 1), 2635–2641.

21.

Wierenga

, Lavender

, Hays

(2018) The potential of calibrated fMRI in the understanding of stress in eating disorders. Neurobiol Stress 9,64–73.

22.

Davis

, Kwong

, Weisskoff

, Rosen

(1998) Calibrated functional MRI: Mapping the dynamics of oxidative metabolism. Proc Natl Acad Sci U S A 95,1834–1839.

23.

Mohtasib

, Lumley

, Goodwin

, Emsley

HCA

, Sluming

, Parkes

(2012) Calibrated fMRI during a cognitive Stroop task reveals reduced metabolic response with increasing age. NeuroImage 59,1143–1151.

24.

Restom

, Bangen

, Bondi

, Perthen

, Liu

(2007) Cerebral blood flow and BOLD responses to a memory encoding task: A comparison between healthy young and elderly adults. Neuroimage 37,430–439.

25.

, Wu

, Xu

, Li

S-J

(2006) Decrease in CMRO2 for memory-encoding tasks in the hippocampus of mild cognitive impairment subjects. Proc Int Soc Mag Reson Med 14,1108.

26.

Sperling

, Dickerson

, Pihlajamaki

, Vannini

, LaViolette

, Vitolo

, Hedden

, Becker

, Rentz

, Selkoe

, Johnson

(201) Functional alterations in memory networks in early Alzheimer’s disease. Neuromolecular Med 12,27–43.

27.

Wierenga

, Dev

, Shin

, Clark

, Bangen

, Jak

, Rissman

, Liu

, Salmon

, Bondi

(2012) Effect of mild cognitive impairment and APOE genotype on resting cerebral blood flow and its association with cognition. J Cereb Blood Flow Metab 32,1589–1599.

28.

Jak

, Bondi

, Delano-Wood

, Wierenga

, Corey-Bloom

, Salmon

, Delis

(2009) Quantification of five neuropsychological approaches to defining mild cognitive impairment. Am J Geriatr Psychiatry 17,368–375.

29.

Wierenga

, Benjamin

, Gopinath

, Perlstein

, Leonard

, Rothi

LJG

, Conway

, Cato

, Briggs

, Crosson

(2008) Age-related changes in word retrieval: Role of bilateral frontal and subcortical networks. Neurobiol Aging 29,436–451.

30.

Wierenga

, Stricker

, McCauley

, Simmons

, Jak

, Chang

Y-L

, Nation

, Bangen

, Salmon

, Bondi

(2011) Altered brain response for semantic knowledge in Alzheimer’s disease. Neuropsychologia 49,392–404.

31.

Shin

, Liu

, Wong

, Jung

(2012) Pseudocontinuous arterial spin labeling with optimized tagging efficiency. Magn Reson Med 68,1135–1144.

32.

Noll

, Fessler

, Sutton

(2005) Conjugate phase MRI reconstruction with spatially variant sample density correction. IEEE Trans Med Imaging 24,325–336.

33.

Liu

, Wong

(2005) A signal processing model for arterial spin labeling functional MRI. Neuroimage 24,207–215.

34.

Restom

, Behzadi

, Liu

(2006) Physiological noise reduction for arterial spin labeling functional MRI. Neuroimage 31,1104–1115.

35.

Liu

, Glover

, Mueller

, Greve

, Brown

(2013) An introduction to normalization and calibration methods in functional MRI. Psychometrika 78,308–321.

36.

D’Agostino

, Wolf

, Belanger

, Kannel

(1994) Stroke risk profile: Adjustment for antihypertensive medication. The Framingham Study. Stroke 25,40–43.

37.

Pakkenberg

, Gundersen

(1997) Neocortical neuron number in humans: Effect of sex and age. J Comp Neurol 384,312–320.

38.

Geerligs

, Renken

, Saliasi

, Maurits

, Lorist

(2015) A brain-wide study of age-related changes in functional connectivity. Cereb Cortex 25,1987–1999.

39.

Park

, Polk

, Park

, Minear

, Savage

, Smith

(2004) Aging reduces neural specialization in ventral visual cortex. Proc Natl Acad Sci U S A 101,13091–13095.

40.

Glorioso

, Pfenning

, Lee

, Bennett

, Sibille

, Kellis

, Guarente

(2019) Rate of brain aging and APOE ɛ4 are synergistic risk factors for Alzheimer’s disease. Life Sci Alliance 2,e201900303.

41.

Hays

, Zlatar

, Meloy

, Bondi

, Gilbert

, Liu

, Helm

, Wierenga

(2019) APOE modifies the interaction of entorhinal cerebral blood flow and cortical thickness on memory function in cognitively normal older adults. Neuroimage 202,116162.

42.

Hays

, Zlatar

, Meloy

, Bondi

, Gilbert

, Liu

, Helm

, Wierenga

(2020) Interaction of APOE, cerebral blood flow, and cortical thickness in the entorhinal cortex predicts memory decline. Brain Imaging Behav 14,369–382.

43.

Zlatar

, Bischoff-Grethe

, Hays

, Liu

, Meloy

, Rissman

, Bondi

, Wierenga

(2016) Higher brain perfusion may not support memory functions in cognitively normal carriers of the ApoE ɛ4 allele compared to non-carriers. Front Aging Neurosci 8,151.

44.

Bondi

, Houston

, Eyler

, Brown

(2005) fMRI evidence of compensatory mechanisms in older adults at genetic risk for Alzheimer disease. Neurology 64,501–508.

45.

Han

, Houston

, Jak

, Eyler

, Nagel

, Fleisher

, Brown

, Corey-Bloom

, Salmon

, Thal

, Bondi

(2007) Verbal paired-associate learning by APOE genotype in non-demented older adults: fMRI evidence of a right hemispheric compensatory response. Neurobiol Aging 28,238–247.

46.

Dickerson

, Salat

, Bates

, Atiya

, Killiany

, Greve

, Dale

, Stern

, Blacker

, Albert

, Sperling

(2004) Medial temporal lobe function and structure in mild cognitive impairment. Ann Neurol 56,27–35.

47.

Sicard

, Duong

(2005) Effects of hypoxia, hyperoxia, and hypercapnia on baseline and stimulus-evoked BOLD, CBF, and CMRO2 in spontaneously breathing animals. NeuroImage 25,850–858.

48.

Zappe

, Uludağ

, Oeltermann

, Uğurbil

, Logothetis

(2008) The influence of moderate hypercapnia on neural activity in the anesthetized nonhuman primate. Cereb Cortex 18,2666–2673.

49.

Ances

, Leontiev

, Perthen

, Liang

, Lansing

, Buxton

(2008) Regional differences in the coupling of cerebral blood flow and oxygen metabolism changes in response to activation: Implications for BOLD-fMRI. Neuroimage 39,1510.

50.

Buxton

(2010) Interpreting oxygenation-based neuroimaging signals: The importance and the challenge of understanding brain oxygen metabolism. Front Neuroenergetics 2,8.

51.

Chiarelli

, Bulte

, Gallichan

, Piechnik

, Wise

, Jezzard

(2007) Flow-metabolism coupling in human visual, motor, and supplementary motor areas assessed by magnetic resonance imaging. Magn Reson Med 57,538–547.

52.

Lin

A-L

, Fox

, Yang

, Lu

, Tan

L-H

, Gao

J-H

(2008) Evaluation of MRI models in the measurement of CMRO2 and its relationship with CBF. Magn Reson Med 60,380–389.

53.

Aanerud

, Borghammer

, Chakravarty

, Vang

, Rodell

, Jónsdottir

, Møller

, Ashkanian

, Vafaee

, Iversen

, Johannsen

, Gjedde

(2012) Brain energy metabolism and blood flow differences in healthy aging. J Cereb Blood Flow Metab 32,1177–1187.

54.

De Vis

, Hendrikse

, Bhogal

, Adams

, Kappelle

, Petersen

(2015) Age-related changes in brain hemodynamics; A calibrated MRI study. Hum Brain Mapp 36,3973–3987.

55.

Ibaraki

, Shinohara

, Nakamura

, Miura

, Kinoshita

(2010) Interindividual variations of cerebral blood flow, oxygen delivery, and metabolism in relation to hemoglobin concentration measured by positron emission tomography in humans. J Cereb Blood Flow Metab 30,1296–1305.

56.

Peng

S-L

, Dumas

, Park

, Liu

, Filbey

, McAdams

, Pinkham

, Adinoff

, Zhang

, Lu

(2014) Age-related increase of resting metabolic rate in the human brain. Neuroimage 98,176–183.

APOE Genotype Modifies the Association of Fusiform Gyrus Cerebral Metabolic Rate of Oxygen Consumption and Object Naming Performance

Abstract

Background:

Objective:

Methods:

Results:

Conclusion:

Keywords

INTRODUCTION

MATERIALS AND METHODS

Participants

Apolipoprotein E genotyping

Cognitive testing

Functional task

Hypercapnic task

Calibrated fMRI acquisition

Data analysis

Statistical analyses

Data and code availability statements

RESULTS

Group differences in demographic, cognitive, and brain variables

APOE and FG %ΔCMRO2

Interaction of APOE and FG %ΔCMRO2 on object naming

Vascular risk, APOE, and FG %ΔCMRO2, %ΔCBF, and %ΔBOLD

DISCUSSION

Footnotes

ACKNOWLEDGMENTS

FUNDING

CONFLICT OF INTEREST

DATA AVAILABILITY

References

APOE and FG %ΔCMRO₂

Interaction of APOE and FG %ΔCMRO₂ on object naming

Vascular risk, APOE, and FG %ΔCMRO₂, %ΔCBF, and %ΔBOLD