Dietary Fat Intake and Its Influence on the Association Between Liver Fat and Brain Choline Metabolism in Young Saudi Women

Summary

Objective:

Fat-rich nutrition may influence the neurochemical consequences of hepatic lipid accumulation. This study examined whether percent energy from fat modifies the association between liver fat and the brain (Cho/Cr) ratio in healthy young women.

Methods:

A total of 109 healthy Saudi women aged 18–25 years underwent brain magnetic resonance spectroscopy (¹H-MRS) to quantify Cho/Cr ratios. Liver fat was estimated using MRI-based proton density fat fraction (PDFF), and lipid intake was assessed using a validated food frequency questionnaire. Multivariable regression models tested the interaction between liver fat and fat-derived energy intake, followed by simple slopes analysis, regression diagnostics, and false discovery rate (FDR) correction.

Results:

A modest but statistically significant interaction was observed between liver fat and reported fat intake in relation to the brain Cho/Cr ratio (p = 0.037). The association between liver fat and Cho/Cr appeared stronger at higher levels of fat consumption (+1 SD) and weaker at lower intake (-1 SD). Regression diagnostics indicated no multicollinearity or heteroscedasticity, and FDR correction supported the stability of the interaction.

Conclusion:

Fat intake may modify the relationship between liver fat and brain choline metabolism in young women. These findings suggest that nutritional context could play a role in liver–brain metabolic coupling, although the effect size is modest and should be interpreted with caution.

Introduction

The liver and brain are metabolically intertwined through a complex network of hormonal, inflammatory, and lipid-signaling pathways. This bidirectional communication—often referred to as the liver–brain axis—has gained increasing attention in the context of metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as non-alcoholic fatty liver disease (NAFLD). The MASLD affects nearly one-third of the global population and is increasingly recognized as a systemic disorder with extrahepatic consequences [1,2]. Emerging evidence suggests that hepatic steatosis may influence central nervous system (CNS) function, including neuroinflammation, neurotransmitter balance, and cognitive performance [3,4]. Chronic systemic inflammation, oxidative stress, and altered lipid-signaling associated with MASLD are now considered potential contributors to impaired brain health.

High-fat dietary patterns play a significant role in influencing both hepatic and neural metabolism [5]. Diets high in fat contribute to hepatic lipid accumulation, disrupt circadian regulation of liver function, and alter brain reward circuitry [6,7]. In animal models, saturated fat intake impairs dopamine neurotransmission in the nucleus accumbens, which is crucial for motivation and feeding behavior [8]. Human studies similarly show that dietary fat restriction or excess can affect activity in brain reward regions, demonstrating the sensitivity of dopaminergic pathways to nutritional status [9,10,11]. Dietary fat also influences the production of gut-derived short-chain fatty acids (SCFAs), which mediate the gut–liver–brain axis and regulate both hepatic lipid metabolism and central neurotransmission [12]. Collectively, these findings suggest that nutritional status may contribute to liver fat accumulation and potentially influence its neurobiological correlates.

¹H-MRS provides a non-invasive tool to quantify brain metabolite ratios, including (Cho/Cr) and lipid-to-creatine (Lipid/Cr). Elevated Cho/Cr is considered a marker of membrane turnover and neuroinflammation, while lipid/Cr may reflect altered lipid metabolism or gliosis. Previous studies have linked these ratios to metabolic syndromes, insulin resistance, and early cognitive decline [13,14]. However, the extent to which dietary fat intake may modify the relationship between liver fat and brain metabolite ratios remains poorly understood.

Understanding this interaction is particularly relevant in young populations, where early metabolic changes may precede overt clinical symptoms [15,16]. The present study focused on women aged 18–25 years to minimize age-related metabolic variability and reduce potential confounding factors such as early insulin resistance, hormonal transitions, or chronic disease. In Saudi Arabia and other Middle Eastern countries, dietary patterns characterized by high saturated fat intake have contributed to rising rates of MASLD among young adults [17,18]. Identifying modifiable nutritional factors that influence liver–brain coupling could inform preventive strategies and personalized interventions [19].

This study therefore aimed to evaluate whether fat consumption modifies the association between liver fat and brain metabolite ratios in young Saudi women. We hypothesized that dietary fat intake may influence the relationship between liver fat and the brain Cho/Cr ratio.

Methods

This cross-sectional study included 109 healthy young Saudi women aged 18–25 years, recruited from Princess Nourah bint Abdulrahman University (PNU), Riyadh. This age range was selected to minimize age-related metabolic variation and reduce potential confounding factors such as early insulin resistance or hormonal changes. Eligibility criteria required participants to be free of diagnosed metabolic, hepatic, or neurological disorders and not taking medications known to affect lipid metabolism or brain function. Ethical approval was obtained from the Institutional Review Board at PNU (IRB Log Number: 23–0257), and written informed consent was obtained from all participants. Data collection was conducted between August 2023 and September 2024.

Brain magnetic resonance spectroscopy (¹H-MRS)

Brain metabolite ratios were quantified using single-voxel ¹H-MRS on a 3T MRI scanner (Siemens Magnetom Skyra, Germany). Voxels were positioned in the left parietal white matter. Acquisition parameters included an echo time (TE) of 135 ms, repetition time (TR) of 2000 ms, and 128 signal averages. Voxel placement for ¹H-MRS acquisition is illustrated in Fig. 1(b). Spectral data were processed using LCModel software (version 6.3), with metabolite concentrations expressed as ratios to creatine (Cr) [20]. The primary outcome measure was the (Cho/Cr) ratio. Spectra exceeding quality thresholds (Cramér–Rao lower bounds >15% or linewidth >0.07 ppm) were excluded prior to analysis.

Fig. 1.

(a) Axial abdominal MRI slice showing the location of the liver voxel used for ¹H-MRS acquisition. The region of interest (ROI) within the right hepatic lobe is indicated by a white box. (b) Axial brain MRI slice illustrating the voxel placement for ¹H-MRS in the left parietal white matter, marked by a white box. ROI: region of interest, ¹H-MRS: proton magnetic resonance spectroscopy.

Liver fat estimation (MRI-PDFF)

Liver fat percentage was estimated using abdominal magnetic resonance imaging with proton density fat fraction (PDFF) mapping, a validated non-invasive biomarker of hepatic steatosis [21]. Regions of interest (ROIs) were placed in the right hepatic lobe, avoiding major vessels and bile ducts. The anatomical localization of liver fat quantification is shown in Fig. 1(a). The PDFF values were averaged across 3 slices, and liver fat was treated as a continuous variable in all statistical models.

Dietary fat intake

Dietary fat consumption was assessed using a validated food frequency questionnaire (FFQ) developed for Saudi populations [17,22]. Frequency codes (0–6) were converted to servings per day using a standardized scale. Portion sizes and nutrient composition were derived from the Saudi Food Composition Tables. Total fat intake (g/day) was converted to percent energy from fat (%fat) using Atwater factors (9 kcal/g for fat). Percent energy from fat was used as the primary dietary variable in all analyses.

Statistical analysis

Descriptive statistics were reported as mean ± standard deviation (SD) or median [interquartile range, IQR] for continuous variables. Multivariable linear regression models were used to evaluate the association between liver fat and brain Cho/Cr ratio, with percent energy from fat included as a covariate. An interaction term (Liver Fat × %Fat) was added to test for effect modification.

Simple slopes analysis was conducted to probe the interaction at ±1 SD of fat intake levels. Variance inflation factors (VIF) were calculated to assess multicollinearity (threshold <5), and Breusch–Pagan tests [23] were applied to evaluate heteroscedasticity. False Discovery Rate (FDR) correction at q = 0.05 was used to control for multiple comparisons [24]. All analyses were performed using Python (version 3.10) with the statsmodels, scipy, and penguin packages. A 2-sided p-value <0.05 was considered statistically significant.

Results

A total of 109 healthy young adult participants were included in the analysis. Liver fat percentage ranged from 0.1% to 60%, with a mean of 13.2% (SD 11.6%). Percent energy derived from dietary fat ranged from 32.5% to 52.3%, with a mean of 42.1% ± 3.9%. The mean brain choline-toc-reatine (Cho/Cr) ratio was 0.71 ± 0.12. Spectral quality met predefined criteria across participants, and spectra exceeding Cramér–Rao lower bounds (CRLB) of 15% or linewidths greater than 0.07 ppm were excluded prior to analysis. A representative ¹H-MRS spectrum from the left parietal white matter is shown in Fig. 2, illustrating the major detectable metabolites including choline (Cho), creatine (Cr), N-acetylaspartate (NAA), myo-inositol (mI), and the glutamate–glutamine complex (Glx). Only the Cho/Cr ratio was included in the present analysis.

Fig. 2.

Representative ¹H-MRS spectrum acquired from the left parietal white matter. Peaks corresponding to choline (Cho, ~3.2 ppm), creatine (Cr, ~3.0 ppm), and lipid (~1.3 ppm) are annotated along the chemical-shift axis. Additional metabolites—including N-acetylaspartate (NAA), myo-inositol (mI), and the glutamate–glutamine complex (Glx)—are also visible in the spectrum but were not included in the present analysis. The x-axis represents the chemical shift in parts per million (ppm), and the y-axis reflects the relative metabolite signal intensity (unitless). ¹H-MRS: proton magnetic resonance spectroscopy, Cho: choline, Cr: creatine, NAA: N-acetylaspartate, mI: myo-inositol, Glx: glutamate–glutamine.

Multivariable linear regression revealed a modest but statistically significant interaction between liver fat and dietary fat intake in relation to the brain Cho/Cr ratio (β 0.0003, SE = 0.0001, p = 0.037), indicating that the association between liver fat and Cho/Cr varied according to fat-derived energy intake (Table 1). Simple slopes analysis showed that the slope of liver fat predicting Cho/Cr was steeper at higher levels of fat consumption (+1 SD; β 0.0064) and weaker at lower intake (-1 SD; β 0.0038), as shown in Table 2.

Table 1.

Multivariable regression model testing the interaction between liver fat and percent energy from fat (% energy from fat) in relation to the brain Cho/Cr ratio. The interaction term reached statistical significance (p = 0.037), indicating a modest dependence of the liver fat–Cho/Cr association on dietary fat intake.

Predictor	β (SE)	p-value
Liver fat (%)	0.0051 (0.0026)	0.053
% Energy from fat	0.0012 (0.0009)	0.187
Liver fat × % Energy from fat	0.0003 (0.0001)	0.037

SE: standard error.

Table 2.

Simple slopes analysis showing the association between liver fat and brain Cho/Cr ratio at low (-1 SD) and high (+1 SD) levels of percent energy from fat. The slope was numerically higher at greater fat intake, suggesting a modest variation in the liver fat–Cho/Cr relationship across dietary fat levels.

%Fat level (±1 SD)	Slope (β)
Low (-1 SD)	0.0038
High (+1 SD)	0.0064

SD: standard deviation.

Regression diagnostics indicated no evidence of multicollinearity, with variance inflation factors (VIF) for all predictors below 1.05. Breusch–Pagan tests for heteroscedasticity were non-significant (p > 0.18 for all predictors), supporting the assumption of homoscedasticity (Table 3). False Discovery Rate (FDR) correction confirmed that the interaction between liver fat and percent energy from fat remained statistically significant at q = 0.05 as shown in Table 4.

Table 3.

Regression diagnostics for multicollinearity and heteroscedasticity. Variance inflation factors (VIF) for all predictors were below 1.05, suggesting no notable multicollinearity. Breusch–Pagan test p-values were non-significant, consistent with the assumption of homoscedasticity.

Predictor	VIF	Breusch–Pagan p
Liver fat (%)	1.03	0.18
% Energy from fat	1.03	0.18
Interaction term	1.04	0.19

Table 4.

False discovery rate (FDR) correction applied to the interaction model testing liver fat × percent energy from fat (% energy from fat) in relation to the brain Cho/Cr ratio. The interaction term remained statistically significant after adjustment (FDR-adjusted p = 0.037), indicating that the finding was retained under multiple-comparison correction.

Test	Raw p	FDR-adjusted p	Significant
Interaction: Liver Fat × %Fat	0.037	0.037	Yes

Discussion

This study provides preliminary evidence that dietary fat intake may influence the association between liver fat and brain choline metabolism in young women. The observed interaction between liver fat and fat-derived energy intake in predicting the brain Cho/Cr ratio suggests a potential modulation of hepatic–neural coupling by nutritional status. Given the modest effect size and borderline statistical significance, these findings should be interpreted cautiously. Nonetheless, they point to the importance of considering dietary context when examining liver–brain metabolic interactions, particularly in populations at risk for early metabolic dysregulation.

The liver–brain axis is increasingly recognized as a bidirectional communication network involving hormonal, inflammatory, and metabolic signaling pathways. Fat intake levels have emerged as a potential modulator of this axis. In animal models, high-fat diets disrupt hepatic circadian rhythms and alter central reward signaling, leading to changes in feeding behavior and motivation as Darcey et al. [9] reported. Wallace and Fordahl [7] demonstrated that saturated fat intake impairs dopamine neurotransmission in the nucleus accumbens, linking peripheral lipid burden to central metabolic changes. While our findings do not establish causality, they are consistent with the broader literature suggesting that dietary lipids may influence the brain's metabolic response to peripheral lipid accumulation.

High-fat diets have also been associated with neuroinflammation, mitochondrial dysfunction, and altered choline metabolism in both preclinical and clinical studies. Srokowska et al. [25]. reported that prolonged exposure to high-fat diets increases brain choline levels and inflammatory markers, suggesting a mechanistic link between dietary lipids and neurochemical stress. Similarly, Cavaliere et al. [26] observed that high-fat feeding induces neuroinflammation and mitochondrial impairment in the cerebral cortex of mice In line with these observations, our data indicate that individuals with higher fat intake levels show a steeper liver fat–Cho/Cr slope, although the magnitude of this effect is modest. This pattern may reflect increased sensitivity of brain choline metabolism to peripheral lipid accumulation, but further research is needed to clarify the underlying mechanisms.

The gut microbiota also plays a critical role in mediating dietary effects on both liver and brain. Short-chain fatty acids (SCFAs), produced by bacterial fermentation of dietary fiber, influence hepatic lipid metabolism and central neurotransmission. Li et al. [12] showed that SCFAs modulate both hepatic steatosis and brain function through the gut–liver–brain axis. Although our study did not directly measure SCFAs or microbiome composition, the observed interaction may reflect downstream effects of altered microbial signaling in response to dietary fat. Future studies incorporating microbiome profiling and SCFA quantification could help clarify these pathways.

A major strength of this study is the use of a well-characterized cohort of young, healthy females free from diagnosed metabolic or neurological disorders. Most existing studies focus on older adults or clinical populations, where confounding factors such as insulin resistance, medication use, and comorbidities obscure early metabolic signals [14,15,27]. By focusing on a metabolically intact population, we were able to detect subtle interactions that may precede clinical symptoms. Moreover, ¹H-MRS provides a sensitive, non-invasive approach for detecting brain metabolic alterations. Htun et al. [27] validated the use of MRS biomarkers in young adults at risk for obesity, supporting the utility of this approach in early detection and intervention.

Study limitations

While the study benefits from a homogeneous and healthy cohort, several limitations should be acknowledged. First, the cross-sectional design limits causal inference, and the observed associations cannot determine temporal directionality. Second, dietary fat intake was assessed using a self-reported food frequency questionnaire, which is subject to recall bias and may not fully capture day-to-day variability in nutrient consumption. Third, the sample consisted exclusively of young Saudi women, which enhances internal validity but limits generalizability to males, older adults, or other populations. Fourth, only one brain region and a limited set of metabolites (primarily Cho/Cr) were examined; additional regions and markers such as NAA/Cr or Glx/Cr may provide a more comprehensive understanding of hepatic–neural interactions. Finally, potential mechanistic mediators—including inflammatory markers, hepatokines, gut microbiota composition, and short-chain fatty acids—were not measured, restricting interpretation of the biological pathways underlying the observed interaction. Future longitudinal and multimodal studies incorporating objective dietary biomarkers and broader neuroimaging metrics are needed to validate and extend these findings.

The possibility that dietary fat intake modulates the neurobiological impact of liver fat has potential implications for dietary counselling and public health. In regions like Saudi Arabia, where high-fat dietary patterns are prevalent among youth, early screening for hepatic and neural metabolic stress may be warranted. Interventions aimed at reducing saturated fat intake could help mitigate hepatic steatosis and its potential neurochemical correlates, although further evidence is needed to confirm these pathways. These findings also support the integration of neuroimaging biomarkers into nutritional epidemiology and metabolic health surveillance [16,17,18,27].

In conclusion, this study provides preliminary evidence that fat-rich diets may influence the association between liver fat and brain choline metabolism in young women. The observed interaction between liver fat and fat-derived energy intake in predicting the Cho/Cr ratio suggests a possible nutritional modulation of hepatic–neural coupling, although the effect size is modest and should be interpreted with caution. These findings support the broader concept of a diet-sensitive liver–brain axis and point to potential neurochemical correlates of high-fat dietary patterns.

By focusing on a metabolically intact cohort, we were able to detect subtle interactions that may precede clinical symptoms. The use of ¹H-MRS enabled non-invasive quantification of brain metabolite ratios, positioning Cho/Cr as a potential—though not definitive—indicator of systemic metabolic stress. In regions with rising rates of MASLD and high dietary fat consumption, these results underscore the value of considering nutritional factors when evaluating early metabolic risk and highlight the potential relevance of integrated hepatic and neural health assessments.

Future research should investigate the mechanistic pathways underlying this interaction, including the roles of hepatocytes, gut-derived metabolites, and inflammatory mediators. Longitudinal studies incorporating objective dietary biomarkers, multimodal neuroimaging, and cognitive assessments will be essential to clarify the temporal dynamics and clinical significance of liver–brain metabolic coupling in young populations.

AI use disclosure statement

Artificial intelligence tools were used only for language editing and grammar improvement during manuscript preparation. No AI tools were used for data collection, data analysis, image creation, scientific interpretation, or decision-making in the research process.

Disclosure

The authors have no conflict of interests, and the work was not supported or funded by any drug company.

Funding acknowledgment

The research was supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2026R852), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.

Ethical statements

Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Princess Nourah bint Abdulrahman University (IRB Log Number: 230257, March 22, 2023).

Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Consent for Publication: Written informed consent was obtained from participants for publication of anonymized data and accompanying images.

Data Availability Statement: Data are available upon reasonable request from the corresponding author. Data are not publicly available due to institutional privacy regulations and ethical restrictions.

Acknowledgment

This research is supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2026R852), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. The authors acknowledge Scribendi (www.scribendi.com) for the English language editing.

Disclosure

This manuscript has not been presented at any conference, scientific meeting, or symposium. No part of the work has appeared in conference proceedings or journal supplements.