Abstract
Objectives
The underlying mechanism of major depressive disorder (MDD) with eating disorder (ED) remains unclear. We aim to clarify the characteristic changes of THs and brain neurometabolic alterations in ED, and to explore their relationships.
Methods
The study included 28 individuals with MDD and ED, 81 individuals with MDD without ED, and 37 age-matched healthy controls (HCs). Serum TH levels were assessed, and 1H proton magnetic spectroscopy was utilized to determine the N-acetylaspartic acid to creatine (NAA/Cr) and choline-containing compounds to creatine ratios in the prefrontal cortex, anterior cingulate cortex, and thalamus. Subsequently, differential analysis, receiver operating characteristic (ROC) analysis, and correlation analysis were performed to explore their characteristics and interrelationships.
Results
In both MDD with ED and MDD without ED cohorts, free tri-iodothyronine (FT3) levels were significantly lower compared to HCs, whereas free thyroxine (FT4) and total thyroxine (TT4) levels were elevated. Significantly lower NAA/Cr ratios were observed in the right thalamus and higher NAA/Cr ratios in the left cerebellum in both MDD with ED and MDD without ED compared to HCs. Neurometabolic factors and THs levels achieved an ROC curve area of 0.830 in differentiating MDD with ED from MDD without ED. In addition, serum FT3 and TT4 levels showed a positive correlation with NAA/Cr in the eft cerebellum in cases of MDD with ED.
Conclusions
Our findings reveal concurrent thyroid hormone irregularities and neurometabolic changes in the thalamic-cerebellum circuitry in MDD with ED, providing preliminary insights into the neurobiology of abnormal eating behaviors in depression. Given the cross-sectional design, these results are exploratory and require validation.
Level of evidence Level IV, cross-sectional study.
Keywords: Major depressive disorder, Eating disorder, Thyroid hormones, Neurometabolic, Cerebellum
Introduction
Major depressive disorder (MDD) is a prevalent psychiatric condition marked by a significant and enduring depressive state, associated with high incidence and suicide risk. According to the 2019 China Epidemic Survey [1], the lifetime prevalence of depression is 3.4%, with an incidence rate of 6.2%. In the last 10 years, there has been a significant rise in psychiatric disorders like eating disorders (EDs), which are often comorbid with depression, among young adults [2, 3]. EDs encompass conditions characterized by negative perceptions of food, body shape, and weight, accompanied by behaviors such as restricted eating, binge eating, excessive exercise, self-induced vomiting, and laxative use, including anorexia nervosa and bulimia nervosa [4]. A survey of American college students from 2013 to 2021 indicated a significant rise in ED risk prevalence from 15 to 28% [5]. EDs are prevalent psychiatric conditions linked to elevated mortality rates. In addition, they were highly significant risk factors for each other [6]. Studies have found that the risk of EDs in patients with MDD is 1.54 times higher than that in normal people [7]. The comorbid mechanisms remain unclear, but abnormalities in the hypothalamic–pituitary–thyroid (HPT) axis [8] and abnormal neural activation patterns [9] may significantly influence the biological mechanisms of emotion regulation and eating behaviors, as well as their pathophysiology and treatment. Understanding the neurobiological mechanisms underlying comorbid EDs in depression is crucial, as it may lead to identifying new therapeutic targets to improve treatment outcomes.
Extensive experimental evidence suggests that the HPT axis affects metabolic activity in the brain, promotes serotonergic neurotransmission and affects brain neurotransmitter systems, such as dopamine and norepinephrine, and is associated with depression [10]. In patients with MDD, Hypothalamic thyroid-stimulating hormone releasing hormone acts as a homeostatic neuromodulator of monoaminergic neurotransmission and can maintain the efficiency of the dopamine system [11]. Our earlier research indicated significant reductions in total triiodothyronine (TT3) and thyroid-stimulating hormone (TSH) levels in patients. In addition, changes in TSH, free tri-iodothyronine (FT3), and free thyroxine (FT4) levels, while remaining within normal limits, were associated with depression severity [12].
The role of thyroid hormones (THs) in regulating mood in patients is yet to be clarified. These patients present with obvious HPT axis alterations that have significant effects on metabolic homeostasis [13]. Patients with anorexia nervosa had lower levels of FT3, FT4, and FT3/FT4 ratio, and TSH was associated with more severe depressive symptoms [8].Contrast analyses showed increasing levels of total cholesterol and TSH in binge eating disorder subjects, whereas bulimia nervosa subjects remained stable, which was significantly different at the final follow-up [14].
Therefore, HPT axis alterations may be involved as neuroendocrine factors in regulating EDs comorbid MDD. Although research indicates a bidirectional relationship between the two variables, there is a lack of sufficient neuroimaging findings in MDD patients. Neuroimaging techniques can help investigate the underlying mechanisms of eating abnormalities in individuals with major depression. Investigating the link between thyroxine and alterations in brain chemical metabolism in depressed patients with EDs is crucial for comprehending the physical and mental health of this population. However, cerebral neurometabolic processes in patients with eating abnormalities have not been elucidated. Neuroimaging studies have been extensively utilized to detect structural and functional abnormalities in critical brain regions associated with MDD [15]. A recent meta-analysis of task-based fMRI in individuals with anorexia nervosa revealed heightened activity in the left superior temporal gyrus and bilateral anterior cingulate cortex during reward-related tasks. Bilaterally decreased activity in the cerebellum and bilateral precuneus, aberrant neural activation patterns of anorexia nervosa may be associated with the limbic system, prefrontal cortex, temporal gyrus, insula, and cerebellum, and exhibit consistent patterns of brain activity associated with food, physical, emotional, cognitive, and reward processes [16]. Changes in these brain regions are critical to the performance of feeding function. Studies have found that subjects with high food addiction have higher activation in the thalamus, hypothalamus, midbrain, putamen, and occipital cortex [17]. Research utilizing diffusion tensor imaging has explored microstructural changes in the white matter of patients with restrictive EDs, revealing a trend of negative correlation between fractional anisotropy behind the thalamus and ED symptoms. This suggests that the integrity of limbic thalamic-cortical circuit projections is crucial for taste pathways and projections in prefrontal and reward-related regions, with alterations potentially disrupting food-related cognitive processing in adolescents [18]. Several studies have shown that the cerebellum, striatum, amygdala, and hypothalamus play important roles in eating behavior and appetite regulation [19]. The thalamus and cerebellum are the core components of the reward neural circuit. Therefore, the activation of these brain regions may be closely related to depression accompanied by abnormal eating.
1H proton magnetic spectroscopy (1H-MRS) enables in vivo quantification of metabolite concentrations in brain fluids. 1H-MRS is an important fMRI imaging method that can be used to study brain metabolites, including N-acetylaspartic acid (NAA), choline-containing compounds (Cho), creatinine/phosphocreatine (Cr), glutamate, gamma-aminobutyric acid, and inosine. NAA serves as a marker for neuronal density and survival, with reductions indicating neuronal damage. Variations in Cho levels signify membrane transition disorders. Cr, being a stable marker of brain energy metabolism, is utilized as a reference for assessing NAA and Cho metabolites [20]. NAA and Cho serve as primary markers for examining fundamental metabolic neuronal processes in MDD [21]. Cho serves as a precursor for phospholipid synthesis, aids in cholesterol transport, and functions as an endogenous compound for acetylcholine release, all of which are crucial for maintaining cell membrane integrity and myelination [22]. 1H-MRS research indicates that adolescents with depression and non-suicidal self-injury may experience thalamic executive dysfunction and altered NAA and Cho metabolism [21]. A significant association was found between improvement in depressive symptoms and a decrease in the Cho/NAA ratio mediated by NAA. Neuroimaging studies have identified that various brain regions and networks are involved in depression. In conclusion, 1H-MRS could be a promising method for investigating the mechanisms underlying impaired feeding function in depression patients. Brain biochemical metabolism has yet to be utilized to explore the mechanisms underlying depression linked to eating abnormalities.
To address these concerns, we assessed MDD with and without EDs Thyroxine in patients with MDD and HCs, measuring neurometabolism. Specifically, 3.0 Tesla (T)1H-MRS was used to explore the changes of biochemical metabolites in the thalamus, PFC, ACC, and cerebellum between a group of drug-free MDD patients with and without ED. We detected and quantified the ratios of NAA, Cho, and Cr in vivo. We hypothesize that abnormalities in THs and neuronal metabolism of PFC, LN, and thalamus in depressed patients may be associated with EDs in MDD, and hopefully find relevant neurobiological evidence for MDD with ED.
Materials and methods
Participants
This study employed a cross-sectional design. We prospectively enrolled 109 patients diagnosed with MDD from the psychiatry department of the First Affiliated Hospital of Jinan University in Guangzhou, China, between January 2019 and December 2023. In addition, 37 right-handed HCs matched by age, gender, and education were recruited through local advertisements during the same period. Two experienced psychiatrists diagnosed each patient using the Structured Clinical Interview for DSM-5 (SCID-P). The clinical state was evaluated using the 24-item Hamilton Depression Rating Scale (HDRS) and the Young Mania Rating Scale (YMRS). The SCOFF questionnaire was used to screen for EDs [23]. The SCOFF questionnaire consists of 5 questions, each corresponding to a core feature of eating disorders: (1) Do you make yourself Sick because you feel uncomfortably full? (2) Do you worry you have lost Control over how much you eat? (3) Have you recently lost more than one stone (6.35 kg) in a 3-month period? (4) Do you believe yourself to be Fat when others say you are too thin? (5) Would you say that Food dominates your life? The SCOFF questionnaire uses a simple dichotomous scoring system: each “yes” answer scores 1 point, and each “no” answer scores 0 points. A total score of ≥ 2 suggests a probable case of anorexia nervosa or bulimia nervosa, warranting further clinical evaluation. The inclusion criteria were: (1) age 18–45 years to minimize the impact of vascular disease and aging; (2) diagnosis of MDD according to DSM-5 criteria; and (3) a HDRS-24 total score of ≥ 20 and a YMRS total score of < 7. Exclusion criteria included subjects using psychotropic drugs, receiving psychotherapy or related physical therapy prior to enrollment, having neurological illnesses or severe diseases, experiencing head injuries or other DSM-5 Axis I psychiatric disorders, being pregnant or having contraindications for MRI scans, and having current or past thyroid disease or treatment, including anti-thyroid drug therapy or 131I treatment. All patients met the inclusion criteria.
The criteria for “MDD with ED” required that depressive symptoms be the primary symptom, fulfilling the DSM-5 diagnostic criteria for MDD, with ED symptoms as recent accompanying symptoms. ED symptoms were evaluated using SCOFF. The score is composed of five dichotomous questions. Each “yes” response is awarded one point. A positive SCOFF score is indicated by two or more affirmative responses. Patients with MDD were categorized into two groups according to the ED item scores from the HDRS-24 and SCOFF. Among all the participants, 28 patients with ED items score ≥ 2 in SCOFF were divided into the MDD with ED group (MDD with ED symptoms, n = 28).In addition, 81 patients with ED items scores < 2 were divided into the MDD without ED group (MDD without ED symptoms, n = 81).
Thirty-seven right-handed volunteers, matched by age, gender, and education, were recruited through local advertisements to serve as healthy controls (HCs). Two experienced psychiatrists conducted diagnostic interviews using the DSM-5 to ensure health controls had no current or past psychiatric illnesses or substance abuse/dependence, either personally or in first-degree relatives. Health controls exhibited a total HDRS-24 score of less than 7 and a SCOFF score below 2. All HCs met the inclusion criteria, and none were excluded. All participants provided signed informed consent. This study was approved by the Medical Research Ethics Committee of the First Affiliated Hospital of Jinan University, China.
Procedure
Participants underwent neuropsychological assessment and TH detection within 48 h of initial contact. Following neuropsychological evaluation, blood sampling, and MRI scanning, patients received suitable treatment based on their condition and preferences.
Serum TH level detections
Blood samples were collected from all participants through a routine venipuncture (5 ml) between 7 and 8 am before initiating any treatment. The concentrations of FT3, TT3, FT4, total thyroxine (TT4), and TSH were measured using the direct chemiluminescence method. All samples from an individual were analyzed using the same laboratory procedure. The normal serum TH levels are: FT3 (3.09–7.42 pmol/L), TT3 (1.01–2.48 nmol/L), FT4 (7.64–16.03 pmol/L), TT4 (69.97–152.52 nmol/L), and TSH (0.49–4.91 mIU/L).
Image acquisition and preprocessing
Participants were positioned supine with ear foam pads and plugs to reduce noise and minimize head motion. An experienced radiologist, unaware of the participant’s group assignment, conducted both MRI and 1H-MRS.MRI and 1H-MRS data were acquired using a 3.0 T GE Discovery MR 750 system with an 8-channel phased array head coil (GE Healthcare, Milwaukee, WI).
First, routine axial T1-weighted fluid attenuation inversion recovery (T1 Flair) (repetition time [TR] = 1750 ms, echo-time [TE] = 24 ms, numbers of excitation [NEX] = 1, field of view [FOV] = 240 × 240 mm, spatial matrix = 320 × 256, slice thickness = 5 mm, gap = 1.5 mm, acquisition time = 1 min and 22 s) and fast spin echo T2 -weighted MR images (TR = 8400 ms, TE = 145 ms, NEX = 1, FOV = 240 × 240 mm, spatial matrix = 256 × 256, slice thickness = 5 mm, gap = 1.5 mm, acquisition time = 2 min and 15 s) were obtained to confirm the absence of any brain tumors, vascular disease, and other cerebral illnesses.
Then, spectra were acquired by using the two-dimensional (2D) multivoxel technique to evaluate brain metabolites. We used the axial T2-weighted MR images for anatomic localization (TR = 3500 ms, TE = 102 ms, NEX = 2, FOV = 240 × 240 mm, spatial matrix = 256 × 256, slice thickness = 5 mm without a gap, acquisition time = 1 min and 45 s). For 1H-MRS studies, the volume of interest (VOI) was focused on the anterior cingulate cortex (ACC), as shown in Fig. 1. The VOI, including 49 nominal voxels (7.5 × 7.5 × 10 mm3), was placed in a uniform manner by the same investigator. The total volume of the VOI was approximately 27.6 cm3, comprising 49 nominal voxels, each with a size of 7.5 × 7.5 × 10 mm3. A single section of 2D multi-voxel 1H-MRS was then acquired by using a point-resolved spectroscopy sequence (PRESS) with water suppression via chemical shift selective saturation (CHESS) pulse (Jia et al., 2015). The following acquisition parameters were applied: TR = 1000 ms; TE = 144 ms; NEX = 1; FOV = 180 × 180 mm; spatial matrix = 240 × 240; slice thickness = 10 mm; a nominal voxel size = 7.5 × 7.5 × 10 mm3. Additional saturation bands were placed outside the VOI to minimize lipid contamination from the scalp. To achieve an optimal full-width half maximum of 10 Hz, automatic pre-scanning was performed before each spectroscopic scan. General quality standard spectra with a line width > 10 Hz or water suppression < 98% efficiency were excluded. Each detection time was 5 min and 28 s. The total acquisition time of each participant for the 1H-MRS sequence was 16 min and 24 s. Voxels were repositioned in the bilateral ACC. The analysis of the spectral dataset was performed with the function in GE Advantage Workstation AW4.2_07. The ratio of NAA/Cr and Cho/Cr in bilateral ACC was used to analyze brain biochemical changes. Regions of interest (ROIs) were identified in ACC (Fig. 1A); PWM and thalamus (Fig. 1B), and cerebellum (Fig. 1C) by a skilled radiologist, who relied on anatomical landmarks for accurate localization. During the post-processing phase, spectral data analysis was conducted using GE's Advantage Workstation ADW4.2 Func Tool. By integrating distribution maps of various metabolites with MRI images, 1H-MRS curves were derived. The RI scanner’s built-in software automatically handled baseline calibration, signal averaging, and metabolite identification. Biochemical peak areas of NAA, Cho, and Cr (Fig. 1D) in the ACC, PWM, thalamus, and cerebellum were measured, and the ratios of NAA/Cr and Cho/Cr were calculated. All voxel placements and data analyses were performed by trained radiologists who were blinded to the participants' diagnoses, ensuring uniformity in sampling and analysis procedures.
Fig. 1.
Magnetic resonance localization imaging. A ACC, B PWM&thalamus, C Cerebellum, D Biochemical peak areas of NAA, Cho, and Cr. NAA/Cr: N-acetylaspartate to creatine, Cho/Cr: Choline-containing compounds to creatine
Statistical analysis
Data analysis was conducted using SPSS for Windows, version 26.0 (Chicago, IL, USA). Categorical data are presented as frequencies, while continuous data are expressed as means with standard deviations. The Shapiro–Wilk test assessed the normal distribution of all indicators, including demographic, clinical characteristics, THs, and neurometabolic data. Levene’s test was used to evaluate the equality of error variances. A two-tailed significance level of p < 0.05 was established. Comparisons of the demographic and clinical characteristics among three groups (MDD with ED, MDD without ED, and HCs) were conducted using the Kruskal–Wallis test analysis for data with non-normal distribution for the mean values. The Mann–Whitney U test was employed to compare clinical characteristics, including age of onset, number of episodes, time since the last depressive episode, and illness duration, between the two patient groups due to the non-normal distribution of mean values. A chi-square test was conducted to assess gender differences across the three groups. A one-way ANOVA or Kruskal–Wallis test was employed to compare the THs and neurometabolic ratios (NAA/Cr and Cho/Cr) of the ACC across the three groups, with Bonferroni post hoc tests applied for significant differences. Pearson’s correlation was applied to normal variables, and Spearman’s correlation was used for skewed variables to assess the relationship between THs and the neurometabolic ratios (NAA/Cr and Cho/Cr) in the ACC. Receiver operating characteristic (ROC) analysis was conducted to evaluate the discriminatory ability of thyroid hormones and neurometabolites for distinguishing MDD patients with ED from those without ED. For thyroid hormones, a combined marker was generated using binary logistic regression, with the predicted probability serving as the composite indicator. For neurometabolites, the NAA/Cr ratio was used as the marker. A combined marker integrating both thyroid hormone and neurometabolite measures was then constructed using the same logistic regression approach. The area under the curve (AUC) was calculated for each marker to assess its discriminatory power, with significance tested against the null hypothesis of AUC = 0.5.
Results
Demographic and clinical data
Table 1 presents the demographic and clinical characteristics of the study participants. There were 28 MDD patients with ED symptoms (MDD with ED group), 81 MDD patients without ED symptoms (MDD without ED group), and 37 healthy subjects (HCs). The three groups showed no significant differences in sex, age, or years of education (all p > 0.05). Also, no significant differences were found for the age of onset, the number of episodes, the time period since the last depressive episode, and the duration of illness between the two patient groups (all p > 0.10). Significant main group effects were observed for the HDRS-24, Y-MRS, and SCOFF scores. Post-hoc Bonferroni correction indicated higher HDRS-24 score, and Y-MRS score in MDD with ED group and MDD without ED group than in HCs (all p < 0.005). In addition, the MDD with ED group showed a higher SCOFF score than the MDD without ED group and HCs.
Table 1.
Demographic and clinical data of subjects (mean ± SD/median [interquartile range])
| MDD with ED (n = 28) | MDD without ED (n = 81) | HC (n = 37) | Z/F/χ2 | p | p value e | |
|---|---|---|---|---|---|---|
| Age (year) | 24.14 ± 6.81 | 24.81 ± 5.86 | 24.73 ± 6.341 | 0.126 | 0.882 a | – |
| Sex (male/female) | 8/20 | 24/57 | 14/23 | 0.931 | 0.628 b | – |
| Years education (year) | 14.82 ± 1.91 | 14.35 ± 2.23 | 15.00 ± 2.44 | 1.261 | 0.286 c | – |
| Age of onset (year) | 21.36 ± 8.18 | 21.49 ± 6.85 | – | − 0.275 | 0.784 d | – |
| Total duration (month) | 35.54 ± 44.62 | 33.80 ± 41.84 | – | − 0.129 | 0.898 d | – |
| Duration of illness (month) | 12.86 ± 25.19 | 14.01 ± 27.58 | – | − 0.281 | 0.779 d | – |
| Number of episodes | 1.89 ± 1.07 | 1.70 ± 0.98 | – | − 0.957 | 0.339 d | – |
| 24-item HDRS score | 27.54 ± 5.26 | 26.30 ± 3.71 | 1.97 ± 2.46 | 581.361 | < 0.001 a,* | > HC MDD with ED, MDD without ED e |
| YMRS score | 1.18 ± 1.44 | 1.36 ± 1.69 | 0.38 ± 0.89 | 5.645 | 0.004 a,* | MDD with ED, MDD without ED > HC e |
| SCOFF | 2.32 ± 0.612 | 0.00 ± 0.00 | 0.00 ± 0.00 | − 10.305 | < 0.001 a,* | MDD with ED > HC e |
MDD Major depressive disorder, ED Eating disorder, HC: Healthy control; HDRS-24: 24-item Hamilton Depression Rating Scale; YMRS: Young Mania Rating Scale;
a One-way ANOVA analyses;
b X2 test;
c Kruskal–Wallis test;
d Mann–Whitney U test;
e Bonferroni post hoc test;
* p < 0.05 significant
Comparisons of THs between the MDD and HC groups
Table 2 shows the comparisons of serum levels of FT3, TT3, FT4, TT4, and TSH between the MDD with ED group, the MDD without ED group, and HCs. FT3 levels were lower in both MDD with ED and MDD without ED compared to HC (p = 0.016). The level of FT4 in patients with MDD without ED was lower than that in patients with HC (p = 0.035). TT4 levels were significantly higher in both MDD with ED and MDD without ED compared to HC (p < 0.001), whereas no significant differences were found in other TH levels between the MDD and HC groups after Bonferroni correction (p = 0.949, p = 0.642). It should be noted that all subjects included in this study had euthyroid function, and although group differences were mentioned above, serum THs levels of all participants were still within the normal range.
Table 2.
Comparison of sex hormones between MDD with ED, MDD without ED and HC group (mean ± SD/median [interquartile range])
| MDD with ED (n = 28) | MDD without ED (n = 81) | HC (n = 37) | Z/F/χ2 | pa | p value b | |
|---|---|---|---|---|---|---|
| FT3 (pmol/L) | 5.31 ± 0.70 | 5.30 ± 0.73 | 5.68 ± 0.55 | 4.244 | 0.016 a,* | MDD with ED, MDD without ED < HC b |
| TT3 (nmol/L) | 1.71 ± 0.32 | 1.73 ± 0.29 | 1.73 ± 0.25 | 0.053 | 0.949 a | |
| FT4 (pmol/L) | 11.31 ± 1.39 | 11.64 ± 1.65 | 10.85 ± 1.35 | 3.443 | 0.035 a,* | MDD without ED > HC b |
| TT4 (nmol/L) | 114.26 ± 23.47 | 108.93 ± 20.35 | 94.20 ± 16.81 | 9.524 | < 0.001 a,*** | MDD with ED, MDD without ED > HC b |
| TSH (mIU/L) | 1.91 ± 1.19 | 2.06 ± 1.40 | 2.23 ± 1.43 | 0.444 | 0.642 a |
MDD Major depressive disorder, ED Eating disorder, HC Healthy control, FT3 Free tri-iodothyronine, TT3 Total tri-iodothyronine, FT4 Free thyroxine, TT4 Total thyroxine, TSH Thyroid-stimulating hormone
aOne-way ANOVA analyses;
bBonferroni post hoc test
*p < 0.05; **p < 0.005; ***p < 0.001
Comparative analysis of neurometabolic factors among depressive patients with and without EDs and HCs
Table 3 presents the comparison of NAA/Cr and Cho/Cr ratios in the bilateral lenticular nucleus, thalamus, prefrontal cortex, ACC, and cerebellum among MDD patients with and without EDs, as well as HCs. In individuals with MDD and ED, the NAA-to-Cr ratio was lower in the right thalamus (F = 3.158, p = 0.046) compared to HCs, while the NAA-to-Cr ratio was higher in the left cerebellum (F = 5.172, p = 0.007) in individuals with MDD, regardless of ED status, compared to HCs.
Table 3.
Comparisons of biochemical metabolite ratios among MDD with ED, MDD without ED and HC group (mean ± SD)
| MDD with ED (n = 28) | MDD without ED (n = 81) | HC (n = 37) | Z/F/χ2 | pa | p value b | |
|---|---|---|---|---|---|---|
| Left lenticular nucleus | ||||||
| NAA/Cr | 1.46 ± 0.56 | 1.41 ± 0.31 | 1.49 ± 0.33 | 0.562 | 0.571a | |
| Cho/Cr | 1.19 ± 1.38 | 0.93 ± 0.37 | 0.87 ± 0.42 | 1.964 | 0.144 a | |
| Right lenticular nucleus | ||||||
| NAA/Cr | 1.50 ± 0.28 | 1.56 ± 0.38 | 1.56 ± 0.32 | 0.365 | 0.695 a | |
| Cho/Cr | 0.87 ± 0.35 | 0.95 ± 0.50 | 0.91 ± 0.65 | 0.235 | 0.791 a | |
| Left thalamus | ||||||
| NAA/Cr | 2.04 ± 0.40 | 2.13 ± 0.35 | 2.06 ± 0.34 | 0.754 | 0.472 a | |
| Cho/Cr | 0.91 ± 0.18 | 0.88 ± 0.19 | 0.91 ± 0.20 | 0.390 | 0.678 a | |
| Right thalamus | ||||||
| NAA/Cr | 1.91 ± 0.40 | 2.02 ± 0.31 | 2.11 ± 0.26 | 3.158 | 0.046 a,* | MDD with ED < HC b,* |
| Cho/Cr | 0.82 ± 0.23 | 0.90 ± 0.17 | 0.87 ± 0.14 | 2.308 | 0.103 a | |
| Left PFC | ||||||
| NAA/Cr | 2.19 ± 0.28 | 2.16 ± 0.46 | 2.13 ± 0.32 | 0.164 | 0.849 a | |
| Cho/Cr | 1.20 ± 0.19 | 1.14 ± 0.22 | 1.15 ± 0.17 | 0.804 | 0.449 a | |
| Right PFC | ||||||
| NAA/Cr | 2.17 ± 0.42 | 2.23 ± 0.39 | 2.20 ± 0.31 | 0.172 | 0.842 a | |
| Cho/Cr | 1.10 ± 0.20 | 1.15 ± 0.21 | 1.15 ± 0.20 | 0.748 | 0.475 a | |
| Left ACC | ||||||
| NAA/Cr | 1.75 ± 0.28 | 1.83 ± 0.34 | 1.81 ± 0.21 | 0.970 | 0.382 a | |
| Cho/Cr | 1.10 ± 0.16 | 1.12 ± 0.19 | 1.07 ± 0.16 | 1.117 | 0.330 a | |
| Right ACC | ||||||
| NAA/Cr | 1.75 ± 0.30 | 1.83 ± 0.31 | 1.84 ± 0.19 | 1.086 | 0.340 a | |
| Cho/Cr | 1.06 ± 0.15 | 1.11 ± 0.20 | 1.12 ± 0.17 | 1.138 | 0.324 a | |
| Left cerebellum | ||||||
| NAA/Cr | 1.57 ± 0.35 | 1.54 ± 0.36 | 1.34 ± 0.27 | 5.172 | 0.007 a,** | MDD without ED, MDD with ED > HC b,* |
| Cho/Cr | 0.93 ± 0.15 | 0.98 ± 0.22 | 0.98 ± 0.21 | 0.528 | 0.591 a | |
| Right cerebellum | ||||||
| NAA/Cr | 1.51 ± 0.34 | 1.50 ± 0.36 | 1.36 ± 0.36 | 2.033 | 0.135 a | * |
| Cho/Cr | 0.93 ± 0.18 | 0.91 ± 0.18 | 0.95 ± 0.27 | 0.513 | 0.600 a | |
MDD Major depressive disorder, ED Eating disorder, HC Healthy control, PFC Prefrontal cortex, ACC Anterior cingulate cortex, NAA/Cr N-acetylaspartate to creatine, Cho/Cr Choline-containing compounds to creatine;
a One-way ANOVA analyses;
b Bonferroni post hoc test;
* p < 0.05;
** p < 0.005;
*** p < 0.001
Correlation between minor variation in THs and neurometabolic factors in MDD with ED group
Pearson’s correlation analysis was performed to examine the relationships between serum thyroid hormone levels (FT3, FT4, TT3, TT4, and TSH) and neurometabolite ratios (NAA/Cr, Cho/Cr) across all four brain regions (left cerebellum, ACC, PWM, and thalamus) in patients with MDD and ED. The complete correlation matrix is presented in Table 4. As shown in Fig. 2, significant positive correlations were observed only in the left cerebellum: FT3 levels were positively correlated with NAA/Cr ratios (r = 0.432, p = 0.022; Fig. 2A), and TT4 levels were positively correlated with NAA/Cr ratios (r = 0.394, p = 0.038; Fig. 2B). No other significant correlations were found between any thyroid hormone and neurometabolite ratios in the left cerebellum or any other brain region (all p > 0.05). These region-specific findings suggest a selective association between peripheral thyroid hormones and cerebellar neurometabolism in MDD patients with comorbid ED symptoms.
Table 4.
Correlation between thyroid hormone and brain biochemical metabolism in depressed patients with ED
| FT3 | TT3 | FT4 | TT4 | TSH | |
|---|---|---|---|---|---|
| Left lenticular nucleus | |||||
| NAA/Cr |
r = -0.172 p = 0.0380 |
r = 0.073 p = 0.712 |
r = -0.088 p = 0.655 |
r = 0.063 p = 0.751 |
r = 0.349 p = 0.069 |
| Cho/Cr |
r = -0.088 p = 0.657 |
r = 0.148 p = 0.453 |
r = -0.253 p = 0.193 |
r = -0.115 p = 0.560 |
r = 0.530** p = 0.004 |
| Right lenticular nucleus | |||||
| NAA/Cr |
r = -0.060 p = 0.763 |
r = -0.177 p = 0.368 |
r = -0.012 p = 0.951 |
r = 0.173 p = 0.378 |
r = -0.116 p = 0.557 |
| Cho/Cr |
r = -0.123 p = 0.535 |
r = -0.223 p = 0.254 |
r = 0.113 p = 0.467 |
r = -0.033 p = 0.869 |
r = 0.154 p = 0.435 |
| Left thalamus | |||||
| NAA/Cr |
r = 0.259 p = 0.183 |
r = 0.170 p = 0.387 |
r = 0.086 p = 0.662 |
r = 0.217 p = 0.267 |
r = 0.144 p = 0.464 |
| Cho/Cr |
r = 0.289 p = 0.136 |
r = 0.289 p = 0.136 |
r = 0.015 p = 0.941 |
r = 0.234 p = 0.232 |
r = 0.040 p = 0.839 |
| Right thalamus | |||||
| NAA/Cr |
r = -0.041 p = 0.836 |
r = -0.023 p = 0.908 |
r = -0.059 p = 0.767 |
r = -0.135 p = 0.493 |
r = 0.437* p = 0.020 |
| Cho/Cr |
r = -0.093 p = 0.637 |
r = -0.207 p = 0.289 |
r = 0.049 p = 0.803 |
r = -0.306 p = 0.114 |
r = 0.335 p = 0.081 |
| Left PFC | |||||
| NAA/Cr |
r = 0.251 p = 0.198 |
r = 0.204 p = 0.298 |
r = 0.101 p = 0.608 |
r = -0.209 p = 0.286 |
r = 0.053 p = 0.790 |
| Cho/Cr |
r = 0.167 p = 0.395 |
r = -0.084 p = 0.670 |
r = 0.170 p = 0.388 |
r = -0.095 p = 0.632 |
r = 0.022 p = 0.913 |
| Right PFC | |||||
| NAA/Cr |
r = -0.064 p = 0.746 |
r = 0.308 p = 0.111 |
r = -0.426* p = 0.024 |
r = -0.352 p = 0.066 |
r = -0.022 p = 0.912 |
| Cho/Cr |
r = -0.212 p = 0.280 |
r = -0.027 p = 0.892 |
r = -0.180 p = 0.359 |
r = -0.284 p = 0.143 |
r = -0.028 p = 0.888 |
| Left ACC | |||||
| NAA/Cr |
r = 0.099 p = 0.616 |
r = 0.079 p = 0.691 |
r = 0.185 p = 0.346 |
r = 0.121 p = 0.540 |
r = -0.338 p = 0.079 |
| Cho/Cr |
r = 0.138 p = 0.484 |
r = 0.176 p = 0.369 |
r = 0.333 p = 0.083 |
r = 0.223 p = 0.253 |
r = 0.102 p = 0.605 |
| Right ACC | |||||
| NAA/Cr |
r = 0.216 p = 0.269 |
r = 0.283 p = 0.144 |
r = 0.203 p = 0.301 |
r = 0.195 p = 0.321 |
r = -0.029 p = 0.884 |
| Cho/Cr |
r = 0.011 p = 0.954 |
r = 0.194 p = 0.324 |
r = 0.096 p = 0.627 |
r = 0.094 p = 0.633 |
r = -0.032 p = 0.871 |
| Left cerebellum | |||||
| NAA/Cr |
r = 0.432* p = 0.022 |
r = 0.146 p = 0.458 |
r = 0.257 p = 0.186 |
r = 0.394* p = 0.038 |
r = 0.180 p = 0.361 |
| Cho/Cr |
r = -0.049 p = 0.804 |
r = 0.023 p = 0.908 |
r = -0.124 p = 0.529 |
r = 0.048 p = 0.809 |
r = 0.143 p = 0.469 |
| Right cerebellum | |||||
| NAA/Cr |
r = 0.315 p = 0.102 |
r = 0.378* p = 0.048 |
r = 0.050 p = 0.801 |
r = 0.285 p = 0.141 |
r = 0.120 p = 0.541 |
| Cho/Cr |
r = 0.066 p = 0.739 |
r = 0.034 p = 0.863 |
r = -0.198 p = 0.314 |
r = -0.065 p = 0.743 |
r = 0.523** p = 0.004 |
FT3 Free tri-iodothyronine, TT3 Total tri-iodothyronine, FT4 Free thyroxine, TT4 Total thyroxine, TSH Thyroid-stimulating hormone, PFC Prefrontal cortex, ACC Anterior cingulate cortex, NAA/Cr N-acetylaspartate to creatine, Cho/Cr Choline-containing compounds to creatine; * p < 0.05; ** p < 0.005; *** p < 0.001
Fig. 2.
Correlation between FT3 levels, TT4 levels, and left cerebellar NAA/Cr in depressed patients with ED. A Correlation between FT3 levels and NAA/Cr in the left cerebellum in MDD with eating disorders; FT3: Free tri-iodothyronine; NAA/Cr: N-acetylaspartate to creatine. B Correlation between TT4 levels and NAA/Cr in the left cerebellum in MDD with eating disorders; TT4: Total thyroxine; NAA/Cr: N-acetylaspartate to creatine
ROC analysis of THs and neurometabolites for distinguishing MDD from ED
Based on the finding of significant differences in serum THs levels and neurometabolic ratios between the MDD with ED and MDD without ED groups, we hypothesized that these peripheral and central measures might serve as sensitive biomarkers for MDD with ED. To test this, we performed ROC analysis to evaluate their ability to distinguish MDD patients with ED from those without ED. The combination of FT3 and TT4 yielded an AUC of 0.640 (p = 0.027) (Fig. 3A). According to Fig. 3B, the AUC of left cerebellar NAA/Cr in distinguishing MDD with or without ED was 0.798 (p < 0.001). According to Fig. 3C, the AUC of the combination of FT3 and TT4 and left cerebellar NAA/Cr in distinguishing MDD with or without ED was 0.830 (p < 0.001). These findings suggest that THs and neurometabolites may serve as promising biomarkers to aid in the identification and diagnosis of MDD patients with comorbid ED symptoms.
Fig. 3.
ROC analysis of thyroid hormone and brain biochemical metabolism in depressed patients with eating disorders. A ROC analysis of diagnostic performance of the combination of FT3 and TT4 for depression with eating disorder. Note: ROC: Receiver operating characteristic; FT3: Free tri-iodothyronine; TT4: Total thyroxine; NAA/Cr: N-acetylaspartate to creatine; AUC: Area under the ROC curve; B ROC analysis of diagnostic performance of left cerebellar NAA/Cr for depression with eating disorder. Note: ROC: Receiver operating characteristic; AUC: Area under the ROC curve; C ROC analysis of diagnostic performance of the combination of FT3 and TT4 and left cerebellar NAA/Cr for depression with eating disorder. Note: ROC: Receiver operating characteristic; FT3: Free tri-iodothyronine; TT4: Total thyroxine; NAA/Cr: N-acetylaspartate to creatine; AUC: Area under the ROC curve
Discussion
The purpose of our study was to compare differences in TH levels among patients with MDD with or without ED and HCs, and to explore the potential effects of small variations in THs in patients with depression with eating abnormalities. We observed minor variations in FT3, FT4, and TT4 levels within the normal range in cases of depression accompanied by abnormal eating. Furthermore, THs can serve as a promising biomarker for the diagnosis of patients with depression accompanied by eating abnormalities. In MDD with ED, NAA/Cr was lower in the right thalamus compared to HCs, while the NAA-to-Cr ratio was higher in the left cerebellum.
Our study found that although FT3, FT4, and TT4 levels were still within the normal range in depression with eating abnormalities, FT3 levels were lower than those of HCs, and FT4 and TT4 levels were higher than those of HCs. Why do such tiny THs variants occur in depression accompanied by eating abnormalities? The problem can be related to two important factors: depression and changed appetite. A comparative study indicates that thyroid dysfunction is more prevalent in mood disorder patients. Elevated serum FT3 levels and thyroid comorbidities heighten the risk of readmission for MDD, implying that TH fluctuations are common and potentially linked to depression onset [24]. Elevated FT3 levels correlate with an increased risk of readmission following MDD hospitalization [24]. Early-onset affective disorder patients have a 2.097 times higher likelihood of developing hypothyroidism and exhibit a 12.7% greater prevalence rate compared to late-onset patients [25]. Subclinical hypothyroidism increases the risk of depression with psychotic symptoms ninefold [26]. Depressive symptoms were found to be significantly inversely associated with FT4, FT3, and TT3 in a survey of depression in 12,502 Americans [27]. From the above studies, it can be inferred that the alteration of TH levels is common in depression, and the severity of depression is closely related to the alteration of TH levels, which may not reach obvious or subclinical dysfunction, but only manifest as fluctuations within the normal range.
On the other hand, altered appetite appears to be associated with altered TH secretion. Patients with acute anorexia nervosa had lower levels of FT3, FT4, and TSH, and a decrease in FT3 was associated with the severity of depressive symptoms, and an increase in the FT3/FT4 ratio was associated with the recovery of depressive symptoms with weight regain [8]. The central nervous system might modulate D2 and D3 expression to precisely adjust hypothalamic TH levels, thus influencing appetite regulation [28].
Therefore, in summary, the decrease in FT3 may be related to the decrease in THs secretion and negative feedback inhibition induced by abnormal feeding. Furthermore, changes in THs were more pronounced in patients with depression with abnormal eating than in HCs. The findings indicate that these could influence MDD patients, potentially causing EDs, while axis functional changes may regulate depressive moods in those with eating abnormalities. Although the relationship between THs and appetite changes is not fully understood, these changes may indicate a possible compensatory mechanism.
Altered brain biochemistry and metabolism may contribute to abnormal eating behaviors in patients with MDD. The data indicate a reduction in the NAA/Cr ratio in the right thalamus of MDD patients. The neurobiological model of MDD highlights the parathalamic nucleus (PSTN), a small nucleus at the posterolateral margin of the hypothalamus, as a key region involved in appetite regulation and the modulation of food intake and anorexia signals [29]. Barbier et al proposed a regulatory role for the insula–PSTN network in ingestion behavior, and PSTN activity signals fasting in sick or dangerous situations [30]. PSTN neurons are attuned to the hedonic aspects of visceral, gustatory, and emotional experiences, reacting to arousal and distress stimuli. In patients with bulimia nervosa, there is a disruption of functional connectivity between resting-state hypothalamic subregions and frontoparietal networks, default-mode networks, somatosensory, and visual networks [31]. Meanwhile, Lv et al The study revealed that adolescent MDD patients exhibited elevated NAA/Cr ratios in the right thalamus and increased Cho/Cr ratios in the right putamen compared to HCs. In the right thalamus, the NAA/Cr ratio showed a negative correlation with happiness response time and a positive correlation with the intensity of sadness, anger, and fear. In facial emotion processing tasks, the MDD group demonstrated notably lower accuracy and perceived intensity for happiness, while exhibiting higher accuracy and perceived intensity for aversion, sadness, and fear [15]. Executive dysfunction may be related to abnormal NAA metabolism in the left thalamus and ACC in adolescents with MDD with non-self-injurious suicide [21]. In conclusion, thalamic activity is involved in encoding aversive memories, driving avoidance behavior, and modulating impulsive versus rewarding actions, such as eating behavior [32]. The thalamus showed greater activity in processing negative expressions, associated with anger and fear regulation.
The study revealed a significant increase in the NAA/Cr ratio in the left cerebellum of the MDD group. The cerebellum potentially plays a crucial role in behavioral and associative modifications linked to the cognitive regulation of appetite. Cerebellar volume was directly associated with inadequate diet and overeating situations, and cerebellar atrophy and cell loss were associated with poor duration and treatment success of AN [33]. Schmahmann et al. determined the association between the posterior cerebellum and cognition, and proposed the theory of distance disorder of thinking (DoT) [34]. This theory proposes that the cerebellum refines cognitive processes by integrating multimodal inputs and optimizing behavioral adjustments, akin to its role in enhancing motor coordination. Within the DoT framework, the cerebellum may similarly influence intake and appetite-related stimuli by developing rigid, repetitive, or inflexible personal models, memories, and expectations based on previous food interactions. Stimulation-salient cerebellar regions (e.g., Lobule VI may be linked to appetite changes and EDs through a negative association with food. In the MRI Popper study of patients with MDD, increased glutamate in the cerebellum was observed, with the highest connectivity between the medial cerebellum and the ventral tegmental area (VTA). Cerebellar projections to the VTA were associated with appetite behavior [35]. Our findings, using 1H-MRS, suggest that depressed patients with eating abnormalities may experience impaired neuronal viability and integrity in the thalamus and cerebellum, which could disrupt regional brain function and connectivity, contributing to eating abnormalities. Research on abnormal eating behaviors in MDD patients is scarce and may be linked to the cerebellar reward loop. In the future, the sample size will continue to be expanded to further study its mechanism.
FT3 and TT4 levels were associated with the NAA/Cr ratio in the left cerebellum of MDD patients exhibiting eating abnormalities. This association was absent in MDD patients without eating abnormalities and in HCs. Structural and functional abnormalities in affective and cognitive control networks contribute to emotional and cognitive dissonance in MDD. Hypothyroidism patients exhibit reduced gray matter volume in the bilateral cerebellum compared to normal controls, indicating that TH deficiency may cause structural brain abnormalities and cerebellar gray matter atrophy. The human cerebellum may play a role in regulating appetite and controlling responses to external food cues through the prefrontal-cerebellar pathway [36]. We also found that TH levels and neurometabolic factors are potential predictors in patients with depression with eating abnormalities, with an AUC of 0.83. Reliable predictors are valuable for patient selection and treatment. As mentioned earlier, these brain regions, including the thalamus and cerebellum, are involved in appetite regulation and inhibitory control and may be involved in thyroxine dysfunction.
A major strength of our study is that we specifically focused on patients with MDD with accompanying eating abnormalities and observed small changes in their THs within the normal range. Our results fill the research gaps of changes in THs in MDD patients with abnormal eating and the effects of small variations in THs within the normal range on depression and appetite, and explore the characteristics of neurometabolic changes in their brains. Despite the promising results of our study, some limitations must be noted. A major limitation is our small sample size. Therefore, our study may not be sufficient to detect small effects, and some findings may have been missed. Secondly, we did not perform a comparative analysis of MDD patients of different genders and subtypes, because the effects of THs on emotion and cognition may be influenced by age, gender, and different subtypes. Finally, we did not deeply explore the mechanism of THs in the normal range of mood and appetite in patients with depression accompanied by eating abnormalities. Future studies with larger samples at multiple ages are needed to explore and elucidate the effects and mechanisms of THs on emotional and brain neurometabolic alterations throughout the life cycle of MDD.
Strengths and limitations
This study has several strengths, including the simultaneous assessment of thyroid function and region-specific neurometabolites using 1H-MRS in MDD patients with ED, and the integration of endocrine and neuroimaging markers to explore the thalamic-cerebellar circuitry. However, several limitations should be acknowledged. First, structural MRI data were acquired using conventional 2D sequences with a relatively thick slice thickness (5 mm), primarily for clinical screening and MRS voxel placement; high-resolution 3D T1-weighted sequences were not included. This limits our ability to perform quantitative morphometric analyses and to investigate potential relationships between metabolic alterations and subtle structural brain changes. Future studies employing 3D T1-weighted sequences are warranted to integrate metabolic and structural data. Second, the sample size was relatively modest, particularly in the MDD with ED group (n = 28), which may reduce statistical power and limit generalizability. Third, the cross-sectional design precludes any causal or temporal inferences regarding the relationship between thyroid dysfunction, neurometabolic disturbances, and eating behaviors. The severity of ED symptoms and the timing of depressive episodes were also not strictly operationalized, further limiting causal interpretations. These limitations collectively indicate that the present findings should be viewed as preliminary and require confirmation in larger, prospective, and more rigorously phenotyped cohorts.
Level of evidence
This study is classified as Level IV evidence (cross-sectional case–control study), according to the Oxford Centre for Evidence-Based Medicine (OCEBM) Levels of Evidence [37].
Conclusions
Taken together, our findings suggest that depressed patients with ED may exhibit concurrent thyroid hormone irregularities and NAA/Cr metabolic abnormalities in the thalamus and cerebellum. FT4 levels and the right thalamic NAA/Cr ratio appear to distinguish MDD patients with and without ED, and in those with ED, FT3 and TT4 abnormalities are associated with altered left cerebellar NAA/Cr metabolism. These exploratory observations provide preliminary insights into the neurobiological underpinnings of eating disturbances in depression. However, given the cross-sectional design and limited sample size, all findings should be considered hypothesis-generating and await validation in future prospective investigations before any clinical translation can be considered.
Acknowledgements
We would like to express our gratitude to all the participants who contributed to this study. We also thank the staff at the First Affiliated Hospital of Jinan University for their support and assistance in data collection and management. Special thanks go to Professor Yanbin Jia and Dr. Shuming Zhong for their valuable contributions and technical support.
Authors’ contributions
Yanbin Jia, Shuming Zhong, Yangyu Wu designed the trial. Yangyu Wu, Shuming Zhong, Jianzhao Zhang, Shunkai Lai, Ying Wang, Yiliang Zhang, Dong Huang, Guanmao Chen, Shuya Yan, Pan Chen, Xiaodan Lu, Xiaodong Song, Qilin Zhong, Rongxu Zhang carried out the statistical analyses. Yangyu Wu drafted the initial manuscript. All authors interpreted the data, revised the paper critically for important intellectual content, approved the final version, and agreed to be accountable for all aspects of the work.
Funding
This study was funded by the National Key R&D Program of China (no. 2022YFB4500600), the National Natural Science Foundation of China (no. 82271564), the Science and Technology Projects of Guangzhou (no. 2024A04J4178), and the Guangdong Basic and Applied Basic Research Foundation (2026A1515011397). The founders have not played any roles in study design, data collection, analysis, manuscript writing, and the decision to publish.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of the First Affiliated Hospital of Jinan University, Guangzhou, China (KY-2023–223); all subjects signed a written informed consent form. Not all or part of this information is published elsewhere; the manuscript has not yet been considered for publication in any other journal; all the authors have personally and actively participated in the actual work of the manuscript and will bear the common and individual responsibility for their content.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yangyu Wu, Qilin Zhong and Jianzhao Zhang these authors made equal contributions to this work.
References
- 1.Huang Y, Wang Y, Wang H, Liu Z, Yu X, Yan J, Yu Y, Kou C, Xu X, Lu J, Wang Z, He S, Xu Y, He Y, Li T, Guo W, Tian H, Xu G, Xu X, Ma Y, Wang L, Wang L, Yan Y, Wang B, Xiao S, Zhou L, Li L, Tan L, Zhang T, Ma C, Li Q, Ding H, Geng H, Jia F, Shi J, Wang S, Zhang N, Du X, Du X, Wu Y (2019) Prevalence of mental disorders in China: a cross-sectional epidemiological study. Lancet Psychiatry 6:211–224. 10.1016/S2215-0366(18)30511-X [DOI] [PubMed] [Google Scholar]
- 2.Iwajomo T, Bondy SJ, de Oliveira C, Colton P, Trottier K, Kurdyak P (2021) Excess mortality associated with eating disorders: population-based cohort study. Br J Psychiatry J Ment Sci 219:487–493. 10.1192/bjp.2020.197 [DOI] [PubMed] [Google Scholar]
- 3.Swinbourne J, Hunt C, Abbott M, Russell J, St Clare T, Touyz S (2012) The comorbidity between eating disorders and anxiety disorders: prevalence in an eating disorder sample and anxiety disorder sample. Aust N Z J Psychiatry 46:118–131. 10.1177/0004867411432071 [DOI] [PubMed] [Google Scholar]
- 4.Ward ZJ, Rodriguez P, Wright DR, Austin SB, Long MW (2019) Estimation of Eating Disorders Prevalence by Age and Associations With Mortality in a Simulated Nationally Representative US Cohort. JAMA Netw Open 2:e1912925. 10.1001/jamanetworkopen.2019.12925 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Daly M, Costigan E (2022) Trends in eating disorder risk among US college students, 2013–2021. Psychiatry Res 1(317):114882 [DOI] [PubMed] [Google Scholar]
- 6.Puccio F, Fuller-Tyszkiewicz M, Ong D, Krug I (2016) A systematic review and meta-analysis on the longitudinal relationship between eating pathology and depression. Int J Eat Disord 49:439–454. 10.1002/eat.22506 [DOI] [PubMed] [Google Scholar]
- 7.Fursland A, Watson HJ (2014) Eating disorders: A hidden phenomenon in outpatient mental health? Int J Eat Disord 47:422–425. 10.1002/eat.22205 [DOI] [PubMed] [Google Scholar]
- 8.Wronski M-L, Tam FI, Seidel M, Mirtschink P, Poitz DM, Bahnsen K, Steinhäuser JL, Bauer M, Roessner V, Ehrlich S (2022) Associations between pituitary-thyroid hormones and depressive symptoms in individuals with anorexia nervosa before and after weight-recovery. Psychoneuroendocrinology 137:105630. 10.1016/j.psyneuen.2021.105630 [DOI] [PubMed] [Google Scholar]
- 9.Val-Laillet D, Aarts E, Weber B, Ferrari M, Quaresima V, Stoeckel LE, Alonso-Alonso M, Audette M, Malbert CH, Stice E (2015) Neuroimaging and neuromodulation approaches to study eating behavior and prevent and treat eating disorders and obesity. NeuroImage Clin 8:1–31. 10.1016/j.nicl.2015.03.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Fukao A, Takamatsu J, Arishima T, Tanaka M, Kawai T, Okamoto Y, Miyauchi A, Imagawa A (2020) Graves’ disease and mental disorders. J Clin Transl Endocrinol 19:100207. 10.1016/j.jcte.2019.100207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Duval F, Mokrani M-C, Erb A, Danila V, Lopera FG, Foucher JR, Jeanjean LC (2021) Thyroid axis activity and dopamine function in depression. Psychoneuroendocrinology 128:105219. 10.1016/j.psyneuen.2021.105219 [DOI] [PubMed] [Google Scholar]
- 12.Feng G, Kang C, Yuan J, Zhang Y, Wei Y, Xu L, Zhou F, Fan X, Yang J (2019) Neuroendocrine abnormalities associated with untreated first episode patients with major depressive disorder and bipolar disorder. Psychoneuroendocrinology 107:119–123. 10.1016/j.psyneuen.2019.05.013 [DOI] [PubMed] [Google Scholar]
- 13.Calcaterra V, Magenes VC, Siccardo F, Hruby C, Basso M, Conte V, Maggioni G, Fabiano V, Russo S, Veggiotti P, Zuccotti G (2023) Thyroid dysfunction in children and adolescents affected by undernourished and overnourished eating disorders. Front Nutrition. 29(10):1205331 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mathisen TF, Sundgot-Borgen J, Rosenvinge JH, Bratland-Sanda S, Svendsen M, Pettersen G, Vrabel K, Friborg O (2023) Metabolic profile in women with bulimia nervosa or binge-eating disorder before and after treatment: secondary analysis from the randomized PED-t trial. Eat Weight Disord-Stud Anorex Bulim Obes 28:41. 10.1007/s40519-023-01567-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lv S, Zhong S, Zhang S, Lai S, Wang Y, Zhao H, Zhang Y, Luo Y, Yan S, Ran H, Chen P, Chen G, Jia Y (2022) Correlations between facial emotion processing and biochemical abnormalities in untreated adolescent patients with major depressive disorder: A proton magnetic resonance spectroscopy study. J Affect Disord 296:408–417. 10.1016/j.jad.2021.08.129 [DOI] [PubMed] [Google Scholar]
- 16.Zhong S, Su T, Gong J, Huang L, Wang Y (2023) Brain functional alterations in patients with anorexia nervosa: A meta-analysis of task-based functional MRI studies. Psychiatry Res 327:115358. 10.1016/j.psychres.2023.115358 [DOI] [PubMed] [Google Scholar]
- 17.Guzzardi MA, Garelli S, Agostini A, Filidei E, Fanelli F, Giorgetti A, Mezzullo M, Fucci S, Mazza R, Vicennati V, Iozzo P, Pagotto U (2018) Food addiction distinguishes an overweight phenotype that can be reversed by low calorie diet. Eur Eat Disord Rev J Eat Disord Assoc 26:657–670. 10.1002/erv.2652 [DOI] [PubMed] [Google Scholar]
- 18.Olivo G, Wiemerslage L, Swenne I, Zhukowsky C, Salonen-Ros H, Larsson E-M, Gaudio S, Brooks SJ, Schiöth HB (2017) Limbic-thalamo-cortical projections and reward-related circuitry integrity affects eating behavior: A longitudinal DTI study in adolescents with restrictive eating disorders. PLoS ONE 12:e0172129. 10.1371/journal.pone.0172129 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Amianto F, D’Agata F, Lavagnino L, Caroppo P, Abbate-Daga G, Righi D, Scarone S, Bergui M, Mortara P, Fassino S (2013) Intrinsic Connectivity Networks Within Cerebellum and Beyond in Eating Disorders. The Cerebellum 12:623–631. 10.1007/s12311-013-0471-1 [DOI] [PubMed] [Google Scholar]
- 20.Mao N, Fang J, Xie H, Liu X, Jiang X, Wang G, Cui M, Wang B, Liu Q (2016) Correlation between neurochemical metabolism and memory function in adolescent patients with depression: A multi-voxel 1H magnetic resonance spectroscopy study. Psychiatry Clin Neurosci 70:167–174. 10.1111/pcn.12372 [DOI] [PubMed] [Google Scholar]
- 21.Zhang Y, Lai S, Wu W, Wang Y, Zhu Y, Chen G, Qi Z, Chen P, Miao H, Yan S, Ran H (2022) Associations between executive function impairment and biochemical abnormalities in depressed adolescents with non-suicidal self-injury. J Affect Disorder. 1(298):492–499 [DOI] [PubMed] [Google Scholar]
- 22.Haley AP, Gonzales MM, Tarumi T, Tanaka H (2013) Dyslipidemia links obesity to early cerebral neurochemical alterations. Obesity 21(10):2007–2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Morgan JF, Reid F, Lacey JH (1999) The SCOFF questionnaire: assessment of a new screening tool for eating disorders. BMJ 319:1467–1468. 10.1136/bmj.319.7223.1467 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Yang L, Yang X, Yang T, Wu X, Sun P, Zhu Y, Su Y, Gu W, Qiu H, Wang J, Chen J, Fang Y (2021) The effect of thyroid function on the risk of psychiatric readmission after hospitalization for major depressive disorder. Psychiatry Res 305:114205. 10.1016/j.psychres.2021.114205 [DOI] [PubMed] [Google Scholar]
- 25.Zhao Y, Wang N, Chen J, Guan Q, Yuan X, Shen Y, Zhang X, Hao W, Zhang R, Zhai D (2022) Association between early-onset affective disorders and hypothyroidism in a larger number of psychiatric drug-free patients. J Affect Disord 299:31–36. 10.1016/j.jad.2021.11.061 [DOI] [PubMed] [Google Scholar]
- 26.Peng P, Wang Q, Ren H, Zhou Y, Hao Y, Chen S, Wu Q, Li M, Wang Y, Yang Q, Wang X, Liu Y, Ma Y, Li H, Liu T, Zhang X (2023) Association between thyroid hormones and comorbid psychotic symptoms in patients with first-episode and drug-naïve major depressive disorder. Psychiatry Res 320:115052. 10.1016/j.psychres.2023.115052 [DOI] [PubMed] [Google Scholar]
- 27.Ma Y, Wang M, Zhang Z (2024) The association between depression and thyroid function. Front Endocrinol 30(15):1454744 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Amin A, Dhillo WS, Murphy KG (2011) The Central Effects of Thyroid Hormones on Appetite. J Thyroid Res 2011:1–7. 10.4061/2011/306510 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shah T, Dunning JL, Contet C (2022) At the heart of the interoception network: Influence of the parasubthalamic nucleus on autonomic functions and motivated behaviors. Neuropharmacology 204:108906. 10.1016/j.neuropharm.2021.108906 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Barbier M, Chometton S, Pautrat A, Miguet-Alfonsi C, Datiche F, Gascuel J, Fellmann D, Peterschmitt Y, Coizet V, Risold P-Y (2020) A basal ganglia-like cortical–amygdalar–hypothalamic network mediates feeding behavior. Proc Natl Acad Sci 117:15967–15976. 10.1073/pnas.2004914117 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wang J, Wu G, Wang M, Li W, Wang Y, Ren X, Wei X, Yang Z, Li Z, Wang Z, Chen Q, Zhang P, Tang L (2023) Exploring the thalamus: a crucial hub for brain function and communication in patients with bulimia nervosa. J Eat Disord 11:207. 10.1186/s40337-023-00933-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kim AJ, Anderson BA (2020) The effect of concurrent reward on aversive information processing in the brain. Neuroimage 217:116890. 10.1016/j.neuroimage.2020.116890 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sader M, Waiter GD, Williams JHG (2023) The cerebellum plays more than one role in the dysregulation of appetite: Review of structural evidence from typical and eating disorder populations. Brain Behav 13:e3286. 10.1002/brb3.3286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Schmahmann JD, Guell X, Stoodley CJ, Halko MA (2019) The Theory and Neuroscience of Cerebellar Cognition. Annu Rev Neurosci 42:337–364. 10.1146/annurev-neuro-070918-050258 [DOI] [PubMed] [Google Scholar]
- 35.Hoffman LJ, Foley JM, Leong JK, Sullivan-Toole H, Elliott BL, Olson IR. An in vivo dissection, and analysis of socio-affective symptoms related to cerebellum-midbrain reward circuitry in humans. bioRxiv. 2024: 2023-09. [DOI] [PMC free article] [PubMed]
- 36.Marron EM, Viejo-Sobera R, Cuatrecasas G, Redolar-Ripoll D, Lorda PG, Datta A, Bikson M, Magerowski G (2005) Alonso-Alonso M (2019) Prefronto-cerebellar neuromodulation affects appetite in obesity. Int J Obes 43:2119–2124. 10.1038/s41366-018-0278-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Jeremy Howick, Iain Chalmers, Paul Glasziou, Trish Greenhalgh, Carl Heneghan, Alessandro Liberati, Ivan Moschetti, Bob Phillips, and Hazel Thornton. “The 2011 Oxford CEBM Levels of Evidence (Introductory Document)”. Oxford Centre for Evidence-Based Medicine. https://www.cebm.ox.ac.uk/resources/levels-of-evidence/ocebm-levels-of-evidence
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.