Abstract
Associations between brain and obesity are bidirectional: changes in brain structure and function underpin over-eating, while chronic adiposity leads to brain atrophy. Investigating brain-obesity interactions across the lifespan can help better understand these relationships. This study explores the interaction between obesity and cortical morphometry in children, young adults, adults, and older adults. We also investigate the genetic, neurochemical, and cognitive correlates of the brain-obesity associations. Our findings reveal a pattern of lower cortical thickness in fronto-temporal brain regions associated with obesity across all age cohorts and varying age-dependent patterns in the remaining brain regions. In adults and older adults, obesity correlates with neurochemical changes and expression of inflammatory and mitochondrial genes. In children and older adults, adiposity is associated with modifications in brain regions involved in emotional and attentional processes. Thus, obesity might originate from cognitive changes during early adolescence, leading to neurodegeneration in later life through mitochondrial and inflammatory mechanisms.
Similar content being viewed by others
Introduction
The multifaceted nature of obesity, driven by an association between genetic and environmental factors1,2,3, has recently become a focus of scientific research. Obesity is highly heritable and genes related to excess weight gain are predominantly expressed in the central nervous system4. Given that obesity is increasingly recognised as a disorder marked by brain changes that affects individuals across the lifespan, from children to older adults5,6,7,8,9,10,11, understanding its neural correlates may provide critical insights into targeted interventions and preventive strategies tailored to specific age groups.
Increasing evidence suggests that chronic obesity can have significant effects on brain structure, function, connectivity, and thus cognition12,13. Mechanisms involved are constituents of the metabolic syndrome, namely type 2 diabetes, hypertension, inflammation, and dyslipidemia, likely exerting their effects in part via cerebrovascular disease14. Simultaneously, certain brain phenotypes are thought to constitute vulnerability factors for excess weight gain, possibly through behaviours such as impulsivity, uncontrolled eating, or executive dysfunction5,6,15,16. Because obesity-related genes affect neurodevelopment, and should therefore be present in early life, while degenerative, inflammatory or vascular brain changes accrue over time in individuals with excess weight, the neural correlates of obesity are likely to differ throughout the lifespan.
Brain map repositories allow researchers to gain a better understanding of neural changes associated with conditions like obesity. The Allen Human Brain Atlas (AHBA), a rich repository of microarray gene expression in the human brain, presents a unique opportunity to investigate the genetic undercurrents of obesity17,18. It allows for identification of cellular and molecular processes that underpin some brain changes19,20,21. Furthermore, recent integration of multiple positron emission tomography (PET) datasets by Hansen and colleagues22 enables us to deepen our understanding of neurochemical systems potentially involved in brain changes related to excess weight. Finally, automated meta-analytic tools such as Neurosynth allow for identification of brain activation patterns related to different cognitive processes, thus enabling an indirect investigation of altered cognition in obesity23.
Building upon this foundation, we aim to provide a comprehensive account of the interplay between brain structure, genetic factors, neurocognitive phenotypes, and obesity across the lifespan. Integrating multimodal data, including neuroimaging, neurochemistry, gene expression, and cognition across multiple age groups, will help uncover mechanisms related to obesity-brain interactions. The strategy here involves (1) using large-scale datasets (total n > 45,000) from the Adolescent Brain Cognitive Development (ABCD), Human Connectome Project (HCP; young adults), Human Connectome Project - Aging (HCP-A), and the UK Biobank (UKBB; older adults) to investigate cortical thickness changes related to obesity at different ages; (2) finding commonalities between those datasets and AHBA/PET/Neurosynth data on gene expression/neurotransmitter receptor and transporter distribution/cognitive processes in the brain. We focus on different age groups to explore whether obesity-related brain changes have different significance and neurocognitive underpinnings across the lifespan, as previously suggested5.
Materials and methods
Participants
Adolescent Brain Cognitive Development sample
To investigate brain correlates of obesity in early adolescence, we used the ABCD sample (data release 4.0) - a longitudinal, multi-site study24,25,26. Study procedures were approved by review boards of all participating institutions and written parental informed consent and child assent were collected for all participants. All ethical regulations relevant to human research participants were followed. We excluded all related participants, individuals with outlying BMI values (below 10 and above 50 kg/m2), and included only individuals with complete cortical thickness data. The final sample consisted of 9521 children (mean age=10 years, SD = 0.5; mean BMI = 18.92 kg/m2, SD = 4.22; 4543 girls, Fig. S1). BMI values were converted to standard deviation scores (BMI SDS) for further analysis using Center for Disease Control and Prevention’s growth charts based on individual age and sex (mean BMI SDS = 0.45, SD = 1.18).
Human Connectome Project
To determine cortical thickness correlates of obesity in young adults, we used a sample from the HCP - a study conducted at Washington University27. We excluded one of each pair of twins and individuals with endocrine disorders10. Study procedures were approved by the relevant review board and all participants signed written informed consent forms. All ethical regulations relevant to human research participants were followed. The final sample included in our study comprised 814 participants (mean age=29 years, SD = 4; mean BMI = 26.60 kg/m2, SD = 5.28; 421 women, Fig. S1).
Human Connectome Project – Aging
We used a sample of the Human Connectome Project - Aging (HCP-A) to determine brain correlates of obesity in a group of middle-aged individuals. As an extension of the HCP project, the HCP-A project recruited participants between the age of 36 and 100 years at 4 sites in the USA28,29. Study procedures were approved by the relevant review boards and all participants gave written informed consent. All ethical regulations relevant to human research participants were followed. To minimise the overlap with the age distribution of the UK Biobank sample below, we only included individuals aged between 36 and 50 years. The final sample size was 228 (mean age = 43 years, SD = 4; mean BMI = 27.71 kg/m2, SD = 5.15; 130 women, Fig. S1).
UK Biobank
To derive correlations between cortical thickness and obesity measures in older adults, we used the UKBB sample - a large-scale, multi site study from the UK30,31. The current study was performed under UKBB study ID 45551. Prior to all analyses, we excluded individuals with neurological disorders and we only included individuals with full neuroimaging data. All participants gave written informed consent and the study was approved by the North-West Multi-Centre Research Ethics Committee. All ethical regulations relevant to human research participants were followed. The final size for this sample was 36,333 individuals (mean age = 64 years, SD = 8; mean BMI = 26.28 kg/m2, SD = 4.23; 18,696 women, Fig. S1).
Neuroimaging data
Adolescent Brain Cognitive Development
Neuroimaging data were collected using 3T Siemens, General Electric, or Philips magnetic resonance imaging (MRI) scanners at 22 different sites in the United States. Data collection was harmonised by using standardised hardware and adjusting scanning sequences. Detailed imaging protocols can be found in prior publications24. Structural T1-weighted images with 1mm3 isotropic voxel size were used to obtain cortical morphometry measures. Here, we used cortical thickness data processed and provided by the ABCD initiative32. Data for each parcel of the Desikan-Killiany (DK; 68 parcels) atlas33 were processed using FreeSurfer 5.3.034 after correcting for gradient non-linearity distortions. For this study, we used parcels corresponding to the ones from the Desikan-Killiany-Tourville (DKT; 62 parcels) atlas35, thus omitting 3 parcels for each hemisphere for consistency with other data samples. Post-processing quality control was provided by the ABCD initiative in the form of fail/pass ratings. In the final sample, we excluded individuals who failed this step.
Human Connectome Project
Neuroimaging data were collected using a 3T Connectome Siemens Skyra MRI scanner. Protocol details can be found in previous publications27. Here, we utilised structural T1-weighted images with 0.7mm3 isotropic voxel size (https://www.humanconnectome.org/hcp-protocols-ya-3t-imaging). We used minimally processed data provided by the HCP36 after gradient non-linearity distortion correction. We obtained cortical thickness data for each parcel of the DKT atlas after running FreeSurfer’s 6.0.1 -recon_all function. We then visually inspected cortical segmentation and excluded all participants with unsatisfactory results.
Human Connectome Project – Aging
Neuroimaging data were collected using 3 T Siemens Prisma scanners at 4 sites using harmonised data acquisition protocols. Imaging protocols are described elsewhere29. Here, we utilised structural T1-weighted images with 0.8mm3 isotropic voxel size. We used cortical thickness data processed with FreeSurfer 6.0.0 as provided by the HCP initiative. Quality control was conducted manually to ensure proper quality of cortical segmentation. Individuals who did not pass quality control were excluded from the final sample. Finally, cortical thickness data were parcellated using the DKT atlas.
UK Biobank
Neuroimaging data were collected using 3T Siemens Skyra scanners at 3 sites in the UK. Imaging protocol details are described online at https://biobank.ctsu.ox.ac.uk/crystal/refer.cgi?id=2367 and in previous publications31. We used T1-weighted structural images with a 0.8mm3 isotropic voxel size. Here, we used cortical thickness imaging-derived phenotypes for each parcel of the DKT atlas provided by the UKBB37,38. Data were obtained using FreeSurfer 6.0.0 and quality control was performed by the UKBB initiative.
Positron emission tomography data
To analyse spatial correspondence between obesity maps and PET neurotransmitter receptor/transporter maps, we used curated data from Hansen et al. 22, as implemented in the ‘neuromaps’ package39. ‘Neuromaps’ is a toolbox that contains several curated brain maps and software tools that allow researchers to make comparisons between different brain maps. The dataset used here contains atlas data from over 1,200 individuals of 19 neurotransmitter systems (receptors and transporters): dopamine D1 receptor40, D2 receptor41,42,43,44, dopamine transporter45,46, serotonin 1a receptor47,48, serotonin 1b receptor47,48,49, serotonin 2a receptor47,48, serotonin 4 receptor48, serotonin 6 receptor50, serotonin transporter47,48, noradrenaline transporter51,52, α4β2 nicotinic receptor53, cannabinoid receptor 154,55, GABAa receptor45,56, histamine H3 receptor57, muscarinic M1 receptor58, metabotropic glutamate receptor 559,60, μ-opioid receptor61,62, and vesicular acetylcholine transporter63,64. For a detailed list of studies from which the data are derived see Supplementary Data 3 in ref. 22. Original data were parcellated using the DKT atlas for further processing using ‘neuromaps’39.
Allen Human Brain Atlas dataset
To explore the relationship between gene expression in brain tissue and obesity maps, we used gene expression data from the AHBA18. Briefly, this dataset contains microarray gene expression data from 6 post-mortem brains. The average age of brain donors is 43 years (range: 24–57 years). Analysis of this dataset was restricted to the left hemisphere, as the right hemisphere data are available from only 2 post-mortem brains. We used the ‘abagen’ toolbox for preprocessing of the genetic data according to previous recommendations65,66, which resulted in expression maps of 15,633 genes. We mapped the data to the DKT parcellation for further analyses.
BrainSpan dataset
Because the average age of AHBA donors is 43 years and gene expression in the brain is different in childhood67, we replicated our analysis comparing brain gene expression and obesity maps from the ABCD dataset using a sample of donors from the BrainSpan dataset (www.brainspan.org)68. We selected 5 specimens obtained from children around the age of the ABCD dataset (mean age: 10 years; range: 4-15 years) and analysed gene expression of 13,786 genes. To enable comparison between datasets, we only selected genes whose expression data were also available in the AHBA atlas. In the BrainSpan dataset, only 11 cortical regions of interest are available. We mapped these regions onto 11 DKT parcels as in ref. 69 for further analyses.
Neurosynth cognitive processes dataset
Next, we investigated how brain changes in obesity were associated with cognitive function. To this end, we used data from Neurosynth, a meta-analytic database of over 15,000 functional MRI studies (www.neurosynth.org)23. Every brain voxel regional value extracted from the database represents the probability of said voxel being activated in a task corresponding to a specific meta-analytic term. As previously70, we used 123 cognitive process terms from the Cognitive Atlas (www.cognitiveatlas.org)71. Maps of all cognitive processes can be found at https://github.com/netneurolab/hansen_receptors. A list of included cognitive processes can be found in the supplement to ref. 22. Data were parcellated according to the DKT atlas for further analyses.
Statistics and reproducibility
Scripts associated with this study can be found at https://github.com/FilipMorys/Obesity_maps and in ref. 72.
Associations between BMI and cortical thickness
We used linear regression analyses to investigate the association between each parcel of the DKT atlas and BMI, separately for each sample. In the analysis, we corrected for age, sex, imaging site (for multisite studies), education (parental education in the ABCD sample), socioeconomic status (household income in the ABCD sample, individual income in the HCP and HCP-A samples, Townsend deprivation index in the UKBB sample73), and additionally for scanning date in the UKBB, as per previous recommendations74. We also repeated the analyses without regressing out the socioeconomic status indicators. Obesity maps were created as t-values of the associations between cortical thickness and BMI for each parcel of the DKT atlas.
Neurotransmitter-cortical thickness spatial correspondence analysis
Correspondence analysis between obesity maps and neurotransmitter maps were performed on the DKT parcellated data using ‘neuromaps’. We used Pearson correlation analysis to determine correlation coefficients, and permutation spin tests (n = 10,000) for parcellated data to determine p-values while accounting for spatial autocorrelation in brain data75,76. Spin tests project data onto a sphere and rotate it to create spatially constrained permuted datasets for significance testing75. Correlation coefficients are recalculated for each of the permuted datasets thus creating a null distribution against which the true correlation coefficient is compared. Here, the significance threshold α was 0.05, meaning that the correlation coefficient was deemed significant if its absolute value was higher than 95% of the coefficients in the null distribution. Analyses were corrected for multiple comparisons using false discovery rate correction within each sample77.
Gene-cortical thickness spatial correspondence analysis
We used partial least squares (PLS) regression analysis to investigate the correspondence between gene expression and obesity maps. PLS finds latent components that maximise covariance between 2 datasets - here genes from the AHBA/BrainSpan and obesity map per sample. Each latent component thus represents a gene expression pattern (gene weights) that is related to the obesity brain map. Significance of each latent component was calculated using permutation tests (n = 10,000) used to create a null distribution while accounting for spatial autocorrelation between cortical thickness data based on scripts at https://github.com/KirstieJane/NSPN_WhitakerVertes_PNAS2016/ and https://github.com/gecthomas/QSM_and_AHBA_transcription_in_PD78,79,80. In each PLS analysis, we tested 5 latent components. The strength of contribution of each gene to the latent components was assessed using bootstrapping (randomly resampling rows of brain and genetic data matrices with replacement; n = 20,000) and expressed as bootstrap ratios (BR) - a ratio of the original gene weights to the standard error estimated from bootstrapping. BRs are akin to z-scores under some conditions81,82. Genes with absolute values of BR higher than 3 (comparable to p < 0.001) were deemed as robustly contributing to the latent components and were retained for further analyses. Individual gene weights obtained from PLS for each latent component were summarised for each brain parcel and expressed as gene scores, thus yielding a brain map of a latent component-related gene expression pattern. If the general correlation between gene scores and obesity maps is positive, then the genes positively contributing to the latent component represent sets with higher expression in brain areas positively related to BMI, while the opposite is true for the genes negatively associated with the latent components.
Gene overrepresentation analysis
Here, we investigated biological processes and cellular components that were related to the gene sets associated with obesity maps for each sample. To this end, we performed statistical overrepresentation analysis using the WEB-based GEne SeT AnaLysis Toolkit (http://webgestalt.org)83. All genes considered in our analyses were used as a reference list and lists of the most robustly contributing genes were used as data of interest. Lists were created separately for genes positively and negatively associated with latent components. False discovery rate (FDR) correction as implemented in WebGestalt was used to correct for multiple comparisons in the gene overrepresentation analysis. For ease of interpretation, we considered 10 processes/components most significantly associated with the gene sets.
Cognitive process-cortical thickness spatial correspondence analysis
As the last step, we tested how functional brain activations for 123 different cognitive processes correspond to obesity maps. Here, we used PLS analysis similarly to the previous step, where the gene matrix was substituted with a matrix of cognitive processes and their activation likelihood for each parcel of the DKT atlas. The remaining steps were identical to those in section ‘Gene-cortical thickness spatial correspondence analysis’. Because the Neurosynth database does not differentiate between activations and deactivations, we cannot interpret the directionality of associations between obesity and cognitive maps.
Results
Associations between cortical thickness and BMI
We performed general linear model analyses to explore the correlations between cortical thickness and BMI across participants in 4 different age groups. Overall, we found negative associations with BMI in the ABCD sample, with the most pronounced effects in the fronto-temporal regions (Supplementary Data 1–4, Fig. 1). Throughout the ageing process, in all tested samples, we observed a negative relationship between BMI and cortical thickness in the temporal and inferior frontal brain regions. Further, in adult samples, we observed some positive associations between BMI and cortical thickness, predominantly in the dorsal frontal and occipital regions of the brain. The effects of BMI on cortical thickness were almost identical when socioeconomic status was not regressed out (Fig. S2).
a Relationship between BMI and cortical thickness in the ABCD sample; b Relationship between BMI and cortical thickness in the HCP sample; c Relationship between BMI and cortical thickness in the HCP-A sample; d Relationship between BMI and cortical thickness in the UKBB sample; e correlations between obesity maps and neurotransmitter systems (chosen neurotransmitter maps derived from largest samples - see Fig. S3 for all maps); * denotes significant associations. BMI body mass index, CT cortical thickness, GABAa gamma-aminobutyric acid receptor a. 5HT1b - serotonin 1b receptor. M1 - muscarinic M1 receptor. NET - norepinephrine transporter. 5HT6 - serotonin 6 receptor. a4b2 NR - α4β2 nicotinic receptor. mGluR5 - metabotropic glutamate receptor 5. 5HT2a - serotonin 2a receptor. D1 - dopamine D1 receptor. H3 - histamine H3 receptor. 5HTT - serotonin transporter. VAChT - vesicular acetylcholine transporter. CB1 - cannabinoid receptor 1. 5HT4 - serotonin 4 receptor. DAT - dopamine transporter. MOR - μ-opioid receptor. D2 - dopamine D2 receptor. 5HT1a - serotonin 1a receptor.
Correspondence between obesity maps and neurotransmitter systems
We compared obesity maps in different age groups with neurotransmitter maps from ref. 22 and found no significant associations in the ABCD, HCP, or HCP-A sample. We found significant negative correlations between obesity maps and dopamine transporter, D2 receptor, vesicular acetylcholine transporter, and serotonin 1a receptor, and positive associations with GABAa receptor and serotonin 1b receptor in the UKBB sample (Fig. 1, Fig. S3, Supplementary Data 5). Supplementary analysis correlating age and association strength between PET data and obesity maps showed that age of the PET study samples did not influence the correlations with different age groups of the tested cohorts (Fig. S4).
Correspondence between obesity maps and gene expression
Using PLS, we identified latent components relating obesity maps to gene expression maps from the AHBA and BrainSpan atlases. We did not find any significant components for the ABCD or HCP samples (with either AHBA or BrainSpan atlases). We found one significant component for the HCP-A sample (p < 0.001, r = 0.654; Fig. 2) that positively related the cortical thickness-BMI map to a latent component of gene expression patterns. This component explained 42.75% of variance in the brain data. Similarly, we found one significant component for the UKBB sample (p < 0.001, r = 0.869; Fig. 3). This component explained 75% of variance in the brain data. All remaining components did not reach statistical significance. Positive correlations between gene scores and obesity brain maps mean that genes negatively related to the component (negative genes) were expressed to a higher degree in brain regions with negative BMI-cortical thickness associations and to a lower degree in brain regions with positive BMI-cortical thickness associations, with an opposite relationship for genes positively related to the component (positive genes). Overall, we found increasing correspondence between obesity maps and gene expression with increasing age of the tested sample.
a Gene scores plotted on the brain showing a summary map of gene expression related to BMI-CT relationship in the HCP-A sample; b Relationship between BMI-CT values and gene scores from the PLS analysis; c–f Gene overexpression analysis terms associated with sets of significant genes in the PLS analysis in the HCP-A sample. BMI - body mass index. CT cortical thickness, FDR false discovery rate.
a Gene scores plotted on the brain showing a summary map of gene expression related to BMI-CT relationship in the UKBB sample; b Relationship between UKBB values and gene scores from the PLS analysis; c–f Gene overexpression analysis terms associated with sets of significant genes in the PLS analysis in the UKBB sample. BMI body mass index, CT cortical thickness, FDR false discovery rate.
Gene overexpression analysis
Using the WebGestalt tool83, we analysed the sets of genes associated with significant PLS components from the previous step to interpret biological processes and cellular components to which the gene expression is related. In the HCP-A sample, lower cortical thickness (negative genes) was significantly associated with regulation of biosynthetic processes, chromosome organisation, and chromosomal cell components, while higher cortical thickness (positive genes) was associated with metabolic processes, creation of ATP, and mitochondria (Fig. 2). In the UKBB sample, lower cortical thickness (negative genes) was associated with inflammation-related processes, such as leukocyte degranulation or general immune response, while higher cortical thickness (positive genes) was related to biosynthetic and metabolic processes, chromatin organisation, and cytoskeleton, chromosome, and nuclear body cell components (Fig. 3).
Correspondence between obesity maps and brain signatures of cognitive processes
Significant associations between obesity maps and Neurosynth cognitive maps were found in the ABCD and UKBB samples. In the ABCD sample, the first component explained 36% of variance in the brain data (p < 0.001, ρ = 0.601; Fig. 4). Cognitive processes related to obesity-associated brain changes in this sample are depicted in Fig. 4 and include visual perception, attention, intelligence, decision making, impulsivity, risk taking, and emotion regulation. In the UKBB sample, the first component explained 48.5% variance in the brain data (p < 0.001, ρ = 0.697). Cognitive processes identified in this analysis are depicted in Fig. 5 and include processes similar to the ones identified in the ABCD sample - spatial attention, visual perception, working memory, emotion regulation, fear, arousal, and anxiety. In the HCP and HCP-A samples we did not find any significant PLS components for cognitive processes (p > 0.05).
a Relationship between BMI-CT values and term scores from the PLS analysis; b term scores plotted on the brain showing a summary map of cognitive terms related to BMI-CT relationship in the ABCD sample; c cognitive terms significantly associated with obesity map in the ABCD sample (colours correspond to positive and negative associations with obesity maps as depicted in b). BMI body mass index, CT cortical thickness.
a Relationship between BMI-CT values and term scores from the PLS analysis; b term scores plotted on the brain showing a summary map of cognitive terms related to BMI-CT relationship in the UKBB sample; c cognitive terms significantly associated with obesity map in the UKBB sample (colours correspond to positive and negative associations with obesity maps as depicted in b). BMI body mass index, CT cortical thickness.
Discussion
Neuroanatomical and cognitive correlates of obesity change across the lifespan. In this study, we aimed to identify structural (cortical thickness), genetic, neurochemical, and cognitive underpinnings of obesity in different age groups. We showed that obesity is consistently related to lower cortical thickness in the temporal and inferior frontal brain regions across all age groups, with the associations becoming more pronounced with age. We also found positive relationships between cortical thickness and BMI in adults, predominantly in the dorsal frontal and occipital brain regions. In middle-aged adults, positive associations were related to gene expression related to mitochondrial processes, while in older adults, lower cortical thickness was related to inflammatory processes. We also showed here that cortical thickness changes in human obesity share some similarities with brain distributions of serotonin, dopamine, acetylcholine, and GABA neurotransmitter systems, and that these associations are only visible in older adults. Finally, we found that in children and older adults, brain changes related to obesity are located in areas involved in emotional and attentional cognitive processes (Fig. 6). Overall, and in keeping with previous studies, our results show different and more pronounced brain changes related to obesity with increasing age6,10,11,12,16,84.
Average t-values per lobe are depicted in the line plot.
We investigated lifespan changes of brain-obesity associations by testing samples of different ages. Our premise is that certain brain features in obesity underpin behaviours that lead to over-eating and excess weight gain, and that these could be present throughout life. On the other hand, negative effects of chronic adiposity on the brain should accumulate over the lifespan5,6,12. Our findings of associations between cortical thickness and obesity are mostly in line with these hypotheses, given that we found predominantly fronto-temporal cortical thinning, increasing with age. These findings are also supported by studies investigating grey matter volume and surface area showing that across different age groups obesity is negatively related to changes predominantly in the frontal but also temporal cortices6,7,8,12,84,85,86,87,88,89.
While we cannot rule out adiposity-related neurodegeneration in the 10-11-year-old ABCD cohort, cortical thickness effects in children could also reflect differences in brain maturation, which may underpin a vulnerability factor for obesity through behavioural changes, as shown previously6. Conversely, higher cortical thinning with age in adults might point to an ongoing process of neurodegeneration associated with excess weight and related comorbidities14. According to our gene expression analyses, these changes in the brain in adults might be specifically related to mitochondrial function and inflammatory processes. This is in line with a large body of literature suggesting that obesity is related to systemic inflammation, which leads to neurodegeneration and could be related to mitochondrial dysfunction12,90,91.
Neurotransmitter systems related to obesity in adulthood in this study have been previously reported as involved in control of food intake and obesity in humans and animals92,93,94,95,96,97,98,99. While spatial associations do not prove a mechanistic relationship between neurotransmitter signalling and obesity, our findings suggest some degree of involvement of those neurotransmitters - as a cause or a consequence. One reason for associations of those systems with obesity being present only in the oldest age group might be that they are a consequence of an obesity-related neurodegeneration that increases with age. Ageing is associated with neural changes within distinct neurotransmitter systems, and this might be accentuated by obesity100,101. Speculatively, this finding could suggest a vicious cycle, where obesity-related brain pathology leads to changes in the brain that promote further increased food intake and thus further brain pathology.
Finally, we characterised cognitive correlates of obesity in children and older adults, showing that obesity-related cortical thickness changes are associated with cognitive processes predominantly involved in attention and emotional processing. Previously, obesity and risk for obesity in childhood were associated with altered attention for food and non-food stimuli and emotional processing and eating compared to control groups102,103,104. In adults, emotional processing is often associated with obesity in the form of a phenotype labelled uncontrolled eating105. At the same time, adult individuals with obesity show attentional bias for food that seems to predict weight gain106,107. Together, attentional and emotional processes alongside their underlying neural correlates might constitute a vulnerability factor for weight gain that remains stable across lifespan.
Some limitations of this study have to be considered: while the ABCD and UKBB samples were large, the HCP and HCP-A samples were smaller and the results might therefore be less reliable. In addition, main analyses compared obesity maps with gene expression, neurotransmitter, and cognitive maps from other samples that differed in age, sex distribution, and likely other factors that we did not account for. These analyses are purely correlational and need to be interpreted with caution. Importantly, all analyses were performed using the DKT atlas which includes 62 cortical parcels and might therefore not distinguish between certain smaller specialised cortical areas that might be pertinent in obesity research. Relatedly, we did not investigate the involvement of subcortical structures in obesity across different ages, which, given the subcortical contributions to obesity12,86,88, is an important topic for future investigations.
In sum, we show that brain and genetic correlates of obesity change across the lifespan, supporting our hypothesis that BMI-related neural endophenotypes might have a different significance in children and adults. This is a cross-sectional study, so causality cannot be inferred. However, in connection with previous studies we believe that accelerated brain maturation in children might increase the risk for obesity - potentially through altered emotional and attentional processing - while in adults, obesity might lead to brain degeneration that involves mitochondrial function and neuroinflammation. This study contributes to the knowledge necessary to understand obesity and create targeted prevention and intervention strategies.
Data availability
Data used in this study are available through the ABCD, HCP, and UKBB consortia upon request.
Code availability
Code used to analyse the datasets was deposited at https://doi.org/10.5281/zenodo.1106853572. and at https://github.com/FilipMorys/Obesity_maps.
References
Yengo, L. et al. Meta-analysis of genome-wide association studies for height and body mass index in ~700,000 individuals of European ancestry. Hum. Mol. Genet. 27, 3641 (2018).
Kweon, H. et al. Human brain anatomy reflects separable genetic and environmental components of socioeconomic status. Sci. Adv. 8, eabm2923 (2022).
Llewellyn, C. & Wardle, J. Behavioral susceptibility to obesity: gene-environment interplay in the development of weight. Physiol. Behav. 152, 494–501 (2015).
Locke, A. E. et al. Genetic studies of body mass index yield new insights for obesity biology. Nature 518, 197–206 (2015).
García-García, I., Morys, F. & Dagher, A. Nucleus accumbens volume is related to obesity measures in an age-dependent fashion. J. Neuroendocrinol. 32, e12812 (2020).
Morys, F. et al. Neuroanatomical correlates of genetic risk for obesity in children. Transl. Psychiatry 13, 1 (2023).
Morys, F., Shishikura, M. & Dagher, A. Population-based research in obesity – an overview of neuroimaging studies using big data approach. Curr. Opin. Endocr. Metab. Res. 23, 100323 (2022).
Adise, S. et al. Multimodal brain predictors of current weight and weight gain in children enrolled in the ABCD study ®. Dev. Cogn. Neurosci. 49, 100948 (2021).
Hamer, M. & Batty, G. D. Association of body mass index and waist-to-hip ratio with brain structure. Neurology 92, e594–e600 (2019).
Vainik, U. et al. Neurobehavioral correlates of obesity are largely heritable. Proc. Natl Acad. Sci. 115, 9312–9317 (2018).
Ronan, L., Alexander-Bloch, A. & Fletcher, P. C. Childhood obesity, cortical structure, and executive function in healthy children. Cereb. Cortex 30, 2519–2528 (2019).
Morys, F., Dadar, M. & Dagher, A. Association between midlife obesity and its metabolic consequences, cerebrovascular disease, and cognitive decline. J. Clin. Endocrinol. Metab. 106, e4260–e4274 (2021).
Morys, F. et al. Obesity-associated neurodegeneration pattern mimics Alzheimer’s disease in an observational Cohort Study. J. Alzheimers Dis. 91, 1059–1071 (2023).
García-García, I. et al. Mechanisms linking obesity and its metabolic comorbidities with cerebral grey and white matter changes. Rev. Endocr. Metab. Disord. 1, 1–11 (2022).
Opel, N. et al. Brain structural abnormalities in obesity: relation to age, genetic risk, and common psychiatric disorders: Evidence through univariate and multivariate mega-analysis including 6420 participants from the ENIGMA MDD working group. Mol. Psychiatry 34, 36 (2020).
Westwater, M. L., Vilar-López, R., Ziauddeen, H., Verdejo-García, A. & Fletcher, P. C. Combined effects of age and BMI are related to altered cortical thickness in adolescence and adulthood. Dev. Cogn. Neurosci. 40, 100728 (2019).
Shen, E. H., Overly, C. C. & Jones, A. R. The Allen Human Brain Atlas: comprehensive gene expression mapping of the human brain. Trends Neurosci. 35, 711–714 (2012).
Hawrylycz, M. J. et al. An anatomically comprehensive atlas of the adult human brain transcriptome. Nature 489, 391–399 (2012).
Rahayel, S. et al. Mitochondrial function-associated genes underlie cortical atrophy in prodromal synucleinopathies. Brain 8, 3301–3318 (2023).
Tremblay, C. et al. Brain atrophy progression in Parkinson’s disease is shaped by connectivity and local vulnerability. Brain Commun. 3, fcab269 (2021).
Mueller, K. et al. Overweight and obesity are associated with neuronal injury in the human cerebellum and hippocampus in young adults: a combined MRI, serum marker and gene expression study. Transl. Psychiatry 2, e200 (2012).
Hansen, J. Y. et al. Mapping neurotransmitter systems to the structural and functional organization of the human neocortex. Nat. Neurosci. 25, 1569–1581 (2022).
Yarkoni, T., Poldrack, R. A., Nichols, T. E., Van Essen, D. C. & Wager, T. D. Large-scale automated synthesis of human functional neuroimaging data. 8, 665–670 (2011).
Casey, B. J. et al. The Adolescent Brain Cognitive Development (ABCD) study: imaging acquisition across 21 sites. Dev. Cogn. Neurosci. 32 43–54 (2018).
Garavan, H. et al. Recruiting the ABCD sample: Design considerations and procedures. Dev. Cogn. Neurosci. 32, 16–22 (2018).
Barch, D. M. et al. Demographic, physical and mental health assessments in the adolescent brain and cognitive development study: Rationale and description. Dev. Cogn. Neurosci. 32 55–66 (2018).
Van Essen, D. C. et al. The WU-minn human connectome project: an overview. Neuroimage 80, 62–79 (2013).
Bookheimer, S. Y. et al. The lifespan human connectome project in aging: an overview. Neuroimage 185, 335–348 (2019).
Harms, M. P. et al. Extending the Human Connectome Project across ages: imaging protocols for the Lifespan Development and Aging projects. Neuroimage 183, 972–984 (2018).
Sudlow, C. et al. UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 12, e1001779 (2015).
Miller, K. L. et al. Multimodal population brain imaging in the UK Biobank prospective epidemiological study. Nat. Neurosci. 19, 1523–1536 (2016).
Hagler, D. J. et al. Image processing and analysis methods for the Adolescent Brain Cognitive Development Study. Neuroimage 202, 116091 (2019).
Desikan, R. S. et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 31, 968–980 (2006).
Fischl, B. FreeSurfer. Neuroimage 62, 774–781 (2012).
Klein, A. & Tourville, J. 101 Labeled brain images and a consistent human cortical labeling protocol. Front. Neurosci. 6, 171 (2012).
Glasser, M. F. et al. The minimal preprocessing pipelines for the Human Connectome Project. Neuroimage 80, 105–124 (2013).
Alfaro-Almagro, F. et al. Image processing and Quality Control for the first 10,000 brain imaging datasets from UK Biobank. Neuroimage 166, 400–424 (2018).
Littlejohns, T. J. et al. The UK Biobank imaging enhancement of 100,000 participants: rationale, data collection, management and future directions. Nat. Commun. 11, 2624 (2020).
Markello, R. D. et al. neuromaps: structural and functional interpretation of brain maps. Nat. Methods 19, 1472–1479 (2022).
Kaller, S. et al. Test-retest measurements of dopamine D1-type receptors using simultaneous PET/MRI imaging. Eur. J. Nucl. Med. Mol. Imaging 44, 1025–1032 (2017).
Jaworska, N. et al. Extra-striatal D2/3 receptor availability in youth at risk for addiction. Neuropsychopharmacology 45, 1498–1505 (2020).
Smith, C. T. et al. Partial-volume correction increases estimated dopamine D2-like receptor binding potential and reduces adult age differences. J. Cereb. Blood Flow. Metab. 39, 822–833 (2019).
Sandiego, C. M. et al. Reference region modeling approaches for amphetamine challenge studies with [11C]FLB 457 and PET. J. Cereb. Blood Flow. Metab. 35, 623–629 (2015).
Alakurtti, K. et al. Long-term test-retest reliability of striatal and extrastriatal dopamine D2/3 receptor binding: study with [(11)C]raclopride and high-resolution PET. J. Cereb. Blood Flow. Metab. 35, 1199–1205 (2015).
Dukart, J. et al. Cerebral blood flow predicts differential neurotransmitter activity. Sci. Rep. 8, 4074 (2018).
Sasaki, T. et al. Quantification of dopamine transporter in human brain using PET with 18F-FE-PE2I. J. Nucl. Med. 53, 1065–1073 (2012).
Savli, M. et al. Normative database of the serotonergic system in healthy subjects using multi-tracer PET. Neuroimage 63, 447–459 (2012).
Beliveau, V. et al. A high-resolution in vivo atlas of the human brain’s serotonin system. J. Neurosci. 37, 120 (2017).
Gallezot, J.-D. et al. Kinetic modeling of the serotonin 5-HT(1B) receptor radioligand [(11)C]P943 in humans. J. Cereb. Blood Flow. Metab. 30, 196–210 (2010).
Radhakrishnan, R. et al. Age-related change in 5-HT6 receptor availability in healthy male volunteers measured with 11C-GSK215083 PET. J. Nucl. Med. 59, 1445–1450 (2018).
Ding, Y.-S. et al. PET imaging of the effects of age and cocaine on the norepinephrine transporter in the human brain using (S,S)-[(11)C]O-methylreboxetine and HRRT. Synapse 64, 30–38 (2010).
Hesse, S. et al. Central noradrenaline transporter availability in highly obese, non-depressed individuals. Eur. J. Nucl. Med. Mol. Imaging 44, 1056–1064 (2017).
Hillmer, A. T. et al. Imaging of cerebral α4β2* nicotinic acetylcholine receptors with (-)-[(18)F]Flubatine PET: Implementation of bolus plus constant infusion and sensitivity to acetylcholine in human brain. Neuroimage 141, 71–80 (2016).
Normandin, M. D. et al. Imaging the cannabinoid CB1 receptor in humans with [11C]OMAR: assessment of kinetic analysis methods, test-retest reproducibility, and gender differences. J. Cereb. Blood Flow. Metab. 35, 1313–1322 (2015).
Laurikainen, H. et al. Sex difference in brain CB1 receptor availability in man. Neuroimage 184, 834–842 (2019).
Nørgaard, M. et al. A high-resolution in vivo atlas of the human brain’s benzodiazepine binding site of GABAA receptors. Neuroimage 232, 117878 (2021).
Gallezot, J.-D. et al. Determination of receptor occupancy in the presence of mass dose: [11C]GSK189254 PET imaging of histamine H3 receptor occupancy by PF-03654746. J. Cereb. Blood Flow. Metab. 37, 1095–1107 (2017).
Naganawa, M. et al. First-in-human assessment of 11C-LSN3172176, an M1 muscarinic acetylcholine receptor PET radiotracer. J. Nucl. Med. 62, 553–560 (2021).
DuBois, J. M. et al. Characterization of age/sex and the regional distribution of mGluR5 availability in the healthy human brain measured by high-resolution [(11)C]ABP688 PET. Eur. J. Nucl. Med. Mol. Imaging 43, 152–162 (2016).
Smart, K. et al. Sex differences in [11C]ABP688 binding: a positron emission tomography study of mGlu5 receptors. Eur. J. Nucl. Med. Mol. Imaging 46, 1179–1183 (2019).
Turtonen, O. et al. Adult Attachment System Links With Brain Mu Opioid Receptor Availability In Vivo. Biol. Psychiatry Cogn. Neurosci. Neuroimaging 6, 360–369 (2021).
Kantonen, T. et al. Interindividual variability and lateralization of μ-opioid receptors in the human brain. Neuroimage 217, 116922 (2020).
Aghourian, M. et al. Quantification of brain cholinergic denervation in Alzheimer’s disease using PET imaging with [18F]-FEOBV. Mol. Psychiatry 22, 1531–1538 (2017).
Bedard, M.-A. et al. Brain cholinergic alterations in idiopathic REM sleep behaviour disorder: a PET imaging study with 18F-FEOBV. Sleep. Med. 58, 35–41 (2019).
Arnatkeviciute, A., Fulcher, B. D. & Fornito, A. A practical guide to linking brain-wide gene expression and neuroimaging data. Neuroimage 189, 353–367 (2019).
Markello, R. D. et al. Standardizing workflows in imaging transcriptomics with the abagen toolbox. Elife 10, e72129 (2021).
Kang, H. J. et al. Spatio-temporal transcriptome of the human brain. Nature 478, 483–489 (2011).
Li, M. et al. Integrative functional genomic analysis of human brain development and neuropsychiatric risks. Science 362, eaat7615 (2018).
Paquola, C. et al. A multi-scale cortical wiring space links cellular architecture and functional dynamics in the human brain. PLoS Biol. 18, e3000979 (2020).
Hansen, J. Y. et al. Mapping gene transcription and neurocognition across human neocortex. Nat. Hum. Behav. 5, 1240–1250 (2021).
Poldrack, R. A. et al. The cognitive atlas: toward a knowledge foundation for cognitive neuroscience. Front. Neuroinform. 5, 17 (2011).
Morys, F. et al. Neural correlates of obesity across the lifespan [Analysis code]. Zenodo https://doi.org/10.5281/zenodo.11068535 (2024).
Townsend, P., Phillimore, P. & Beattie, A. Health and Deprivation: Inequality and the North. (Croom Helm, 1988).
Alfaro-Almagro, F. et al. Confound modelling in UK Biobank brain imaging. Neuroimage 224, 117002 (2021).
Alexander-Bloch, A. F. et al. On testing for spatial correspondence between maps of human brain structure and function. Neuroimage 178, 540–551 (2018).
Markello, R. D. & Misic, B. Comparing spatial null models for brain maps. Neuroimage 236, 118052 (2021).
Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. 57, 289–300 (1995).
Whitaker, K. J. et al. Adolescence is associated with genomically patterned consolidation of the hubs of the human brain connectome. Proc. Natl Acad. Sci. USA 113, 9105–9110 (2016).
Thomas, G. E. C. et al. Regional brain iron and gene expression provide insights into neurodegeneration in Parkinson’s disease. Brain 144, 1787–1798 (2021).
Váša, F. et al. Adolescent tuning of association cortex in human structural brain networks. Cereb. Cortex 28, 281–294 (2018).
Mišić, B. et al. Network-level structure-function relationships in human neocortex. Cereb. Cortex 26, 3285–3296 (2016).
Efron, B. & Tibshirani, R. Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy. Stat. Sci. 1, 54–75 (1986).
Liao, Y., Wang, J., Jaehnig, E. J., Shi, Z. & Zhang, B. WebGestalt 2019: gene set analysis toolkit with revamped UIs and APIs. Nucleic Acids Res. 47, W199–W205 (2019).
Laurent, J. S. et al. Associations among Body Mass Index, Cortical Thickness, and Executive Function in Children. JAMA Pediatr. 174, 170–177 (2020).
García-García, I. et al. Neuroanatomical differences in obesity: meta-analytic findings and their validation in an independent dataset. Int. J. Obes. 43, 943–951 (2018).
Zapparoli, L. et al. Neural structural abnormalities behind altered brain activation in obesity: evidence from meta-analyses of brain activation and morphometric data. Neuroimage Clin. 36, 103179 (2022).
Chen, E. Y., Eickhoff, S. B., Giovannetti, T. & Smith, D. V. Obesity is associated with reduced orbitofrontal cortex volume: a coordinate-based meta-analysis. Neuroimage Clin. 28, 102420 (2020).
Herrmann, M. J., Tesar, A. K., Beier, J., Berg, M. & Warrings, B. Grey matter alterations in obesity: A meta-analysis of whole-brain studies. Obes. Rev. 20 464–471 (2019).
Jiang, F. et al. Obesity is associated with decreased gray matter volume in children: a longitudinal study. Cereb. Cortex 33, 3674–3682 (2023).
Lampe, L. et al. Visceral obesity relates to deep white matter hyperintensities via inflammation. Ann. Neurol. 85, ana.25396 (2018).
de Oliveira, L. G., Angelo, Y. S., Iglesias, A. H. & Peron, J. P. S. Unraveling the link between mitochondrial dynamics and neuroinflammation. Front. Immunol. 12, 624919 (2021).
van Galen, K. A., Ter Horst, K. W. & Serlie, M. J. Serotonin, food intake, and obesity. Obes. Rev. 22, e13210 (2021).
Martinelli, I. et al. Obesity-related brain cholinergic system impairment in high-fat-diet-fed rats. Nutrients 14, 1243 (2022).
DiFeliceantonio, A. G. & Small, D. M. Dopamine and diet-induced obesity. Nat. Neurosci. 22, 1–2 (2019).
Horstmann, A., Fenske, W. K. & Hankir, M. K. Argument for a non-linear relationship between severity of human obesity and dopaminergic tone. Obes. Rev. https://doi.org/10.1111/obr.12303 (2015).
Dagher, A. Functional brain imaging of appetite. Trends Endocrinol. Metab. 23, 250–260 (2012).
Karlsson, H. K. et al. Obesity is associated with decreased -opioid but unaltered dopamine D2 receptor availability in the brain. J. Neurosci. 35, 3959–3965 (2015).
Schulz, P. et al. What role does the endocannabinoid system play in the pathogenesis of obesity? Nutrients 13, 373 (2021).
Delgado, T. C. Glutamate and GABA in appetite regulation. Front. Endocrinol. 4, 103 (2013).
Lee, J. & Kim, H.-J. Normal aging induces changes in the brain and neurodegeneration progress: review of the structural, biochemical, metabolic, cellular, and molecular changes. Front. Aging Neurosci. 14, 931536 (2022).
Lotter, L. D. et al. Human cortex development is shaped by molecular and cellular brain systems. Preprint at bioRxiv https://doi.org/10.1101/2023.05.05.539537 (2023).
Mehl, N. et al. Cause or consequence? Investigating attention bias and self-regulation skills in children at risk for obesity. J. Exp. Child Psychol. 155, 113–127 (2017).
Parkinson, K. N., Drewett, R. F., Le Couteur, A. S., Adamson, A. J. & Gateshead Milennium Study Core Team. Do maternal ratings of appetite in infants predict later Child Eating Behaviour Questionnaire scores and body mass index? Appetite 54, 186–190 (2010).
Ashcroft, J., Semmler, C., Carnell, S., van Jaarsveld, C. H. M. & Wardle, J. Continuity and stability of eating behaviour traits in children. Eur. J. Clin. Nutr. 62, 985–990 (2008).
Vainik, U., García‐García, I. & Dagher, A. Uncontrolled eating: a unifying heritable trait linked with obesity, overeating, personality and the brain. Eur. J. Neurosci. 50, 2430–2445 (2019).
Castellanos, E. H. et al. Obese adults have visual attention bias for food cue images: evidence for altered reward system function. Int. J. Obes. 33, 1063–1073 (2009).
Kaisari, P. et al. Top-down guidance of attention to food cues is enhanced in individuals with overweight/obesity and predicts change in weight at one-year follow up. Int. J. Obes. 43, 1849–1858 (2019).
Acknowledgements
This work was supported by a Foundation Scheme award to AD from the Canadian Institutes of Health Research, Fonds de recherche du Québec-Santé postdoctoral fellowship to FM. A portion of the data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development SM (ABCD) Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children aged 9–10 and follow them over 10 years into early adulthood. The ABCD Study® is supported by the National Institutes of Health and additional federal partners under award numbers U01DA041048, U01DA050989, U01DA051016, U01DA041022, U01DA051018, U01DA051037, U01DA050987, U01DA041174, U01DA041106, U01DA041117, U01DA041028, U01DA041134, U01DA050988, U01DA051039, U01DA041156, U01DA041025, U01DA041120, U01DA051038, U01DA041148, U01DA041093, U01DA041089, U24DA041123, U24DA041147. A full list of supporters is available at https://abcdstudy.org/federal-partners.html. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/consortium_members/. ABCD consortium investigators designed and implemented the study and/or provided data but did not participate in the analysis or writing of this report. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIH or ABCD consortium investigators. The ABCD data repository grows and changes over time. The ABCD Study data used in this report came from https://doi.org/10.15154/zw9a-7t30. Data were provided in part by the Human Connectome Project, WU-Minn Consortium (Principal Investigators: David Van Essen and Kamil Ugurbil; 1U54MH091657) funded by the 16 NIH Institutes and Centers that support the NIH Blueprint for Neuroscience Research; and by the McDonnell Center for Systems Neuroscience at Washington University. This research has been conducted using data from UK Biobank, a major biomedical database.
Author information
Authors and Affiliations
Contributions
F.M. and A. Dagher designed the study; F.M., C.T., S.R., J.Y.H., and A. Dai analysed the data; all authors took part in results interpretation and manuscript preparation.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Communications Biology thanks Kathleen Page and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary handling editors: Jessica Peter, George Inglis, and Dario Ummarino. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Morys, F., Tremblay, C., Rahayel, S. et al. Neural correlates of obesity across the lifespan. Commun Biol 7, 656 (2024). https://doi.org/10.1038/s42003-024-06361-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s42003-024-06361-9
- Springer Nature Limited