Association between ambient air pollution and thyroid hormones levels: A systematic review and meta-analysis.

BACKGROUND: Growing studies have focused on the effects of ambient air pollution on thyroid hormones (THs), but the results were controversial. Therefore, a systematic review and meta-analysis was conducted by pooling current evidence on this association. METHODS: Four databases were searched for studies examining the associations of particulate matter [diameter ≤10 µm (PM10) or ≤2.5 µm (PM2.5)] and gaseous [sulfur dioxide (SO2), nitrogen dioxide (NO2), ozone (O3), carbon monoxide (CO)] pollutants with THs levels. Random effects models were used to pool the changes in THs levels with increasing air pollutant concentrations. Subgroup analyses were constructed by region, design, sample size, pollutant concentrations, evaluated methods, and potential risk exposure windows. RESULTS: A total of 14 studies covering 357,226 participants were included in this meta-analysis. The pooled results showed significant associations of exposure to PM2.5, PM10, NO2, SO2, and CO with decreases in free thyroxine (FT4) with percent changes (PC) ranging from -0.593 % to -3.925 %. PM2.5, NO2, and CO were negatively associated with levels of FT4/FT3 (PC: from -0.604 % to -2.975 %). In addition, results showed significant associations of PM2.5 with hypothyroxinemia and high thyroid-stimulating hormone (TSH). Subgroup analyses indicated that PM2.5 and NO2 were significantly associated with FT4 in studies of Chinese, and similar significant findings were found in studies of PM2.5 and FT4/FT3 in areas with higher concentrations of air pollutants and larger samples. PM2.5 exposure in the first trimester was found to be associated with lower FT4 levels in pregnant women. CONCLUSION: Our findings suggest that exposure to air pollution is associated with changes in THs levels. Enhanced management of highly polluted areas, identification of harmful components and sources of PM, and protection from harmful exposures in early pregnancy may be of great public health importance for the population's thyroid function.

Psychosocial predictors of persistent suicidal ideation among Chinese adolescents: A longitudinal study.

BACKGROUND: Suicide is the fourth leading cause of death among adolescents. Studies have shown that persistent suicidal ideation has a more important effect on suicidal behavior. The objective of this study was to identify predictors of persistent suicidal ideation. METHODS: Data were collected from 4225 Chinese middle and high school students. These adolescents were assessed for suicidal ideation at baseline and the second year. We used multinomial logistic regression (n = 4171) for the predictive effect of these factors on persistent suicidal ideation. We controlled for gender, residence, clinical diagnosis, clinical diagnosis family, suicide planning, and suicide attempts. RESULTS: Depressive symptoms are essential in predicting persistent suicidal ideation (OR = 14.0; p < 0.001). Persistent suicidal ideation was predicted by sleep disorders, such as poor sleep quality (OR = 2.3; p = 0.008), difficulty falling asleep (OR = 2.4; p = 0.005), frequently midnight awakening (OR = 1.9; p = 0.044), and frequent nightmares (OR = 2.1; p = 0.040). There was a significant association between concern with persistent ideation and parental-peer alienation (OR for father, 1.9[p = 0.024]; OR for mother, 3.1[p < 0.001]; OR for peer, 2.3[p = 0.003]). LIMITATIONS: All measures are based on self-report rather than objective assessment or clinical diagnostic assessment. CONCLUSIONS: Persistent suicidal ideation had a more important role in influencing suicide planning and attempt. Interventions targeting sleep disorders and attention to attachments in the home and school are particularly important in preventing persistent suicidal ideation in adolescents.

Effects of urban particulate matter on gut microbiome and partial schizophrenia-like symptoms in mice: Evidence from shotgun metagenomic and metabolomic profiling.

BACKGROUND: Epidemiological evidence reported that particulate matter (PM) was associated with increased schizophrenia (SCZ) risk. Disturbance of gut microbiome was involved in SCZ. However, it remains unclear whether PM induces SCZ-like symptoms and how gut microbiome regulates them. Therefore, a multi-omics animal experiment was conducted to verify how urban PM induces SCZ-like behavior and altered gut microbiota and metabolic pathways. METHODS: Using a completely random design, mice were divided into three groups: PM group, control group and MK801 group, which received daily tracheal instillation of PM solution, sterile PBS solution and intraperitoneal injection of MK801 (establish SCZ model), respectively. After a 14-day intervention, feces were collected for multi-omics testing (shotgun metagenomic sequencing and untargeted metabolomic profiling), followed by open field test, tail suspension test, and passive avoidance test. Besides, fecal microbiome of PM group and control group were transplanted into "pseudo-sterile" mice, then behavioral tests were conducted. RESULTS: Similar to MK801 group, mice in PM group showed SCZ-like symptoms, including increased spontaneous activity, excitability, anxiety and decreased learning and spatial memory. PM exposure significantly increased the relative abundance of Verrucomicrobia and decreased that of Fibrobacteres et al. The metabolism pathways of estrogen signaling (estriol, 16-glucuronide-estriol and 21-desoxycortisol) and choline metabolism (phosphocholine) were significantly altered by PM exposure. Verrucomicrobia was negatively correlated with the level of estriol, which was correlated with decreased learning and spatial memory. Fibrobacteres and Deinococcus-Thermus were positively correlated with the level of phosphocholine, which was correlated with increased spontaneous activity, excitability and anxiety. Fecal microbiome transplantation from PM group mice reproduced excitability and anxiety symptoms. CONCLUSIONS: Exposure to PM may affect composition of gut microbiome and alterations of estrogen signaling pathway and choline metabolism pathway, which were associated with partial SCZ-like behaviors. But whether gut microbiome regulates these metabolic pathways and behaviors remains to be determined.

Comparison of three machine learning models to predict suicidal ideation and depression among Chinese adolescents: A cross-sectional study.

BACKGROUND: Machine learning (ML) algorithms based on various clinicodemographic, psychometric, and biographic factors have been used to predict depression, suicidal ideation, and suicide attempt in adolescents, but there is still a need for more accurate and efficient models for screening the general adolescent population. In this study, we compared various ML methods to identify a model that most accurately predicts suicidal ideation and level of depression in a large cohort of school-aged adolescents. METHODS: Ten psychological scale scores and 20 sociodemographic parameters were collected from 10,243 Chinese adolescents in the first or second year of middle school and high school. These variables were then included in a random forest (RF) model, support vector machine (SVM) model, and decision tree model for factor screening, dichotomous prediction of suicidal ideation (yes/no), and trichotomous prediction of depression (no depression, mild-moderate depression, or major depression). RESULTS: The RF model demonstrated greater accuracy for predicting suicidal ideation (mean accuracy (ACC) = 87.3 %, SD = 3.2 %, area under curve (AUC) = 92.4 %) and depressive status (ACC = 84.0 %, SD = 2.8 %, AUC = 90.1 %) than SVM and decision tree models. We have also used the RF model to predict adolescents with both depression and suicidal ideation with satisfactory results. Significant differences were found in several sociodemographic parameters and scale scores among classification groups and differences in six factors between sexes. CONCLUSIONS: This RF model may prove valuable for predicting suicidal ideation, depression, and non-suicidal self-injury among the general population of Chinese adolescents.

Temporal trends of the association between temperature variation and hospitalizations for schizophrenia in Hefei, China from 2005 to 2019: a time-varying distribution lag nonlinear model.

Along with climate change, unstable weather patterns are becoming more frequent. However, the temporal trend associated with the effect of temperature variation on schizophrenia (SCZ) is not clear. Daily time-series data on SCZ and meteorological factors for 15-year between January 1, 2005 and December 31, 2019 were collected. And we used the Poisson regression model combined with the time-varying distribution lag nonlinear model (DLNM) to explore the temporal trend of the association between three temperature variation indicators (diurnal temperature range, DTR; temperature variability, TV; temperature change between neighboring days, TCN) and SCZ hospitalizations, respectively. Meanwhile, we also explore the temporal trend of the interaction between temperature and temperature variation. Stratified analyses were performed in different gender, age, and season. Across the whole population, we found a decreasing trend in the risk of SCZ hospitalizations associated with high DTR (from 1.721 to 1.029), TCN (from 1.642 to 1.066), and TV (TV0-1, from 1.034 to 0.994; TV0-2, from 1.041 to 0.994, TV0-3, from 1.044 to 0.992, TV0-4, from 1.049 to 0.992, TV0-5, from 1.055 to 0.993, TV0-6, from 1.059 to 0.991, TV0-7, from 1.059 to 0.990), but an increasing trend in low DTR (from 0.589 to 0.752). Subgroup analysis results further revealed different susceptible groups. Besides, the interactive effect suggests that temperature variation may cause greater harm under low-temperature conditions. There was a synergy between TCN and temperature on the addition and multiplication scales, which were 1.068 (1.007, 1.133) and 0.067 (0.009, 0.122), respectively. Our findings highlight public health interventions to mitigate temperature variation effects needed to focus not only on high temperature variations but also moderately low temperature variations. Future hospitalizations for SCZ associated with temperature variation may be more severely affected by temperature variability from low temperature environments. The temporal trend is associated with the effect of temperature variation on schizophrenia (SCZ).

Oxidative stress-mediated particulate matter affects the risk of relapse in schizophrenia patients: Air purification intervention-based panel study.

Particulate matter (PM) exposure increased the risk of hospital admission and was related to symptoms of schizophrenia (SCZ). However, there are limited studies on the relationship between PM exposure and SCZ relapse risk, and the underlying biological mechanisms remain unclear. We designed an air purification intervention study under a 16-day real air purifier scenario and another 16-day sham air purifier scenario, with a 2-day washout period. Twenty-four chronic stable male patients were recruited. The oxidative stress biomarkers were measured including serum catalase (CAT), superoxide dismutase (SOD), total antioxidant capacity (T-AOC), malondialdehyde (MDA), and nitric oxide (NO). The relapse risk was evaluated by the early signs scale (ESS). Linear mixed effect models were fitted to establish the associations between PM exposure and ESS and oxidative stress. Mediation model was performed to explore the mediation effect of oxidative stress on the PM-ESS association. Higher concentrations of PM2.5/PM10 exposure were associated with an elevated risk of relapse of SCZ. For each 10 µg/m3 in PM2.5 concentration, the scores of ESS and subscales of incipient psychosis (ESS-IP), depression/withdrawal (ESS-N), anxiety/agitation (ESS-A), and excitability/disinhibition (ESS-D) were increased by 4.112 (95% CI: 3.174, 5.050), 1.516 (95%CI: 1.178, 1.853), 1.143 (95%CI: 0.598, 1.689), 1.176 (95%CI: 0.727, 1.625) and 0.238 (95%CI: 0.013, 0.464), while logCAT, SOD and T-AOC were reduced by 0.039 U/ml (95% CI: 0.017, 0.060), 1.258 U/ml (95% CI: 0.541, 1.975), and 0.076 mmol/l (95% CI: 0.026, 0.126). In addition, pathways of "PM2.5→T-AOC→ESS-A″ and "PM2.5→T-AOC→ESS-D″ were found, with significant T-AOC mediated effects 15.70% (P = 0.02) and 52.99% (P = 0.04). Our findings suggest that PM may increase the risk of anxiety, depression, excitability, and incipient psychosis behaviors in SCZ patients, while reducing the function of the antioxidant system. The decrease of T-AOC may medicate the PM-ESS association in SCZ.

The Relationship Between Suicidal Ideation and Parental Attachment Among Adolescents: The Mediator of Anhedonia and Peer Attachment.

Background: Previous studies had shown that poor quality of early parental attachment is a risk factor for suicide, but few have focus on the mechanism between suicidal ideation and parental attachment. The aim of this study was to explore how parental attachment, anhedonia, and peer attachment were associated with suicidal ideation in adolescents. Method: Participants were enrolled in middle schools, in Hefei, Anhui, China. All participants completed socio-demographic characteristic and standard assessments on parental attachment, peer attachment, anhedonia, and suicidal ideation by paper surveys. The effect of parental attachment on suicidal ideation mediated by anhedonia and peer attachment was analyzed by a structural equation model (SEM) using SPSS AMOS 23.0. Results: The SEM analysis revealed that the standard total effect of parental attachment on suicidal ideation was -0.137 (Z=-27.00, 95% confidence interval [CI; -0.147, -0.127], p<0.001), with a direct effect of parental attachment on suicidal ideation of -0.107 (Z=-21.40, 95% CI [-0.117, -0.098], p<0.001), while the indirect effects were-0.002 (Z=-3.33, 95% CI [-0.003, -0.002], p<0.001) in the pathway of parental attachment-anhedonia-peer attachment-suicidal ideation, -0.019 (Z=-19.00, 95%CI [-0.022, -0.017], p<0.001) in the pathway of parental attachment-anhedonia-suicidal ideation, and-0.008 (Z=-7.00, 95% CI [-0.010, -0.007], p<0.001) in the pathway of parental attachment-peer attachment-suicidal ideation. Conclusion: The study suggested that parental attachment could directly influence suicidal ideation and indirectly influence suicidal ideation via anhedonia and peer attachment. The results emphasized the importance of attachment in infancy and verified the feasibility of intervention on anhedonia and peer attachment to prevent suicidal ideation.

Does the gut microbiome partially mediate the impact of air pollutants exposure on liver function? Evidence based on schizophrenia patients.

Air pollution may alter the composition of gut microbiome and subsequent liver-related metabolic disorders. Schizophrenia was often accompanied by liver dysfunction. But it was still unclear whether air pollutants affected liver function in patients with schizophrenia through gut microbiome. We aimed to clarify the impacts of long-term air pollutants on the gut microbiome and liver function in schizophrenia and to evaluate the intermediary effect of microbiome. Schizophrenia patients were recruited then serum biochemical indicators were tested. Air pollutant exposure in the previous year was retrospectively estimated by inverse distance weighting. The associations among air pollutants, gut microbiome, and liver function indicators in schizophrenia were estimated. Then the mediating effect of gut microbiome was further explored. The results showed that nitrogen dioxide (NO2), carbonic oxide (CO), ozone (O3), particulate matter with aerodynamic diameter ≤10 µm (PM10), and fine particulate matter (PM2.5) explained 2.68%-10.77% of the variation in gut microbiome composition (order level) in schizophrenia (all P < 0.05). Network correlation analysis indicated that air pollutants and liver function indicators were mainly related to Firmicutes, Actinobacteria, and Proteobacteria in schizophrenia. Long-term NO2 exposure significantly increased the levels of gamma-glutamyl transpeptidase (GGT) and glutamic pyruvic transaminase (GPT) in schizophrenia. Coriobacteriales mediated 13.98% and 49.56% (all P < 0.05) of the associations of long-term NO2 with GGT and GPT, respectively. To conclude, long-term NO2 exposure is positively associated with liver dysfunction in schizophrenia, in which gut microbiome plays an intermediary role. The two pathways, "NO2-Coriobacteriales-GGT" and "NO2-Coriobacteriales-GPT", would provide scientific evidence for the intervention of schizophrenia with liver dysfunction.

Effects of different heat exposure patterns (accumulated and transient) and schizophrenia hospitalizations: a time-series analysis on hourly temperature basis.

Growing studies have shown that high temperature is a potential risk factor of schizophrenia occurrence. Therefore, elaborate analysis of different temperature exposure patterns, such as cumulative heat exposure within a time period and transient exposure at a particular time point, is of important public health significance. This study aims to utilize hourly temperature data to better capture the effects of cumulative and transient heat exposures on schizophrenia during the warm season in Hefei, China. We included the daily mean temperature and daily schizophrenia hospitalizations into the distributed lag non-linear model (DLNM) to simulate the exposure-response curve and determine the heat threshold (19.4 °C). We calculated and applied a novel indicator-daily excess hourly heat (DEHH)-to examine the effects of cumulative heat exposure over a day on schizophrenia hospitalizations. Temperature measurements at each time point were also incorporated in the DLNM as independent exposure indicators to analyze the impact of transient heat exposure on schizophrenia. Each increment of interquartile range (IQR) in DEHH was associated with elevated risk of schizophrenia hospitalizations from lag 1 (RR = 1.036, 95% confidence interval (CI): 1.016, 1.057) to lag 4 (RR = 1.025, 95% CI: 1.005, 1.046). Men and people over 40 years old were more susceptible to DEHH. Besides, we found a greater risk of heat-related schizophrenia hospitalizations between 0 a.m. and 6 a.m. This study revealed the adverse effects of accumulated and transient heat exposures on schizophrenia hospitalizations. Our findings need to be further tested in other regions with distinct regional features.

Relationship between cerebellar structure and emotional memory in depression.

BACKGROUND: A few studies have been conducted on the relationship between cerebellar volume and emotional memory or clinical severity in major depressive disorder (MDD). In this study, we aimed to compare the volume and density of the cerebellar gray matter (GM) in patients with MDD and in healthy controls (HCs) and explore the association between these cerebellar parameters and measurements of emotional memory and clinical severity. METHOD: Voxel-based morphometry (VBM) and Individual Brain Atlases using Statistical Parametric Mapping (IBASPM) were used to assess GM density and volume in the cerebellum, respectively, in patients with MDD and the HCs. Indicators of emotional memory performance were measured, including the hit rate (HR), rate of false alarm (FA), precision (Pr = HR - FA) and emotional memory enhancement [∆Pr = Pr(emotion) - Pr(neutral)] values. Beck Depression Inventory (BDI) scores were used to measure the severity of depression. RESULTS: In the patients with MDD, the GM density was decreased in three cerebellar cortical regions and increased in three cerebellar cortical regions (p < .005). The GM volumes in eight cerebellar cortical regions were significantly smaller in the patients with MDD than in the HC subjects (p < .05). In the patients with MDD, the GM volume was correlated with the ∆Pr (p < .05) in two cerebellar cortical regions. The BDI scores were significantly correlated with the relative GM densities (p < .05) in 5 cerebellar cortical regions, and the GM volumes in 13 cerebellar cortical regions were correlated with the BDI scores in patients with MDD. CONCLUSIONS: Emotional memory and the severity of depressive symptoms are associated with structural changes in both the posterior and anterior GM regions in the cerebellum in patients with MDD. These findings could be useful for improving our understanding of the neurobiological mechanisms underlying emotional memory and explaining the abnormalities of the neural correlates that are associated with MDD.

Trait-related decision making impairment in obsessive-compulsive disorder: evidence from decision making under ambiguity but not decision making under risk.

This study aimed to investigate whether deficits in decision making were potential endophenotype markers for OCD considering different phases of the disease. Fifty-seven non-medicated OCD patients (nmOCD), 77 medicated OCD patients (mOCD), 48 remitted patients with OCD (rOCD) and 115 healthy controls were assessed with the Iowa Gambling Task (IGT), which measured decision making under ambiguity, and the Game of Dice Task (GDT), which measured decision making under risk. While the three patients groups showed impaired performance on the IGT compared with healthy controls, all patients showed intact performance on the GDT. Furthermore, the rOCD patients showed a preference for deck B, indicating that they showed more sensitivity to the frequency of loss than to the magnitude of loss, whereas the mOCD patients showed a preference for deck A, indicating that they had more sensitivity to the magnitude of loss than to the frequency of loss. These data suggested that OCD patients had trait-related impairments in decision making under ambiguity but not under risk, and that dissociation of decision making under ambiguity and under risk is an appropriate potential neurocognitive endophenotype for OCD. The subtle but meaningful differences in decision making performance between the OCD groups require further study.

Similarities and Differences in Decision-Making Impairments between Autism Spectrum Disorder and Schizophrenia.

Although individuals with autism spectrum disorders (ASD) and schizophrenia (SCH) share overlapping characteristics and may perform similarly on many cognitive tasks, cognitive dysfunctions common to both disorders do not necessarily share the same underlying mechanisms. Decision-making is currently a major research interest for both ASD and SCH. The aim of the present study was to make direct comparisons of decision-making and disorder-specific underlying neuropsychological mechanisms between the two disorders. Thirty-seven participants with ASD, 46 patients with SCH, and 80 healthy controls (HC) were assessed with the Iowa Gambling Task (IGT), which measures decision-making under ambiguity, and the Game of Dice Task (GDT), which measures decision-making under risk. The results revealed that both the ASD and SCH groups had deficits for both the IGT and the GDT compared with the HC. More importantly, in the IGT, participants with ASD displayed a preference for deck A, indicating that they had more sensitivity to the magnitude of loss than to the frequency of loss, whereas patients with SCH displayed a preference for deck B, indicating that they showed more sensitivity to the frequency of loss than to the magnitude of loss. In the GDT, the impaired performance might be due to the deficits in executive functions in patients with SCH, whereas the impaired performance might be due to the deficits in feedback processing in participants with ASD. These findings demonstrate that there are similar impairments in decision-making tasks between ASD and SCH; however, these two disorders may have different impairment mechanisms.

Dissociation of decision making under ambiguity and decision making under risk: a neurocognitive endophenotype candidate for obsessive-compulsive disorder.

Evidence in the literature suggests that executive dysfunction is regarded as an endophenotype candidate for obsessive-compulsive disorder (OCD). Decision making is an important domain of executive function. However, few studies that have investigated whether decision making is a potential endophenotype for OCD have produced inconsistent results. Differences in the findings across these studies may be attributed to several factors: different study materials, comorbidity, medication, etc. There are at least two types of decision making that differ mainly in the degree of uncertainty and how much useful information about consequences and their probabilities are provided to the decision maker: decision making under ambiguity and decision making under risk. The aim of the present study was to simultaneously examine decision making under ambiguity as assessed by the Iowa Gambling Task (IGT) and decision making under risk as measured by the Game of Dice Task (GDT) in OCD patients and their unaffected first-degree relative (UFDR) for the first time. The study analyzed 55 medication-naïve, non-depressed OCD patient probands, 55 UFDRs of the OCD patients and 55 healthy matched comparison subjects (CS) without a family history of OCD with the IGT, the GDT and a neuropsychological test battery. While the OCD patients and the UFDRs performed worse than the CS on the IGT, they were unimpaired on the GDT. Our study supports the claim that decision making under ambiguity differs from decision making under risk and suggests that dissociation of decision making under ambiguity and decision making under risk may qualify to be a neurocognitive endophenotypes for OCD.

[Time perception in depressed and manic patients].

OBJECTIVE: To investigate the time perception in affective disorders by using neuropsychological tests and to try to elucidate its neurobiochemical mechanism. METHODS: Using a time reproduction task, a comparative study was conducted for 28 depressive patients, 22 manic patients, and 26 age and education level matched healthy persons as healthy controls. RESULTS: Both depressive patients and manic patients are abnormal (P < 0.001), depressive patients over-reproduced the time interval than healthy controls (600 ms/delay 1 s: 1.6 +/- 0.6, P < 0.001; 600 ms/delay 5 s: 1.7 +/- 0.6, P < 0.001; 3 s/delay 1 s: 3.9 +/- 0.9, P < 0.001; 3 s/delay 5 s: 3.9 +/- 0.7, P < 0.001; 5 s/delay 1 s: 5.9 +/- 1.3, P < 0.001; 5 s/delay 5 s: 6.1 +/- 1.3, P < 0.001), yet manic patients under-reproduced the time interval (600 ms/delay 1 s: 0.7 +/- 0.2, P < 0.01; 600 ms/delay 5 s: 0.6 +/- 0.3, P < 0.001; 3 s/delay 1 s: 1.7 +/- 0.5, P < 0.001; 3 s/delay 5 s: 1.8 +/- 0.6, P < 0.001; 5 s/delay 1 s: 2.9 +/- 0.7, P < 0.001; 5 s/delay 5 s: 3.0 +/- 0.8, P < 0.001). The results of time reproduction task in patients were not related to age, education, duration of illness, number of admission (P > 0.05), but had some relation to severity of illness.And the results were positively correlated with the score of HAMD in depressive patients (six times: r = 0.44, 0.46, 0.73, 0.61, 0.55, 0.50, P < 0.05), but negatively with the score of BRMS in manic patients (six times: r = -0.57, -0.54, -0.71, -0.69, -0.80, -0.71, P < 0.05). CONCLUSION: Emotion will affect one's time perception. And the neurotransmitter in brain may participate in the processes of time perception.

Diurnal rhythms of free estradiol and cortisol during the normal menstrual cycle in women with major depression.

To investigate whether depression is accompanied by changes in diurnal rhythms of free estradiol and cortisol in different phases of the menstrual cycle, we measured these two hormone levels in saliva samples collected every 2 h for 24 h from 15 healthy normally cycling women and 12 age-matched normally cycling women suffering from major depression taking antidepressants. The assessments were repeated four times over one menstrual cycle: during menstruation and in the late follicular/peri-ovulating, early to mid-luteal and late luteal phases, respectively. Quantification with a nonlinear periodic regression model revealed distinct diurnal rhythms in free estradiol and free cortisol in all subjects. For the diurnal cortisol rhythm, significant differences were found in the peak-width and ultradian amplitude among different menstrual phases, both in controls and depressed patients, while no significant differences were found between the two groups. The diurnal estradiol rhythm, on the other hand, was quite consistent among different menstrual phases within both groups, while the depressed patients had overall larger amplitudes than controls, which is negatively correlated with disease duration. Significant positive correlations between the two hormone rhythms were found for 24-h mean level (mesor), peak, and trough in late luteal phase, and for ultradian harmonics in early to mid-luteal phase in controls, but only for ultradian harmonics in late follicular/peri-ovulating phase and for acrophase in the menstruation phase in depressed patients. A sub-analysis was also performed in patients who received Fluoxetine (n = 7). The findings implicate a close correlation between the hypothalamic-pituitary-adrenal axis and the hypothalamic-pituitary-gonadal axis, both of which may be involved in depression.