Structural MRI analysis of age-related changes and sex differences in marmoset brain volume

The field of aging biology, which aims to extend healthy lifespans and prevent age-related diseases, has turned its focus to the Callithrix jacchus (common marmoset) to understand the aging process better. This study utilized magnetic resonance imaging (MRI) to non-invasively analyze the brains of 216 marmosets, investigating age-related changes in brain structure; the relationship between body weight and brain volume; and potential differences between males and females. The key findings revealed that, similar to humans, Callithrix jacchus experiences a reduction in total intracranial volume, cortex, subcortex, thalamus, and cingulate volumes as they age, highlighting site-dependent changes in brain tissue. Notably, the study also uncovered sex differences in cerebellar volume. These insights into the structural connectivity and volumetric changes in the marmoset brain throughout aging contribute to accumulating valuable knowledge in the field, promising to inform future aging research and interventions for enhancing healthspan.


Introduction
Biology of aging is an academic field that studies the aging processes of living organisms, and findings in this field promise to offer immense contributions to humanity.This includes extending healthy life expectancy by discerning the aging mechanisms, potentially allowing older individuals to enjoy longer and healthier lives.Aging corelates with various diseases such as cardiovascular diseases, cancer, and neurodegenerative diseases.However, research could elucidate the origins of these illnesses and methods for their prevention and treatment.Socioeconomically, enhancing healthy life expectancy and managing diseases could decrease social security expenses and sustain or even enhance the labor force.Magnetic resonance imaging (MRI) has emerged as an invaluable tool for aging and brain studies; its use allows studies to be performed while highly observing ethical measures in research.Specifically, MRI techniques enable measurements of total intracranial volume via anatomical measurements (Goldszal et al., 1998).In addition, the development of diffusion tensor imaging using diffusion weighted image (DWI), one of the MRI imaging methods (Peter and Basser, 1994), have made visualizing previously unseen neuronal white matter fibers possible (Mori and van Zijl, 2002).Directional anisotropy, as measured by diffusion tensor imaging, defines structural connections between functional cortical regions (He et al., 2007) (Hata and Nakae, 2023).A quantitative diffusion value can also be obtained, which is an indicator of the extent of water molecule diffusion by calculating the diffusion tensor (Pierpaoli and Basser, 1996) (Basser et al., 1994).Although MRI has played a pivotal role in human studies, like the exploration of memory changes with age (Grady et al., 2006) and brain network reorganization in aging (Sala-Llonch et al., 2015), human-based research presents challenges.These encompass individual differences influenced by factors such as age, genetics, diseases and lifestyle, making research subjects non-uniform.Additionally, the necessity for long-term follow-ups involves substantial time and costs.Lastly, the ethical concerns limit in vivo studies of human brain tissue.To address these challenges, researchers have turned to nonhuman primates (NHPs) as models of aging.NHPs, being phylogenetically closer to humans, mirror many of our brain functions and behaviors, making them valuable animal models.
The use of non-human primates in aging research provides valuable insights into understanding the aging process and diseases in humans.In particular, studies on chimpanzees, bonobos, and other large apes (Lowenstine et al., 2016) have revealed similarities and species-specific differences in disease comparisons with humans.Research on rhesus monkeys (Chen and Errangi, 2013) has shown a reduction in gray matter and fractional anisotropy (FA; an indicator of the integrity of fiber tracts) with aging, which aligns with the brain aging process in humans.Velvet monkeys (Frye et al., 2021) have been shown to serve as a natural model of Alzheimer's disease, demonstrating the neuropathological features characterizing early Alzheimer's disease.The gray mouse lemur (Chaudron et al., 2021) has been evaluated for changes in psychomotor and cognitive functions between healthy aging and pathological aging, with approximately 10 % naturally showing pathological aging.This included observed brain atrophy and cognitive deficits similar to these involved in the aging process in humans.These studies demonstrate the crucial role of aging research using non-human primates in elucidating the mechanisms of aging and disease in humans.Recent studies have highlighted the importance of environmental control in the research process, including studies involving Non-Human Primates (NHPs) such as Callithrix jacchus, as a crucial aspect of enhancing the reproducibility and accuracy of outcomes (Suzette and Tardif, 2011).The ability to regulate environmental variables in NHP research allows for a focused examination of their impacts on the incidence of age-related diseases.Furthermore, adjusting for environmental variables could provide insights into how such modifications affect aging and disease management in humans.Furthermore, due to their physiological and anatomical similarities to humans, Callithrix jacchus has garnered attention as an alternative model to primates in biomedical research (Uematsu et al., 2017) (Hata and Nakae, 2023).Although they exhibit differences from old world primates and humans in aspects such as reproduction, endocrine signaling, and immunology, the undeniable display of characteristic primate traits (Abbott and Barnett, 2003) cannot be overlooked.These unique characteristics have led to an expanded use of Callithrix jacchus in biomedical research.
Moreover, Callithrix jacchus exhibits a decline in cognitive function around the age of 8 years (Rothwell et al., 2021) (Suzette andKeith, 2011) (Abbott et al., 2003).These studies have demonstrated changes consistent with these occurring during the human aging process, such as a reduction in gray matter and alterations in white matter integrity.Marmosets, in particular, have been reported to show brain aging-related changes by the age of 8, including a decrease in neurogenesis and β-amyloid deposition in the cerebral cortex, often referred to as 'senescence.'Research comparing young (2-3 years) and older (7-8 years) marmosets has found signs of aging in the latter group, such as a transition from fibrous to fibrocartilaginous in the articular disc (Berkovitz and Pacy 2000), β-amyloid deposition in the cerebral cortex (Suzette and Keith, 2011), and deposition of Alzheimer-like plaques in the cerebral cortex.These findings highlight the marmoset as a valuable model for studying aging and related diseases, offering insights particularly into the pathology of Alzheimer's disease among other age-related changes.
Despite the advantages of using Callithrix jacchus, detailed studies on the changes in their brain structure with aging remain unclear.We utilized small animal MRI scanners to secure longitudinal MRI data of Callithrix jacchus from infancy through adulthood to old age.In a study targeting juvenile subjects, Seki F et al. evaluated marmosets aged up to 30 months post-birth and discovered that the characteristic development of white matter involves an initial volume increase followed by stabilization.Investigations combining diffusion MRI and histological validation of MRI offer the potential to provide foundational data for further analysis of the myelination processes.Therefore, we focused on examining how the total intracranial volume and neural fibers change with aging from infancy.

Animals
In this study, the total intracranial volumes of healthy Callithrix jacchus, aged between 0.9 and 10.3 years (mean: 4.34 ± 2.56 years), were assessed through imaging in 216 subjects.Diffusion-weighted imaging (DWI) was successfully performed in 125 out of these 216 subjects.The variability in acquiring DWI data was partly attributed to the physical condition of the Callithrix jacchus and presence of image artifacts.(Table 1).Each marmoset was anesthetized, and its head immobilized before imaging.In vivo MRI scans were carried out with the animals in a supine position on an imaging stretcher, under anesthesia with isoflurane (1.5-2.5 %; Abbott Laboratories, Abbott Park, IL, USA) mixed with oxygen and air.Throughout the imaging procedure, vital parameters such as heart rate, oxygen saturation, and rectal temperature were consistently monitored.This study was approved by the Animal Experiment Committees of RIKEN Center for Brain Science (CBS) and conducted in accordance with the Guidelines for Conducting Animal Experiments of RIKEN CBS and National Institute of Health Guide for the Care and Use of Laboratory Animals.Every effort was also made to minimize the number of animals used and their suffering.

Ageing
The need for a clear definition and clarification of "aging" and the corresponding ages of marmosets and humans in this study was based on the classification of common marmosets by Adriano de Castro Leão et al. (2009) and the book "The Life Cycle Completed" by Erik Erikson (Hirose and Kobayashi, 2020) the stages of human development.According to de Castro Leão et al. (2009), the life stages of common marmosets are categorized as follows: infancy (0-3 months), juvenile (3-12 months), subadult (13-18 months, reaching sexual maturity but not full adult size, analogous to human adolescence ranging between 10 and 19 years), and adult (over 18 months).Marmosets are often considered 'aged' at approximately 8 years old (Abbott et al., 2003).In comparison, Erikson's human developmental stages (Erik H. Erikson.,1982) are defined as: infancy (0-17 months), early childhood (18 months-3 years), play age (3-5 years), school age (5-13 years), adolescence (13-20 years), young adulthood (20-40 years), adulthood (40-65 years), and old age (65 years and above).By aligning the developmental stages of common marmosets with those of humans, our study aimed to provide a clearer understanding of 'aging' within the context of both species.This alignment allows us to explicitly define 'aging' in our study and state corresponding age ranges in common marmosets, facilitating a more precise comparison with human aging processes.Drawing on the work of Adriano de Castro Leão et al. (2009), which documented age-related changes in young (2-3 years old) and older (7-8 years old) marmosets, our study aimed to further investigate the alterations in intracranial volume due to aging in Callithrix jacchus.We examined the relationship between intracranial volume and body weight, differences in intracranial volume based on sex, and variations in brain volume across different life stages-from 0 to 1.9 years, mirroring human infancy to old age, and from 2 to 7 years, reflecting human adolescence to adulthood.This exploration assessed the impact of sex on brain volume in these primates, thereby expanding upon the foundational insights provided by previous research to offer a more comprehensive understanding of age-related changes in brain volume.

MRI
MRI imaging of the 216 Callithrix jacchus was achieved by a 9.4-T BioSpec 94/30 unit (Bruker Optik GmbH, Ettlingen, Germany) equipped with an 86-mm inner diameter transmit and receive coil.

Brain region
In this study, the term 'brain region' refers to specific areas within the brain that have been identified and categorized based on their anatomical characteristics.We utilized the anatomically segmented atlas of the common marmoset brain, as detailed by Hashikawa et al. (Hashikawa and Nakatomi, 2015), following the methodology outlined by Tournier et al., (2004).This atlas provides a comprehensive anatomical parcellation of the marmoset brain into distinct regions, facilitating precise analysis and comparison.For our analysis, the atlas was applied to MRI data, specifically standardized T2-weighted (T2w) images.This standardization process ensures that the atlas can be accurately aligned with individual brain datasets, regardless of variations in size or orientation among subjects.The initial application of the atlas identified 111 distinct regions in one brain specimen.In a second specimen, regions were consolidated, resulting in a total of 52 distinct areas.This consolidation was performed by combining adjacent or functionally similar regions from the initial set of 111, based on established anatomical criteria.To further categorize the data, we divided the brain into six major anatomical sections: the cerebrospinal fluid, cortex, subcortex, white matter, cerebellum, and brainstem, thereby allowing for a more structured analysis of brain composition and function.In this study, total brain volume was defined as the sum of these sections: cortex, subcortex, white matter, cerebellum, and brainstem.Alignment of the brain atlas with individual MRI datasets was achieved by the ANTs SyN software.This tool facilitates the precise overlay of the atlas onto the T2w images of each subject's brain, ensuring that the anatomical parcellation is consistent across all datasets (Fig. 1A).

T2WI
To correct the T2w images, whole brains were extracted from the  image data using BrainSuite18a (David W. Shattuck, Ahmanson-Lovelace Brain Mapping Center at the University of California).Subsequently, mask images were created, and a registration process was performed to align the standard brain images by mapping region data to each animal's structural images.The ANTs software (Brian B. Avants, University of Pennsylvania) was used for this process.

Statistical analysis
In this study, our primary objectives were to explore the relationships among key continuous variables-brain volume, age, and body weight of the participants-and examine how brain volume could differ between sexes.Based on the scatter plot, we derived the R 2 value, F value, P value, and the regression equation.We used the JMP Pro 16.0.0software for statistical analysis.A General Linear Model (GLM) incorporating covariates was employed to investigate the association between brain volume (measured in cubic millimeters) and various factors, including age, weight, and sex.We assessed the model's goodness-of-fit through Pearson Chi-square and Deviance Chi-square statistics, along with the value of overdispersion, to ascertain the adequacy of the model fit to the data.This advanced statistical approach allowed for a comprehensive analysis that accounted for the effects of multiple covariates.

Fig. 2. Total brain volume and quantitative diffusion
Value.Correlations between total brain volume (cortex, white matter, subcortex, brainstem, and cerebellum) and age, and between diffusion quantiles and age are explored.The total intracranial volume demonstrates a non-normal distribution.In accordance with a previous study by Fjell, A. M., et al. (2010), a nonlinear triple exponential curve is employed to model this relationship.The diffusion indices, including fractional anisotropy (FA), axial diffusivity (AD), and radial diffusivity (RD), are modeled as non-linear biexponential curves due to the non-normal distribution of the data, based on the F value, coefficient of determination (R 2 ), and observed data distribution.The solid lines of brain volumes represent fits to the third order of the data; the diffusion quantiles represent fits to the second order of the data.Dotted lines delimit the associated confidence intervals and visually summarize the fits and the variability around these estimates.Analysis utilizing a generalized linear model (GLM) reveals significant age-related differences in total brain volume (n=216) and RD (n=125).The estimated change in total brain volume is − 0.0005 with a standard error of 0.0002, yielding a p-value of 0.0301.Regarding RD, the estimated change is − 0.0132 with a standard error of 0.0031, resulting in a p-value of <.0001.(Detailed statistical results can be found in Supplementary Table 1.).

Datasets
All datasets are publicly available in the Brain/MINDS Data Portal (https://dataportal.brainminds.jp/marmoset-mri-na216).These datasets are segmented into four sections: in vivo MRI in 216 animals, ex vivo MRI in 91 animals, standard brains, and BMA 2019 Atlas mapping.

Results
From T2W images of the brains of 216 Callithrix jacchus (91 males and 125 females; Table 1), we obtained total intracranial volumes (mm 3 ) that were derived for six regions: the cerebrospinal fluid, cortex, white matter, subcortex, brainstem, and cerebellum.Furthermore, five additional regions, namely the hippocampal formation, thalamus, retrosplenial cortex, putamen, and cingulate, were considered to detail total intracranial volume, diffusion quantification, and images.This was performed to examine the deeper regions of the brain more meticulously and draw comparisons with previous studies on other NHPs.All diffusion quantification values were calculated as the median of each region and were examined in relation to age and weight.

Correlation between brain volume and age with age, weight, and gender
The investigation into the relationship between brain volume (measured in cubic millimeters) and factors such as age, weight, and sex utilized a GLM incorporating covariates, with results shown in Figs. 2, 3, and Supplementary Figures 2-4, as well as Supplementary Tables 1 to 9. Following the methodology by Fjell and Walhovd (2010), brain volumes were modeled using a nonlinear triple exponential curve, evidenced by solid lines for third-order data fits and dotted lines marking confidence intervals, while diffusion indices (FA, [AD, and RD) followed nonlinear double exponential curves for second-order data fits.A goodness-of-fit test confirmed the adequacy of the GLM, which incorporated age, sex, and weight as factors, for both brain volume and diffusion quantiles (FA, [AD, and RD).

Total brain volume
Regarding total brain volume, data showed an initial decrease with age, followed by an increase, and then another decrease, indicating a complex, non-linear pattern.Weight also displayed a non-linear relationship with brain volume: it initially caused a decrease, then an increase at a certain threshold, and decreased again with further weight gain.The GLM confirmed a positive correlation between weight and brain volume (p=0.0022) and a negative correlation with age (p=0.0301),whereas sex did not significantly affect brain volume (p=0.7712).

Cerebrospinal fluid
Regarding cerebrospinal fluid volume, a similar complex, non-linear relationship with age was found, with initial decrease, subsequent increase, and another decrease in volume.Weight affected cerebrospinal fluid volume in a non-linear pattern as well, increasing with weight up to a point before decreasing and then increasing again at very high Fig. 3. Correlations between intracranial volume and age in each region.In accordance with a previous study by Fjell, A. M., et al. (2010), a nonlinear triple exponential curve is employed.The horizontal axis represents age, while the vertical axis represents volume.The solid lines of brain volumes represent fits to the third order of the data.Dotted lines delimit the associated confidence intervals and visually summarize the fits and the variability around these estimates.An analysis performed using a generalized linear model (GLM) reveals significant differences related to volume and age in the cortex, subcortex, white matter, and brainstem (n=216).(Detailed statistical results can be found in Supplementary Table 1.).
weights.However, the model's low R 2 value (0.007) suggests it captures little variability, implying significant influence from other unmeasured factors.GLMs showed that age, weight, and sex did not significantly impact cerebrospinal fluid volume (p-values of 0.13519, 0.39141, and 0.33608, respectively).

Cortex
Age showed a non-linear relationship with cortical volume, decreasing, then increasing, and decreasing again.Weight similarly affected cortical volume, initially decreasing then increasing, and decreasing again.The model indicated a negative association with age and a positive association with weight, while sex showed no effect.

White matter
it exhibited a non-linear pattern with age and weight, whereas volume decreased with age, then increased, and decreased again.Weight increases led to an initial increase in volume, a decrease, and then an increase again.Age and weight were positive predictors.However, sex showed no significance.

Subcortex
It demonstrated a non-linear relationship with both age and weight, whereas volume decreased with age, increased, and decreased again.An increase in body weight initially led to a decrease in volume, then an increase, and a decrease again.

Brainstem
It indicated a non-linear relationship with age and weight, with volume decreasing with age, then increasing, and decreasing again.As body weight increased, volume initially increased, decreased, and increased again.

Subcortex and brainstem
The model revealed that weight positively impacted brain volume, whereas age negatively affected it.Sex did not significantly predict volume changes.

Cerebellum
The R 2 value of 0.032 suggested low explanatory power, indicating that other factors significantly contributed to variations in cerebellar volume.The relationship with body weight showed an initial decrease in volume, an increase, and then a decrease again (Fig. 3).The model indicated a positive impact of body weight and negative impact of age on volume.Females had a slightly smaller cerebellar volume than did males (p = 0.0303).

Hippocampal formation
The value of R 2 is 0.005, explaining only approximately 0.5 % of the total variation, and many other unknown factors may have affected brain volume (Fig. 3).The relationship with body weight showed an initial increase in volume, a decrease at a certain point, and then an increase again.The model indicated a positive impact of body weight on volume, but age and sex did not have a significant effect.

Thalamus
It exhibited a non-linear relationship with age, with volume decreasing in early ages, increasing in middle ages, and decreasing again in older ages, suggesting age-specific sensitivity for brain volume.The relationship with body weight is also complex; volume decreased with initial weight gain, increased with further weight gain, and decreased again with significant weight gain.

Retro splenial cortex
It showed a similar pattern to the thalamus in terms of age and weight.Brain volume decreased with initial age increase, rose to a certain point, and then fell again in older age.Similarly, brain volume decreased with initial weight gain, increased at a specific threshold, and decreased again with further weight gain.
Thalamus, Retro splenial cortex: GLM showed that weight and age had a significant effect, whereas sex did not.

Cingulate
It demonstrated that age significantly affected brain volume variability, with an R^2 value of 0.290 indicating a strong model fit for age as a predictor.Brain volume showed a complex relationship with body weight, decreasing with initial weight gain, increasing within a certain weight range, and decreasing again with further weight increase.The model highlighted a strong negative impact of age on brain volume, whereas body weight and sex showed no significant effects.

Putamen
It revealed a complex pattern with both age and body weight.Brain volume decreased with initial aging, increased, and then decreased again with further aging.The relationship with body weight started with an increase in volume with weight gain, slowed down, and then increased again with further weight gain.The model indicated a positive effect of body weight on brain volume, whereas age and sex showed no significant impacts.4.1.14.Total brain volume from 2 to 6.9 years (Supplementary Figure 2) The factors influencing whole brain volume from the ages of 2-6.9 years were analyzed through a GLM.The model exhibited good fit, as indicated by the Pearson statistic, deviance statistic, and AICc value, all suggesting favorable model fit.Overdispersion was not a concern.Regarding total brain volume, the relationship with age showed an increase in brain volume with age, followed by a decrease, and then an increase again as age further advanced.This suggested that the effect of age on brain volume could vary within certain age ranges.Based on the GLM, body weight emerged as the strongest factor providing a statistically significant positive impact on total brain volume (p<0.0001).Conversely, age and sex showed no statistically significant effect on total brain volume, as illustrated in Supplementary Figure 2.

Brain volumes of ages 0-1.9 and 8+ (Supplementary Figure 1, Supplementary 7)
Age had a statistically significant negative impact on the total volume of the brain as well as on specific labeled sections, namely the cortex and subcortex.This indicated a tendency of brain volume to decrease as age increases.Body weight potentially exerted a small positive effect on the overall volume of the brain, with slight significance in the cortex; a significant positive effect on the cerebellum has also been confirmed.However, the impact of body weight on the cerebrospinal fluid, white matter, and brainstem was not found to be statistically significant.Sex did not have a statistically significant effect on the total brain volume or any other regions.

Age, weight, and sex results for diffusion quantitative values of brain volume
Diffusion metrics (FA, AD, RD) followed a nonlinear double exponential curve of the second data fit.Low R 2 values (for example, below 0.05) suggested a high likelihood that there was no strong relationship between age and these measurements.Scatter plots with high R 2 values (0.2 or above) may indicate that age or weight could be significant predictors for the measurements.However, they still suggest that more than half of the overall variability could be due to other factors.
Diffusion metrics, including FA), AD, and RD, were analyzed using GLMs.The model's overall fit was assessed to ascertain its appropriateness, considering age, sex, and body weight as covariates.Regarding AD, across all examined regions except for the brainstem, cerebellum, and hippocampal formation, the goodness-of-fit statistics-Pearson's Chi-squared and Deviance Chi-squared-demonstrated an excellent fit to the data (Fig. 4, Supplementary Fig ure 5, Supplementary Table 1~5).
For total brain volume, the FA value indicated that sex was a significant predictive factor, suggesting it had a positive effect.However, the effects of age and weight were not statistically significant.The AD value had a very small p-value for age, indicating it had the strongest effect.Age demonstrated a decrease with each increase in age.Regarding sex and weight, neither showed a statistically significant impact.The RD value indicated a negative effect, showing a decrease with each increase in age.Regarding sex and weight, both were not found to have a statistically significant effect.Considering the cerebrospinal fluid, FA values were not influenced by age, weight, or sex.Conversely, AD values were impacted by age and weight but remained unaffected by sex.Similarly, RD values were influenced by age and weight, with sex showing no effect.Regarding the cortex, FA values were not modulated by any of the studied variables.AD values exhibited a negative correlation with age, whereas weight and sex had no impact.Likewise, RD values were negatively correlated with age, without modulation by weight or sex.
In regard to the white matter, the FA values demonstrated sex-based differences, with female subjects exhibiting lower values compared with their male counterparts.AD values were negatively correlated with age but unaffected by weight or sex.RD values followed a similar pattern, being negatively influenced by age without any impact from weight or sex.With regard to the subcortex, the FA values were modulated by age and sex but not by weight.The AD values revealed a negative correlation with age and a positive correlation with weight, without effect from sex.The RD values were negatively influenced by age, with a potential positive modulation by weight without impact from sex.
Considering the brainstem, the FA, AD, and RD values all exhibited a slight negative correlation with age, without effects from weight or sex.Similarly, in the cerebellum and hippocampal formation, the FA, AD, and RD values all demonstrated a negative correlation with age, without being influenced by weight or sex.Regarding the thalamus, the FA values increased with age, whereas AD and RD values decreased.Weight slightly positively correlated with AD values, without observed impact from sex.In the retro splenial cortex, age positively influenced FA values, whereas female subjects showed decreased FA values, and weight positively influenced these values.The AD values were negatively associated with age and positively with weight, without any sex impact.The RD values were negatively influenced by age, without modulation by sex or weight.Considering the cingulate, the influence of age, sex, or weight on FA values was not statistically significant.The AD values were negatively associated with age and showed a slight, nonsignificant positive correlation with weight, without influence from sex.The RD values were negatively influenced by age and positively by weight, without sex effect.In regard to the putamen, age positively influenced the FA values, without impact from sex or weight.AD exhibited a negative association with age and a positive one with weight, without sex effect.RD was negatively influenced by age, without significant impact from weight or sex.

Discussion
Investigations into the effects of aging on brain volume in Callithrix jacchus have revealed areas that aligned with and those that differed from the findings in humans and other non-human primates (NHPs).The observed reduction in total brain volume with aging in Callithrix jacchus, in some aspects, corresponds with research reports on humans (Westlye LT et al., 2010) and other primates (Chen X and Errangi B, 2013), suggesting universal aspects of aging that transcend species.

Total brain volume
This study aimed to elucidate the impact of aging on brain volume in Callithrix jacchus in comparison with the aging processes in humans and NHPs.In humans, brain volume decreases by about 11 % by the late 80 s (Dekaban, 1978), and a similar, albeit more gradual, decrease is observed in chimpanzees and rhesus monkeys (Chen and Errangi, 2013).Callithrix jacchus, as revealed in this study, demonstrated a pattern of initial decrease in brain volume, followed by an increase at a specific point, and then a subsequent decrease.This pattern reflects the widespread trend observed in humans and other NHPs, wherein brain volume typically peaks during early development before beginning to decline.Seki et al. (2017) documented that the brain volume in Callithrix jacchus reaches its peak at 6 months after birth and then gradually decreases.It was found that there is a significant decrease between the ages of 0 and 1.9 years.Given that Callithrix jacchus reaches puberty around 9 months and achieves sexual maturity by 2 years of age, as reported by de Castro Leao et al. (2009) and Schultz-Darken et al. (2016), we defined adulthood as the period from 2 to 6 years old.However, between the ages of 2-6.9 years and over the age of 8, age did not influence brain volume.This result, which differs from humans, emphasizes species-specific patterns of brain development and aging.Body weight was found to have a positive correlation with brain volume increase, a result that is consistent with other non-human primates.

Other brain regions
According to human studies, it has been observed that the volume of various brain regions decreases with age.These regions include the cerebrospinal fluid (Courchesne et al., 2000;Faria et al., 2010;Hasan Fig. 4. Correlation of diffusion quantiles with age across various sites.The correlation between diffusion quantification values and age at each site is shown.The analysis of diffusion quantiles FA, AD, and RD (N=125) utilizing a generalized linear model reveals significant age-related differences across various brain regions.Regarding FA, significant differences are observed in the subcortex, thalamus, retrosplenial cortex, and putamen.Considering AD, notable differences are found in the cerebrospinal fluid, cortex, white matter, subcortex, brainstem, thalamus, retrosplenial cortex, cingulate, and putamen.Similarly, on RD, significant variations are detected in the cerebrospinal fluid, cortex, white matter, subcortex, brainstem, thalamus, retrosplenial cortex, cingulate, and putamen, indicating significant differences across these regions.(Refer to Supplementary Table 1 for detailed statistical results).The horizontal axis represents age, while the vertical axis corresponds to diffusion quantiles.Solid lines represent fits to the second order of the data.Dotted lines delineate the associated confidence intervals, visually summarizing the fits and the variability around these estimates.et al., 2009), cortex (Allen et al.), white matter (Bush and Allman, 2004;Westlye et al., 2010), subcortex (Chen and Errangi, 2013), cerebellum (Fjell and Walhovd, 2010;Raz and Lindenberger, 2005), hippocampal formation (Erickson et al., 2012;Terry. Jernigan et al., 2001;Pruessner et al., 2001), thalamus (Raz et al., 2005;Fama and Sullivan, 2015), and putamen (Abedelahi et al., 2013).
In Callithrix jacchus, it has been found that the cortex, white matter, subcortex, thalamus, and cingulate showed a decrease in volume with age.Specifically for the cingulate, it has been reported that the volume peaks at 6.4 months and then decreases by 10 % at 18 months (Uematsu et al., 2017).Similarly, the brainstem (Fjell and Walhovd, 2010) does not change with age in Callithrix jacchus, paralleling human findings.
Differing from humans, the cerebellum, hippocampal formation, and putamen do not decrease in volume with age in Callithrix jacchus.In rhesus monkeys, the hippocampal formation also does not show a decrease in volume with age (Shamy et al., 2006), and there appears to be no significant loss of neurons with aging (Peters, 2002;Amaral, 1993).However, the results indicate a very low R 2 value for the hippocampal formation (R 2 =0.005), suggesting that most variations in these regions are either unexplained or significantly influenced by other unobserved factors.
Furthermore, although there are studies observing a decrease in the volume of the putamen in rhesus monkeys (Alexander et al., 2008), reports also exist stating that the volume decrease is not significant (Wisco et al., 2008), suggesting differences among species.Lastly, in the retrosplenial cortex, significant functional changes due to aging have been confirmed in humans (Sambataro, F., Murty, V. P., 2010).It cannot be conclusively stated that age has a statistically significant impact on brain volume in Callithrix jacchus.Furthermore, concerning cerebrospinal fluid, it has been demonstrated that human cerebrospinal fluid increases with age (Courchesne, E., Chisum, H.J., 2000) (Faria et al., 2010;Hasan et al., 2009).No age-related changes were indicated in Callithrix jacchus, a finding similarly reported in rhesus monkeys (Wisco and Killiany, 2008).However, the current results highlighted a very low R 2 value (R 2 =0.007) for cerebrospinal fluid, suggesting that most of the variability in these areas could be either unexplained or significantly influenced by factors other than age, weight, and sex.Future research, increasing the number of individuals over the age of 8, is deemed necessary.
Regarding diffusion metrics, Chen X et al. (2013) reported that the average FA in rhesus monkeys decreases with aging.In chimpanzees and humans, an inverted U-shaped curve adequately represents the relationship between FA and age.The average RD in rhesus monkeys increases with age, while in chimpanzees and humans, a U-shaped curve more accurately depicts the relationship between RD and age.In Callithrix jacchus, the illustrated pattern was nonlinear and biexponential, without U-shaped or inverted U-shaped curves, resembling a near-linear mirroring (Fig. 4).Additionally, the study demonstrated that the effects of age, weight, and sex on FA, AD, and RD values vary across different brain regions, particularly revealing that age exerts a negative impact on the AD and RD values in many brain areas.This is likely due to the complex tissue structure of the brain and the accumulation of abnormal proteins, which may inhibit the movement of water molecules along neural fiber directions (Phillips et al., 2019) (Bennett et al., 2010) (Abbott and Barnett, 2003).

Body weight
Analysis of the total brain volume, cortex, white matter, subcortex, brainstem, cerebellum, hippocampal formation, thalamus, and putamen has shown that an increase in body weight positively affected brain volume increase.There are reports that the body weight of Callithrix jacchus doubles from 3 to 12 months and then stabilizes from 12 to 24 months (Sawiak et al., 2018).However, in this study, it was revealed that weight was the strongest factor positively influencing total brain volume between the ages of 2-6.9 years (Supplementary Figure 2), with statistical significance (p<0.0001).Nonetheless, considering these subjects are kept as experimental animals, differences from Callithrix jacchus in the wild may be expected.It was found that the influence of body weight on the diffusion metrics in specific brain regions is limited.Specifically, the AD values in the subcortex and thalamus, the FA value in the retrosplenial cortex, the RD value in the cingulate, and the AD value in the putamen are statistically significantly affected by body weight, but the impact of weight in many other areas was not statistically significant.Chen et al. (2013) reported that changes in RD are sensitive to myelin changes (Song et al., 2002) (Tyszka et al., 2006).Furthermore, changes in AD may be related to shifts in fiber density or axonal diameter (Sheng-Kwei and Sun, 2003) (Tyszka et al., 2006).Thus, with the increase in body weight, potential changes in tissue structure in the subcortex, thalamus, retrosplenial cortex, cingulate, and putamen were suggested.Although pathological results for each individual have not been obtained, the MRI findings support this observation.

Male and female differences
In studies on sex differences, it has been emphasized that in human brains, not only the volume of the temporal lobes and hippocampus but also the overall brain volume is larger in males (Scahill et al., 2003).Waters, Society for Women's Health Research Alzheimer's Disease Network, and Laitner (2021) reported evidence of sex-dimorphic disease effects in both schizophrenia and bipolar disorder, which may also impact Alzheimer's disease (Waters, Society for Women's Health Research Alzheimer's Disease Network, and Laitner, 2021).Thus, research into the sex differences in brains, particularly in Alzheimer's disease, is needed (Carter et al.,2012).In this study on Callithrix jacchus, significant differences in brain volume were observed only in the cerebellum, with females having a slightly smaller average brain volume compared with males, without finding significant effects in other regions (Fig. 5).However, other studies have suggested that there is no sex difference in the total brain volume of marmosets (Uematsu et al., 2017).
In diffusion metrics across various regions, significant differences in FA were observed in the total brain volume, white matter, subcortex, and retrosplenial cortex, with females tending to have lower FA values.This finding was unprecedented in other studies.However, how statistical results reflect the actual magnitude or importance of these effects remains a subject for future research.
Callithrix jacchus has garnered attention as an alternative primate model in biomedical research.Nonetheless, in fields such as reproduction, endocrine signaling, and immunology, it significantly differs from old world primates and humans.However, Callithrix jacchus exhibits classic primate characteristics (Abbott and Barnett, 2003).Such differences have contributed to the increasing use of Callithrix jacchus in biomedical research, traditionally focusing on humans.Although we have examined brain changes related to aging, a comprehensive characterization of Callithrix jacchus in physiological, reproductive, and behavioral contexts is still required.
Previous studies have reported the effects of anesthesia on brain volume measurements (Ansgar and Alex, 2010).In our study, we have used isoflurane as the anesthetic agent, and the T2-weighted imaging used to assess total intracranial accumulation was completed in a time span of 7 minutes and 24 seconds.Additionally, DWI was performed within a brief period of 90 min, leading us to suggest that the measurements of brain volume could not be significantly affected by the anesthesia.

Limitations
The Callithrix jacchus individuals in our study were captive-bred, and while we had a large number of them, the study is cross-sectional in nature.Moreover, the lifespan of Callithrix jacchus ranges from 10 to 15 years, and our oldest subject was 10 years old.Additionally, this study only examined the relationship between age, brain volume, sex differences, and body weight.Consequently, the impact of other factors on aging remains unclear.

Conclusion
Our study investigated the changes associated with aging in Callithrix jacchus, focusing on changes in total brain volume and nerve fibers, the potential relationship between total brain volume and body weight, and sex differences.This comparison is crucial for understanding the broader implications of aging across species and could guide the development of interventions applicable to human health.Our findings revealed significant differences in the brain aging process of Callithrix jacchus, particularly in terms of a decrease in total brain volume and changes in nerve fibers, contrasting with the patterns observed in humans.This study provides new knowledge about the differences in the brain aging process between humans and Callithrix jacchus.This research underscores the importance of cross-species studies in uncovering the complexities of aging, offering implications for both primatology and gerontology.Further investigation of other sites would provide a new perspective on how the aging process in the brain has changed over the course of evolution from an evolutionary biology perspective.This could provide a basis for further exploration of the relationship between evolutionary biology and aging.A limitation to consider is that our oldest marmoset was 10 years old.Therefore, results beyond this age remain unknown.Future research should address the relative stability in total brain volume as Callithrix jacchus age and explore regions not covered in this study.

Fig. 1 .
Fig. 1.MRI images (T2W, FA, AD, RD) and 3D renderings of sagittal sections of each region.The intracranial regions are color-coded.A. Color-coded intracranial region.B. Axial, sagittal, and coronal images on MRI (T2W; T2 weighted image, FA; Fractional Anisotropy, AD; Axial Diffusivity, RD; Radial Diffusivity) C. A sagittal cross section is created from the 3D rendered image of each site.

Fig. 5 .
Fig.5.Sex differences at each site A generalized linear model is used to analyze differences in brain volume between males and females.To visualize these differences, a shadow graph is created, with the vertical axis representing brain volume and the horizontal axis denoting sex. A. Differences in intracranial volume between the sexes are observed only in the cerebellum, without significant differences noted in other regions (males, n=91; females, n=125).B. Significant differences are noted in the diffusion quantification values of the total brain volume, white matter, subcortex, and retrosplenial cortex for FA (vertical axis) among the subjects (males, n=57; females, n=68).No significant differences are found in other diffusion quantification measures.

Table 1
Distribution of marmosets used by age. A: