Brainstem volume and arterial stiffness interact to support memory in midlife

Abstract

Historically regarded as a hub for autonomic control, the brainstem is now increasingly recognized for its role in cognition. However, little information is available regarding its specific contribution to cognitive function, particularly in midlife, when modifiable cardiovascular risks begin to shape brain health trajectories. Because the brainstem houses key nuclei that regulate cardiovascular dynamics and cerebral perfusion, we focused on carotid–femoral pulse wave velocity (cfPWV), a well-established index of central arterial stiffness that directly reflects cerebral hemodynamic load. This study investigated whether brainstem volume is associated with cognitive performance in healthy midlife adults and whether this relation is moderated by cfPWV. A total of 198 adults aged 40–65 years underwent neuropsychological testing, fasting cardiovascular and metabolic assessments, and structural MRI. Linear regression models examined brainstem volume as a predictor of attention, executive function, and memory, adjusting for age, sex, education, and total intracranial volume. Additional models tested the interaction between cfPWV and brainstem volume, controlling for cardiometabolic covariates. Greater brainstem volume was significantly associated with better memory performance (p = .010, FDR q = 0.030), but not attention or executive function. Greater brainstem volume was also associated with lower systolic and pulse pressure. Notably, cfPWV moderated the relationship between brainstem volume and memory, such that the positive association was strongest among individuals with higher arterial stiffness (p = .001, FDR q = 0.002). These findings suggest that the brainstem may serve as a structural hub linking peripheral vascular health and cognitive function in midlife, particularly under conditions of elevated arterial stiffness.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11682-026-01206-x.

Introduction

The brainstem is a structurally and functionally heterogeneous region classically known for its role as cardiovascular and respiratory control centers (Benarroch, 2008). More recently, it has been implicated in early Alzheimer’s disease (AD) pathology (Braak et al., 2011) and higher-order cognitive processes (Mather et al., 2016; Venkatraman et al., 2017). Despite this, most studies examining the brainstem’s role in cognition have focused on older adults or clinical populations. Far less is known about its relevance in midlife, a critical period when modifiable risk factors such as vascular health begin to shape long-term brain health trajectories.

Given the brainstem’s dual roles in managing peripheral physiology and central neural function, the brainstem may serve as an integrative hub for both identifying early risk and understanding mechanisms underlying cognitive decline. The present cross-sectional study investigates whether brainstem volume, derived from standard T1-weighted MRI, is associated with cognitive performance in healthy midlife adults and whether this relation is moderated by vascular health, indexed by central arterial stiffness.

Central arterial stiffness, typically measured by carotid–femoral pulse wave velocity (cfPWV), represents the reference-standard index of arterial stiffness and a key marker of vascular aging and cerebral hemodynamics (Townsend et al., 2015; Van Bortel et al., 2012). We have previously demonstrated that elevated cfPWV in midlife adults has been associated with lower subcortical perfusion and poorer cognitive performance (Strasser et al., 2015; Tarumi et al., 2011, 2013). These findings are consistent with large population-based evidence linking increased aortic stiffness to lower brain volume, greater white matter hyperintensity burden, and reduced memory performance (Mitchell et al., 2011; Tsao et al., 2013). Because cfPWV captures central rather than peripheral stiffness, it reflects the hemodynamic load most directly experienced by the brain and remains the most physiologically meaningful index for identifying early vascular contributions to cognitive risk in midlife.

Building on this rationale, we examined whether structural variability in the brainstem, a region integrating vascular and neural regulation, relates to cognitive function in midlife. Given the brainstem’s dense noradrenergic and autonomic projections to memory-critical structures (Mather et al., 2016; Sara, 2009), greater brainstem integrity may be particularly relevant to cognitive function under conditions of increased vascular burden, when demands on regulatory mechanisms may be greater. Accordingly, we hypothesized that larger brainstem volume would be associated with better episodic memory performance, and that this positive association would be stronger among individuals with higher arterial stiffness. Associations with attention and executive function were evaluated for comparison but were expected to be weaker, as these domains rely primarily on cortical networks less directly influenced by brainstem integrity (Miller & Cohen, 2001; Petersen & Posner, 2012).

Methods

Participants and procedure

The study enrolled 198 adults aged 40–65 years, recruited from Central Texas via flyers, bus and newspaper advertisements, and online postings. Exclusion criteria included a history of neurological disease, major psychiatric disorder, substance-related hospitalization, or MRI contraindications. Participants completed two visits: (1) metabolic and cardiovascular assessments, and (2) cognitive testing and MRI session. A detailed study procedure was previously described (Foret et al., 2021).

Cognitive testing

A comprehensive neuropsychological battery assessing attention, executive function, and memory was administered. Briefly, attention was assessed with the Trail Making Test A, Stroop Color Test, and Stroop Word Test. Executive function was assessed with the Trail Making Test B, Stroop Color-Word test, and Controlled Oral Word Association Test (COWAT). Episodic memory was assessed with the California Verbal Learning Test-II (CVLT-II) short-delay free recall, long-delay free recall, and recognition trials. Raw scores were converted to sample-based z-scores, and domain composites were derived by averaging the relevant subtest scores.

MRI Acquisition and processing

Brain MRI data were acquired on a 3.0 T Siemens Skyra scanner (Siemens Healthineers, Malvern, PA, USA), using a 32-channel head coil between 2011 and 2014. Structural T1-weighted MPRAGE images (256 × 256 matrix, flip angle = 7°, field of view = 24 × 24 cm², 1 mm slice thickness, 0 gap) were processed using the FreeSurfer v7.2.0 pipeline (Fischl, 2012), including motion correction, skull stripping, segmentation, and parcellation. Brainstem boundaries were delineated automatically by the FreeSurfer segmentation algorithm. Following automated segmentation, quality control was conducted by visually evaluating the brainstem segmentation for all participants. Total brainstem volume and total intracranial volume (TIV) were extracted for statistical analyses.

Metabolic and cardiovascular health measurement

Participants arrived after an overnight (≥ 8 h) fast. Height and body weight were measured to calculate body mass index (BMI). Venous blood samples were collected to assess concentrations of triglycerides, high-density lipoprotein (HDL) cholesterol, and glucose using standard enzymatic assays on the Cholestech LDX analyzer (Alere, Hayward, CA, USA). Blood pressure was measured on the right arm using an automated oscillometric device (Omron Healthcare, Kyoto, Japan), and systolic, diastolic, mean, and pulse pressure were derived. Participants also completed health history questionnaires to report medication use, smoking status, and habitual physical activity.

Arterial stiffness was assessed using carotid-femoral pulse wave velocity (cfPWV), obtained by recording pulse waveforms from the carotid and femoral arteries with the Omron VP-1000plus device (Omron Healthcare, Kyoto, Japan). cfPWV was calculated by dividing the distance between recording sites by the pulse transit time. Detailed acquisition and analysis procedures are described in detail (Pasha et al., 2018).

Statistical analyses

Analyses were conducted using R Statistical Software (v4.4.1; R Core Team, 2021) using base R functions and the mediation package (Tingley et al., 2014). Multiple linear regression models examined associations between brainstem volume (cm³) and each cognitive domain (attention, memory, executive function), adjusting for age, sex, education, and TIV. p(FDR) values were Benjamini–Hochberg–corrected within the hypothesis-driven family of brainstem effects across the three cognitive domains (attention, executive function, memory).

Additional models tested brainstem volume as a predictor of blood pressure, adjusting for BMI, glucose, HDL cholesterol, triglyceride concentrations, smoking status, physical activity, and use of blood pressure, cholesterol, and insulin medications. All medications were coded 0 = no, 1 = yes. Interaction terms evaluated whether cfPWV moderated the association between brainstem volume and cognitive outcomes, controlling for the same covariates. p(FDR) values were Benjamini–Hochberg–corrected within the hypothesis-driven family across the three interaction tests (one per outcome).

Finally, an exploratory mediation analysis was conducted to assess whether blood pressure measures mediated the association between brainstem volume and memory performance.

Results

Demographic and clinical characteristics are presented in Table 1, and cognitive domain z-scores and brain volumes are shown in Table 2. Raw cognitive test values and full regression parameters are provided in Supplemental Tables 1–3.

Table 1.

Participant demographic and clinical characteristics

Variable	Number (%) or Mean (SD)
	Overall (N = 198)
Age (in years)
Mean (SD)	49.5 (6.4)
Sex
Female	115 (58.1%)
Male	83 (41.9%)
Education (in years)
Mean (SD)	16.3 (2.7)
Race/Ethnicity
White	119 (60.1%)
Black or African American	19 (9.6%)
Hispanic or Latino	37 (18.7%)
Asian	8 (4.0%)
Other	13 (6.6%)
Missing	2 (1.0%)
Fasting glucose (mg/dL)
Mean (SD)	100 (31.5)
Triglycerides (mg/dL)
Mean (SD)	124 (86.7)
HDL-cholesterol (mg/dL)
Mean (SD)	51.8 (17.9)
Systolic blood pressure (mmHg)
Mean (SD)	121 (14.4)
Diastolic blood pressure (mmHg)
Mean (SD)	73.1 (9.73)
Mean arterial pressure (mmHg)
Mean (SD)	90.6 (12.2)
Pulse pressure (mmHg)
Mean (SD)	48.3 (8.66)
Carotid-femoral pulse wave velocity (cm/s)
Mean (SD)	997 (268)

HDL-cholesterol high-density lipoprotein-cholesterol

Table 2.

Cognitive and imaging measures (domain scores and regional volumes)

Variable	Mean (SD)
	Overall (N = 198)
Attention domain z score
Mean (SD)	0.03 (0.67)
Executive domain z score
Mean (SD)	0.04 (0.67)
Memory domain z score
Mean (SD)	0.00 (0.88)
Total intracranial volume (cm3)
Mean (SD)	1,470 (222)
Brainstem volume (cm3)
Mean (SD)	20.8 (2.66)

After adjusting for age, sex, education, and TIV, greater brainstem volume was associated with better memory performance (β = 0.07, p = .010, FDR q = 0.030). The full model accounted for 9% of the variance in memory performance (adjusted R² = 0.09, F(5, 192) = 4.93, p < .001). No significant associations were found between brainstem volume and attention or executive function (ps > 0.13, full models adjusted R² = 0.03 and 0.08, respectively). In models adjusting for age, sex, TIV, health behaviors (e.g. smoking, physical activity), and metabolic health covariates, higher brainstem volume was associated with lower systolic blood pressure (β = -1.378, p = .012, FDR q = 0.024) and lower pulse pressure (β = -0.097, p = .009, FDR q = 0.024), but not with diastolic (p = .090, FDR q = 0.121) or mean arterial pressure (p = .154, FDR q = 0.154).

In the fully adjusted model, a significant brainstem volume × cfPWV interaction emerged for memory performance (p = .001, FDR q = 0.002), indicating that the positive association between brainstem volume and memory was stronger among individuals with higher arterial stiffness. The full model accounted for 18% of the variance in memory performance (adjusted R² = 0.18, F(17, 115) = 2.75, p < .001). Adding the interaction term explained an additional ~ 3% of variance (ΔR² ≈ 0.03) beyond covariates, relative to the model without the interaction. Simple slope analyses showed that the brainstem–memory association was significant at high cfPWV (+ 1 SD; β = 0.10, p = .02), but not at mean (β = 0.02, p = .63) or low cfPWV (–1 SD; β = − 0.07, p = .21). The moderation is illustrated in Fig. 1, where cfPWV is grouped as high versus low based on the commonly accepted 10 m/s cutoff (The Reference Values for Arterial Stiffness’ Collaboration, 2010).

Fig. 1.

Interaction between brainstem volume and arterial stiffness (cfPWV) predicting memory performance in midlife adults. Greater brainstem volume was associated with higher memory domain z-scores among participants with higher arterial stiffness (red line), but not among those with lower arterial stiffness (blue line). Shaded areas indicate 95% confidence intervals

In exploratory mediation analyses, bootstrapped confidence intervals (5000 resamples) indicated the indirect effect of brainstem volume on memory performance through systolic blood pressure (ACME = − 0.003, 95% CI [–0.028, 0.020], p = .65) or other blood pressure measures was not statistically significant.

Discussion

This study examined the relation between arterial stiffness, brainstem volume, and cognition in a midlife sample. We demonstrated that greater brainstem volume was associated with better verbal episodic memory performance, particularly in the context of elevated arterial stiffness. These findings highlight the association between brainstem volume and memory performance varying by arterial stiffness level, suggesting that brainstem structure may play a particularly important role in memory among individuals with higher vascular burden.

The brainstem is increasingly recognized as a key structure in AD and related neurodegenerative processes. Specific brainstem subnuclei are among the first to exhibit tau-related changes, occurring as early as prepuberty (Braak et al., 2011). Brainstem and subnuclei volumes have been identified as sensitive biomarkers, not only in predicting AD progression in older adults (Dutt et al., 2020) but also in associating with amyloid accumulation in cognitively healthy middle-aged adults (Jacobs et al., 2022). As an early and persistent site of pathological protein deposition, the brainstem is susceptible to downstream axonal degeneration and neuronal loss, processes for which volumetric measures may serve as a proxy (Theofilas et al., 2017). Due to its early involvement in AD progression, sensitivity in midlife populations, and easy accessibility in conventional MRI studies, brainstem volume represents a practical and informative metric for evaluating early neurological risk. This study used brainstem volume as the independent variable, acknowledging that this is a coarse measure given the brainstem’s tissue heterogeneity and the availability of other more specific neuroimaging biomarkers.

Although the locus coeruleus–noradrenergic system housed within the brainstem broadly modulates both arousal and executive attention, its structural integrity has been most consistently linked to episodic memory performance in aging and midlife (Dahl et al., 2022; Mather et al., 2016). Structural indices such as brainstem volume likely reflect the capacity for sustained neuromodulatory output rather than transient state-level engagement, which may explain why brainstem–memory associations were stronger than those for attention or executive function in the present sample. Another possibility to explain the positive association between brainstem volume and memory performance is that the brainstem volume reflects preserved integrity of subcortical neuromodulatory nuclei. Several brainstem subnuclei have been implicated in memory processes (Da Cunha et al., 2003; Takeuchi et al., 2016). Although conventional T1-weighted MR structural images lack both the spatial resolution and tissue contrast needed to reliably visualize individual brainstem nuclei, total brainstem volume may nonetheless serve as a coarse proxy for their overall structural integrity in midlife.

Beyond its direct involvement in memory, the brainstem may also influence memory indirectly through its role in autonomic regulation. In our midlife sample, greater brainstem volume was associated with lower systolic and pulse pressure, consistent with findings from a recent UK-Biobank based lifespan study (Gurholt et al., 2025). This association is anatomically plausible, as the brainstem houses the nucleus tractus solitarius (NTS). This system receives afferent input from arterial baroreceptors and maintains mean arterial pressure (Benarroch, 2008). Because cerebral perfusion pressure is closely tied to mean arterial pressure (Lucas et al., 2010), effective baroreflex regulation may help preserve stable cerebral blood flow, particularly to the hippocampus, a region critical for memory and highly susceptible to hypoperfusion (Laosiripisan et al., 2015). However, our exploratory mediation analyses did not support blood pressure as a mediator of the brainstem-memory association detected in our study.

Lastly, we observed that greater brainstem volume was associated with better memory performance only among individuals with higher central arterial stiffness, whereas no such relationship emerged in those with lower stiffness. An interpretation of this finding is that brainstem structural integrity may become more consequential for cognition under conditions of greater vascular burden. The brainstem contains autonomic and noradrenergic systems involved in regulating blood pressure, cerebral perfusion, arousal, and memory-related processes. Thus, larger brainstem volume may broadly reflect greater integrity of regulatory systems that help maintain cognitive function when vascular regulation is compromised.

We propose two speculative and complementary mechanisms by which arterial stiffening may potentiate this link. First, central arterial stiffness may impair baroreflex sensitivity (Laosiripisan et al., 2015; Monahan et al., 2001), diminishing the ability of the vasculature to buffer fluctuations in blood pressure (Tarumi et al., 2015). In the context of impaired afferent signaling due to arterial stiffening, preserved integrity of medullary relay nuclei may help sustain reflex gain and stabilize blood pressure and subsequently cerebral perfusion. Animal and human lesion studies have consistently shown that structural damage to these brainstem baroreflex nuclei led to impaired baroreflex function, supporting the idea that preserved volume may offer compensatory value (Laosiripisan et al., 2015; Monahan et al., 2001; Riley et al., 2002). Second, central arterial stiffness may impair neurovascular coupling, the mechanism by which cerebral blood flow is directed locally to metabolically active regions. Central arterial stiffness increases transmission of pressure and flow pulsatility to the brain’s microvasculature, promoting endothelial injury and small-vessel dysfunction in high-flow, low-resistance regions such as hippocampus (Mitchell, 2018; Wahl & Clayton, 2024). This vascular milieu is expected to weaken neurovascular coupling as functional hyperemia depends on intact endothelial conduction along the arterial wall (Chen et al., 2014). In this context, greater brainstem volume may reflect preserved integrity of LC (noradrenergic)-related pathways that help maintain dynamic localized blood flow regulation despite vascular stress. Together, these two pathways may explain why brainstem integrity appears more consequential for memory performance under conditions of elevated arterial stiffness.

The study offers several notable strengths. It is among the first to examine the dual roles of brainstem in vascular function and cognition. The use of a comprehensive neuropsychological battery allows for robust domain-specific cognitive assessments, while vascular and metabolic health variables were carefully measured and controlled. Another strength is that the associations and interactions between brainstem volume, memory, and vascular health remained significant after accounting for key covariates known to influence cognitive outcomes (age, education, lifestyle factors, medication use). This suggests specific pathways linking vascular, brain, and cognitive health. However, several limitations must be considered. The cross-sectional design precludes causal inference regarding the directionality of relationships. Though we demonstrate the utility of brainstem volume, future studies need to employ imaging tools with anatomical or functional specificity to assess individual brainstem nuclei characteristics and explain the underlying drive for such associations. Last, the sample was highly educated, potentially limiting the generalizability of findings to the broader population.

Conclusion

In this cognitively unimpaired midlife sample, greater brainstem volume was associated with better memory performance, particularly among individuals with elevated arterial stiffness. These results support the view that brainstem integrity contributes to cognitive function and may be particularly important under conditions of compromised vascular regulation. By demonstrating these associations in midlife, before overt cognitive decline or extensive vascular damage, this study highlights the potential of brainstem volumetrics as an accessible MRI biomarker for identifying individuals at elevated risk for future memory decline. Longitudinal studies incorporating higher-resolution brainstem imaging and functional measures of autonomic regulation will be essential to clarify the mechanisms underlying these associations and to determine whether maintaining brainstem integrity contributes to resilience against cognitive decline in the context of vascular aging.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Author contributions included conception and study design (APH, HT, YL, TW), data processing (YL, TW, MC, CY, IG, JP), statistical analysis (YL, APH), interpretation of results (YL, TW APH, HT), drafting the manuscript or revising it critically for important intellectual content (YL, TW, APH, HT), and approval of the final version to be published and agreement to be accountable for the integrity and accuracy of all aspects of the work (all authors).

Funding

This work was supported by the National Institutes of Health [R01 NS75565 to A.P.H.].

Data availability

De-identified data and analytic code are available from the corresponding author upon reasonable request, in accordance with institutional and funding-agency policies.

Code Availability

All analytic code was written in R (version 4.4.1) and is available from the corresponding author upon reasonable request.

Declarations

Ethics approval

All study procedures were approved by the University of Texas at Austin Institutional Review Board (IRB ID # 2011070025) and conducted in accordance with the 1964 Helsinki Declaration and its later amendments.

Competing interests

The authors declare no competing interests.

Consent for publication

Participants consented to the publication of de-identified data in aggregated form.

Consent to participate

All participants provided written informed consent prior to study participation.