Clock gene signature predicts Insomnia and links to sleep/circadian parameters

Clock gene signature predicts Insomnia and links to sleep/circadian parameters | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Clock gene signature predicts Insomnia and links to sleep/circadian parameters Ana Rita Alvaro, Catarina Carvalhas-Almeida, João Alves, Tiago Davi, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6268489/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Chronic Insomnia is a prevalent sleep disorder that remains difficult to diagnose due to subjective symptoms and heterogeneous presentations. The most severe form, insomnia with short sleep duration (ISSD), is defined by a total sleep time of less than six hours on polysomnography. However, objective assessments are rarely recommended in diagnostic guidelines, highlighting the need for alternative biomarkers. Disruptions in the circadian clock system may contribute to chronic insomnia, though the extent of these effects remains unclear. In this study, we investigate sleep and circadian rhythm-related alterations in chronic insomnia and its subtypes, ISSD and insomnia with normal sleep duration (INSD), by assessing plasma cortisol, wrist and axillary temperature, and clock gene expression in peripheral blood mononuclear cells (PBMCs). Additionally, we use machine learning to identify the most relevant clock genes for detecting insomnia and classifying its subtypes. Chronic insomnia patients exhibited reduced body temperature rhythms, elevated nighttime cortisol levels, and significant alterations in clock genes expression, including in BMAL1 , PER1-2 , REV-ERBα , and REV-ERBβ , compared to controls. Most alterations were more significant in the ISSD. Moreover, associations between clock gene expression, sleep-related parameters and insomnia severity index (ISI) scores were identified. Using machine learning, we identified three genes as sensitive biomarkers distinguishing chronic insomnia from controls and differentiating between ISSD and INSD subtypes. Our findings suggest that circadian markers and machine learning could improve understanding of chronic insomnia and aid biomarker discovery for diagnosis. Health sciences/Biomarkers/Predictive markers Biological sciences/Molecular biology Insomnia machine learning actigraphy circadian clock polysomnography Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Chronic insomnia disorder is a prevalent condition affecting approximately 10% of the global population 1 . It is defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, Text Revision (DSM-5-TR) and the International Classification of Sleep Disorders, 3rd edition, Text Revision (ICSD-3-TR) as persistent nighttime symptoms despite adequate sleep opportunities, accompanied by daytime impairments such as fatigue, occurring at least three nights per week for over three months 2 , 3 . Diagnosis typically relies on self-reported symptoms, limiting objective assessment and standardization. Over time, varying definitions of insomnia and its subtypes, such as primary insomnia or classifications based on sleep duration or individual traits, have led to inconsistencies in research findings and diagnostic challenges 4 – 8 , hindering the identification of reliable biomarkers 9 , 10 . A clinically relevant subtype, insomnia with short sleep duration (ISSD), is characterized by a total sleep time (TST) of less than six hours, as measured by single-night polysomnography (PSG) 7 . ISSD is associated with an increased risk of significant medical and psychiatric comorbidities, adverse cardiometabolic outcomes and neurocognitive impairments 7 , 11 , 12 . Moreover, individuals with ISSD respond less favorably to cognitive-behavioral therapy compared to those with insomnia with normal sleep duration (INSD, TST > 6 hours) 7 , highlighting the need for subtyping insomnia during diagnosis for tailored treatment approaches. Despite the clinical relevance of ISSD and INSD, objective measures such as PSG and actigraphy are rarely incorporated into routine diagnosis of insomnia due to high costs, logistical challenges 13 , and their inability to capture subjective wakefulness experiences 14 . Additionally, European guidelines do not mandate objective sleep measures for diagnosing insomnia 9 . To address these challenges, identifying reliable biological markers could enhance diagnostic precision and inform targeted treatment strategies. Emerging evidence suggests that chronic insomnia may disrupt the circadian clock system, though the extent of these effects remains unclear 15 – 19 . Previous studies have proposed the analysis of clock gene expression profiles in human peripheral blood mononuclear cells (PBMCs) as a promising approach for assessing circadian rhythm 20 – 23 .. Indeed, our prior study has shown through machine learning analysis, that patients with obstructive sleep apnea (OSA), another highly prevalent sleep disorder, can be distinguished from healthy controls based on altered expression patterns of several clock genes 24 , suggesting that clock gene expression may serve as potential biomarker for sleep disorders. In recent years, machine learning approaches have increasingly been utilized to identify biomarkers for different sleep disorders and to monitor treatment responses 24 , 25 . Building on this concept, we investigated the impact of chronic insomnia on the biological clock by analyzing key physiological markers, including plasma cortisol, body temperature, and clock gene expression in PBMCs. We also characterized sleep using PSG, assessed actigraphy-derived circadian parameters, and machine learning-based analysis for identifying biomarkers that enable robust differentiation of insomnia patients from controls and between subtypes. Our findings reveal significant circadian disruptions and altered clock gene expression in chronic insomnia, particularly in ISSD, underscoring the potential of circadian biomarkers for improved diagnosis and treatment. Machine learning analysis enabled robust differentiation between insomnia patients and controls, as well as between ISSD and INSD subtypes, based on clock gene expression. Overall, this work advances our understanding of insomnia’s circadian disruption and highlights the utility of integrating circadian biomarkers with interpretable machine learning techniques to improve insomnia diagnosis, optimize treatment strategies based on subtype, and reduce the societal burden of insomnia, including work absenteeism and healthcare costs. Methods Ethical Approval This study received approval from the Ethical Committees of the Faculty of Medicine, University of Coimbra (CE-162/2021) and the Coimbra Hospital and University Centre (ULSC, OBS.SF.55-2021) in Coimbra, Portugal. All experimental procedures adhered to the ethical guidelines and regulations established in the 1964 Declaration of Helsinki and its subsequent amendments. The study also complied with the provisions of Regulation (EU) 2016/679 of the European Parliament and Council, governing the General Data Protection Regulation (GDPR), as well as Portuguese Law n.º 12/2005 of January 26 and its implementing regulations detailed in Decree-Law n.º 131/2014 of August 29, 2014. Informed consent was obtained from all participants prior to their inclusion in the study. Study design Female and male adult volunteers (age ≥ 18 years) suspected of having chronic insomnia were recruited for a sleep study at the Sleep Medicine Unit of CHUC, Coimbra, Portugal, between January 2022 and September 2023. All selected subjects underwent a detailed screening process, which included a clinical interview conducted by an experienced psychiatrist, validated questionnaires and continuous wrist actigraphy monitoring over two-weeks (ActTrust® 2, Condor Instruments Ltda, SP, Brazil). Participants identified as potential chronic insomnia cases were asked to undergo a single-night, home-based polysomnography (PSG) study (type II) to determine sleep duration subtype and were categorized into two groups: insomnia patients with short sleep duration (ISSD) (TST < 6 hours) and insomnia patients with normal sleep duration (INSD) (TST ≥ 6 hours) 7 . PSG was also used to rule out other sleep disorders. Sleep studies were staged and scored according to the American Academy of Sleep Medicine (AASM) scoring manual v2.4, 2012. Healthy volunteers were invited to participate in a one-night, home-based sleep study using the WatchPAT™ 300 device (Itmar Medical, Gasoxmed, Portugal) to record their respiratory events and sleep parameters. Control subjects of similar age (± 5 years) and including both sexes, were classified as disease-free if met the following criteria: no lifetime history of significant insomnia-associated symptoms, TST ≥ 6 hours, absence of medical or psychiatric comorbidities that could interfere with sleep (e.g. major depressive disorder) verified through clinical history and a structured clinical interview using the SCID-IV; no history of other sleep disorders as confirmed by clinical evaluation and screening PSG; no engagement in shiftwork within the preceding 6 months; no current use of prescription medications of over-the-counter products that could affect sleep; not being pregnant or lactating within the past 6 months. Data from the PSG test (PSG report) and WatchPAT™ 300 (report) were obtained from each participant. An ID was assigned to each subject to ensure data confidentiality and traceability. On the morning following PSG or WatchPAT™ 300 examination, participants completed the following validated questionnaires to assess various sleep and mental health parameters: Insomnia Severity Index (ISI) 26 , Epworth sleepiness scale 27 , STOP-BANG 28 and Hospital Anxiety and Depression Scale (HADS) 29 . During this period, individual data was collected from all participants, including sex, age, body mass index (BMI), lifestyle (sleep/wake routine, diet, meal schedules, physical exercise), and clinical history (comorbidities and medication). Blood samples were collected from all subjects following their sleep studies during hospital visits. Samples were collected at four different times: in the morning (8:00 and 11:00 h), afternoon (15:30 h) and night (21:30 h). Blood was collected for the analysis of cortisol plasma levels, and clock genes expression in PBMCs. Axillary temperature was monitored at the same four time points. Between sampling sessions, participants were allowed to leave the Sleep Unit to continue their usual daily activities, ensuring minimal disruption to their routine. The study design is summarized in Figure 1. Measuring rest:activity and skin temperature rhythms The structure and timing of daily rest:activity patterns were assessed using wrist-worn ActTrust 2 actigraphs (Condor Instruments, SP, Brazil). Participants wore the devices continuously on their non-dominant wrist for 14 consecutive days under normal ambulatory conditions. During this period, participants were instructed to follow their usual routines and record their sleep patterns in a sleep diary. They were advised to remove the actigraphs during activities involving water (e.g. bathing, dishwashing or swimming). The actigraphs were configured with a sampling frequency of 25 Hz to optimize battery life over the recording period. The devices continuously recorded motor activity (using the Proportional Integral Mode, PIM), skin temperature (°C), and light exposure (lux) in one-minute epochs. PIM measures movement intensity by calculating the area under the signal curve for each epoch 30 . The raw actigraphy data was processed using the ActStudio software version 2.1.2 (Condor Instruments, SP, Brazil), which automatically detected sleep-wake cycles, off-wrist periods, sleep parameters, parametric (COSINOR) and non-parametric circadian rhythm analysis (NPCRA) calculations, and export of the study graphs. Off-wrist periods were automatically excluded from the analysis by the software. Sleep parameters included: sleep efficiency (the percentage of time spent asleep while in bed), total sleep time (the cumulative amount of sleep), number of awakenings and wakefulness after sleep onset (WASO). COSINOR analysis included measures of amplitude (the difference between the peak and the mean value of the cosine function), midline estimating statistic of rhythm (MESOR, mean value of a rhythmic variable over a 24-hour period) and acrophase (timing of the peak activity in a circadian rhythm). NPCRA was conducted to assess the circadian structure of rest:activity rhythms, including intradaily variability (IV), interdaily stability (IS), and relative amplitude (RA). The analysis also examined average activity levels during the least active 5-hour period (L5) and the most active 10-hour period (M10). IV reflects rest:activity fragmentation, with higher values denoting greater consistency. IS measures rhythm stability across days, with higher values denoting greater consistency. RA quantifies rhythm strength by comparing activity levels between M10 and L5. Within-subject variability in L5, M10 and RA was assessed using the standard deviation of daily measures. For generating the actigraphy plots for wrist temperature, PIM, TAT, and ZCM, actigraphy data was grouped by the mean per minute, for each participant and displayed on a 24-hour horizontal axis. Data transformations were performed using the Numpy and Pandas Python packages, within the Jupyter environment. The visualizations were generated using the Vega-Lite visualization grammar, with the Altair Python binding. The layered chart has two components: an area chart for the SEM bounds and a line chart for the mean values of the data. Axillary temperature Axillary temperature was measured twice at each of the four time points within 24 h (8:00, 11:00, 15:30 and 21:30 h) in both chronic insomnia patients and control subjects, following the sleep unit routine, through an electronic thermometer. The average between the two temperature measurements was used. Blood samples collection and processing Blood samples (15–20 mL) were collected from chronic insomnia patients and control subjects at four time points within 24 hours (8:00, 11:00, 15:30, and 21:30 h) via the antecubital vein using a 21-gauge needle into K2EDTA tubes (#367839, BD Vacutainer) and subsequently processed as previously described 24 . Plasma and PBMCs aliquots were stored at - 80 ⁰C until further analysis. Cortisol plasma levels assessment Plasma cortisol levels were quantified using human-specific commercial ELISA assays (EH0641 respectively, Fine Test), according to the manufacturer's protocol. Absorbance was measured at 450 nm using a SpectraMax Plus 384 Microplate Reader (Molecular Devices). The concentrations were then calculated by interpolating the results from standard curves generated by plotting the concentration of the standards provided by each kit against their absorbance using a sigmoidal 4-parameter logistic (4-PL) curve, with concentration expressed logarithmically. Clock gene expression analysis Total RNA was extracted from PBMCs using the miRCURY RNA Isolation kit – Cell and Plant (#300,110, Exiqon) according to manufacturer's protocol, with DNAse digestion to prevent genomic DNA contamination. Total RNA was quantified by optical density (OD) using the ND-1000 Nanodrop Spectrophotometer (Thermo Scientific), and purity was assessed by the OD ratio at 260 and 280 nm. RNA samples were converted into cDNA, using the iScript cDNA Synthesis Kit (Bio-Rad), according to manufacturer's instructions and as previously described 24 . The mRNA levels of BMAL1, CLOCK , PER1–3, CRY1–2, REV-ERBα, and REV-ERBβ of all the enrolled subjects were assessed by real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR), using the iCycler iQ™ Real-Time PCR Detection System (Bio-Rad). Primer sequences and qPCR conditions are listed in Table S1. Relative gene expression was calculated according to the ΔCT method 31 . Each assay included a non-template control (NTC), a no reverse transcription (NRT) control, and a standard curve for each target gene that was used in all plates to normalize for interplate variability. Hypoxanthine-guanine phosphoribosyltransferase ( HPRT ), Glyceraldehyde 3-phosphate dehydrogenase ( GAPDH ) and β−2-microglobulin ( β−2 M ) were used as housekeeping genes, as validated previously 24 . Bio-Rad CFX Maestro software (Bio-Rad) was used to automatically determine amplification efficiencies and threshold cycles (CT). mRNA expression data is presented as ΔΔCT values, relative to the average of the ΔCt of all time points of all enrolled subjects, except for machine learning analysis, where the ΔCT values were used. Statistical analysis If not stated otherwise, statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, Inc). Data distribution was assessed using the Shapiro–Wilk test and the quartile–quartile plot (QQ-plot) was visually inspected. Based on data distribution, unpaired t-tests were performed to detect statistically significant differences between control and chronic insomnia group, at each time of the day. For correlation dot plot, pairwise Spearman correlation coefficients were calculated using Origin 2025 (OriginLab, New York, USA). All statistical tests performed were two-sided, with statistical significance set at 0.05. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. No values were excluded. Specific statistical details are described in the figure legends. Rhythmicity analysis For the detection of circadian oscillating genes, harmonic regression fitting was used as implemented in the geoTS R package using the haRmonics function 32 . The dataset was categorized into three groups: (1) controls, (2) ISSD and (3) INSD. Missing values were imputed using the respective group's mean ΔΔCT values normalized to the average of all participants. We then computed the mean gene expression ΔΔCT values normalized to the average of all participants values across four time points (8:00, 11:00, 15:30, and 21:30 hours) for each group and each gene. The summary plot using data from each participant of the study is depicted using the standard error of mean and the figures plotted in R programming software using ggplot2 package (v.3.3.2). A smooth harmonic regression curve was used for connecting the summary plot. For each gene and each group, we fitted the harmonic regression model y(t) = m + a × sin (2 × π × t/ω) + b × cos (2 × π × t/ω), with an ω being a period of 24 hours. The model parameters ( m , a , b ) were estimated using haRmonics function. The amplitude A and acrophase φ were derived using: A = √ ( a 2 + b 2 ), φ = arctan ( b , a ). To visualize the fitted curves, a second-degree polynomial model of the form: y(x) = b 0 + b 1 x + b 2 x 2 was fitted to the harmonic regression outputs at the four sampling time points. We then computed the associated p-values via the F-test, which was obtained using the F-statistic components provided by the summary function in R. Error bars (standard errors) were obtained by dividing the sample standard deviation by the number of participants in each group. For each gene, we generated separate plots illustrating the group-wise mean gene-expression values ΔΔCT at each sampling time, with error bars and the corresponding smoothed polynomial curves for each group. Machine learning To develop and validate a predictive model for classifying insomnia status based on clock gene expression, we used gene expression data from PBMCs collected at multiple time points throughout the day (8:00, 11:00, 15:30, and 21:30 h). For each participant, we derived a single feature vector by averaging gene expression levels across all time points for each gene. We implemented a Support Vector Machine (SVM) using the Scikit-learn library in Python. In each training iteration, we selected a subset of the data containing three genes, representing each participant as a 3D data point with a feature vector of mean expression values for the selected genes. To identify the most predictive genes, we conducted multiple training iterations, testing all possible three-gene combinations from a set of nine genes. The dataset was split into training, validation, and test sets. Hyperparameter tuning was performed using the training and validation sets, varying parameters such as kernel type and regularization through grid search to optimize performance. The ten models achieving the highest validation accuracy were selected. Finally, we evaluated generalization performance by testing the best-performing SVM configuration on the test set. This approach ensured unbiased performance estimates on unseen data. Results Cohort profile: demographic and clinical characteristics During recruitment, 30 potential insomnia patients and 15 potential controls completed the baseline visit. Of these, 14 patients with insomnia and 13 control subjects met the eligibility criteria and were included in the study (Figure 1A and B). The demographic and clinical characteristics are summarized in Table 1. The insomnia group comprised 1 male and 13 females, while the control group included 4 males and 9 females. All participants were Caucasian. There were no statistically significant differences between the groups in terms of age, BMI, STOP-BANG scores, or daytime sleepiness (Table 1). The ISI score was significantly higher in the insomnia group compared to controls. Similarly, Hospital Anxiety and Depression Scale (HADS) scores were significantly elevated in the insomnia group, relative to controls. It should be noted, none of the participants met the clinical criteria for anxiety or depression, as those with diagnosed psychiatric conditions were excluded. Insomnia patients also had higher rates of comorbid conditions, including hypertension, cardiovascular diseases, metabolic disorders (such as diabetes and dyslipidemia), and gastrointestinal conditions. Details of medication use are provided in Table 1. Accurate assessment of sleep duration and quality is essential for understanding insomnia, which is characterized by subjective complaints of poor sleep despite variations in objective measurements. While PSG remains the gold standard for evaluating sleep architecture and diagnosing sleep disorders, actigraphy provides a non-invasive, long-term alternative that also captures circadian parameters 33 . However, the degree to which these methods detect sleep disturbances in chronic insomnia remains a key question. This study aimed to characterize sleep parameters and stratify patients by objective sleep duration based on a single-night PSG and to determine if patients with chronic insomnia exhibited circadian parameter alterations by actigraphy (Figure 1B). Based on TST duration, 8 patients were classified as ISSD, and 6 were categorized with INSD (Figure 1A and Table 2). PSG analysis shows that patients with chronic insomnia and the ISSD subgroup, but not the INSD, show significantly reduced TST compared to controls. Sleep latency was significantly prolonged in the ISSD subgroup compared to controls, while sleep efficiency was significantly reduced in the insomnia group and the ISSD subgroup compared to controls. Additionally, REM percentage (%) was significantly reduced and the number of awakenings was markedly higher in all insomnia groups compared to controls. Importantly, no significant differences were identified in apnea-hypopnea index (AHI) values across groups, confirming that sleep-disordered breathing was not a confounding factor in our sample (Table 2). Our findings confirm that ISSD patients experience the most profound sleep disturbances, including significantly reduced TST, prolonged sleep latency, and lower sleep efficiency, while INSD patients show milder impairments. Lower body temperatures suggest circadian dysfunction in patients with chronic insomnia To further investigate sleep-wake patterns and circadian rhythm stability, we examined actograms, which provide a visual representation of rest-activity cycles over multiple days 33 . These actograms allowed us to assess differences in sleep organization between groups, offering complementary insights to the PSG and actigraphy-derived sleep parameters. Representative actograms revealed marked differences between control participants and patients with insomnia. As expected, control group actograms showed well-organized entrainment to the day/night cycle, with stable phase and amplitude across days, aside from slight weekend phase shifts. In contrast, actograms from ISSD and INSD patients demonstrated increased sleep-onset latency, disrupted sleep maintenance and reduced sleep duration on most nights. Both controls and insomnia patients (ISSD and INSD) exhibited slightly increased sleep periods during weekends (Figure 2A). Quantitative analysis of actigraphy-derived rest-activity rhythms further supported these findings. Insomnia patients demonstrated higher overall activity levels, as indicated by increased average daily activity (PIM) graphs (Figure 2B and 2C), time-above-threshold (TAT) and Zero Crossing Method (ZCM) values (Supplementary Figure S1A and B), suggesting prolonged wakefulness of chronic insomnia patients. We next assessed circadian rhythm parameters, including parametric and non-parametric measures, wrist temperature measurements. However, most parametric and non-parametric measures, including M10, L5, RA, IS, circadian function index (CFI), MESOR, and amplitude did not significantly differ between groups. The only significant circadian rhythm alteration was lower intradaily variability in insomnia patients (Supplementary Figure S1 C-J). Skin temperature rhythms were shown to be disrupted in chronic insomnia, with patients exhibiting lower wrist temperatures compared to controls (Figure 2D), which could be indicative of circadian dysfunction. Despite these minimal differences in circadian rhythm robustness, a strong negative correlation was observed between CFI, IS and ISI scores (Figure 3 E), indicating that worse insomnia symptoms were associated with weaker circadian function. Given the reduced wrist temperature observed in patients with chronic insomnia and the negative correlation observed between actigraphy-derived circadian measures (CFI and IS) and ISI scores, we hypothesized that chronic insomnia might be associated with circadian disruption in biological outputs. To further investigate this, we also assessed the impact of insomnia on the biological clock by measuring axillary temperature and plasma levels of cortisol at four time points across the day (08:00, 11:00, 15:30, and 21:30 h) in both insomnia patients (ISSD and INSD) and control subjects (Figure 3). Significant differences in axillary temperature were identified between insomnia patients and controls. The insomnia group exhibited significantly lower axillary temperatures during the morning time points (08:00 and 11:00 h). Notably, the INSD subgroup mirrored this pattern, while the ISSD subgroup demonstrated significantly lower axillary temperature only at 08:00 h (Figure 3A-C). These diurnal differences in axillary temperature were consistent with wrist temperature data obtained via actigraphy (Figure 2D). We also observed that cortisol levels had distinct circadian pattern of dysregulation in insomnia patients. The overall insomnia group showed significantly elevated cortisol levels at night (21:30 h). Further subgroup analysis indicated that the ISSD subgroup displayed heightened cortisol levels at both 15:30 and 21:30 h, while no significant cortisol alterations were observed in the INSD subgroup (Figure 3A-C). These results show that chronic insomnia patients, specially the ISSD group, show cortisol dysregulation. These data suggest modest circadian dysfunction through actigraphy, as reflected by largely unchanged parametric and non-parametric measures and a significant correlation between circadian metrics and insomnia severity. Additionally, lower wrist and axillary temperatures, accompanied by elevated nighttime cortisol in insomnia patients suggest further circadian misalignment, warranting further investigation. Chronic Insomnia, in particular ISSD, promotes alterations in expression levels and the temporal expression profiles of several clock genes Given the overall impact of chronic insomnia on temperature and cortisol rhythms, we aimed to investigate whether clock genes were affected at the transcriptional level in peripheral blood samples, given their importance in circadian rhythm regulation. To this end, we analyzed the expression levels of nine core-clock genes ( BMAL1, CLOCK, PER1-3, CRY1-2, REV-ERBα, and REV-ERBβ ) at the four time points of the study (8:00, 11:00, 15:30 and 21:30 h) in PBMCs of patients with chronic insomnia and its subgroups, compared to control subjects. Patients with insomnia showed higher expression levels of PER1 (8:00 and 15:30 h) , PER2 (8:00 and 15:30 h) , REV-ERBα (11:00 h), and REV-ERBβ (15:30 h), relative to control subjects. No significant changes were detected in the expression levels of BMAL , CLOCK, PER1, CRY1 and CRY2 in PBMCs of insomnia patients (Figure 4A). Subgroup analysis revealed that the ISSD subgroup exhibited similar alterations in clock gene expression to the total chronic insomnia group. By contrast, the INSD subgroup showed fewer differences, when compared with the insomnia group, with significantly higher expression levels of BMAL , PER1 , and CRY2 only at the morning timepoint (08:00 h) (Figure 4B and C). To better characterize the clock genes expression along the day and part of the night, we then analyzed the alignment of the datasets with a 24-hour rhythmic oscillation using harmonic regression. Specifically, we applied a harmonic regression model with a fixed 24-hour period to the average data from all patients with insomnia, as well as to its subgroups (Supplementary Figure S2 A-C). The results indicate that PER1 shows a marginally significant result (p = 0.048) in the INSD group, suggesting possible differences in circadian rhythms compared to the control group (Supplementary Table S2). Although trends in amplitude and acrophase are apparent, particularly the generally higher amplitudes in insomnia-related groups, the p -values indicated that these differences, except for PER1 in the INSD group, are not statistically significant. Using this methodology of analysis, the results seem to suggest that circadian rhythms of clock genes are relatively stable across the groups in this dataset. Explorative correlation analysis shows significant correlations between insomnia severity and circadian markers To further investigate the relationship between insomnia severity and circadian regulation, we conducted an exploratory correlation analysis to assess the associations between insomnia sleep-related parameters (derived from PSG), actigraphy-derived circadian metrics, and biological circadian markers. Interestingly, age emerged as a significant factor, positively correlating with the expression levels of BMAL , CLOCK , PER1 , CRY1 , and REV-ERBα in patients with insomnia and its subgroups, suggesting that clock gene expression may change with aging in individuals with chronic insomnia (Figure 5A). Our analysis identified strong negative correlations between ISI and expression levels of BMAL , CLOCK , PER2 , CRY1-2, and cortisol . These correlations were even more pronounced when examining the two insomnia subgroups, ISSD and INSD, separately. Additionally, ISI showed significant positive correlations with axillary temperature and PER1 expression, further supporting the involvement of circadian dysfunction in insomnia pathology. TST was significantly negatively correlated with nighttime cortisol levels across all insomnia groups, reinforcing the well-established link between short sleep duration and increased hyperarousal. Moreover, modest associations were also observed between TST and PER2 and REV-ERBβ expression, indicating potential links between sleep duration and molecular circadian rhythms (Figure 5 A-C). Among the actigraphy-derived circadian parameters, the strongest associations were observed between CFI and the expression levels of CLOCK, PER1, PER2, CRY1 , and CRY2 in the insomnia group. These findings suggest that molecular clock disruptions could be reflected in circadian activity patterns, reinforcing the role of intrinsic circadian misalignment in insomnia. Subgroup analysis further highlighted distinct associations in INSD patients. Specifically, axillary temperature was negatively correlated with age, ISI, relative amplitude (RA), and CFI. This suggests that reduced circadian stability (as reflected by lower CFI and RA values) is linked to both insomnia severity and impaired thermoregulation, particularly in INSD patients. Interestingly, the negative correlation between axillary temperature and CFI further supports the role of altered thermoregulation in insomnia-related circadian disruption (Figure 5C). Overall, these findings highlight the complex interplay between insomnia severity, circadian markers, and molecular clock regulation. The observed associations suggest that chronic insomnia, particularly its subtypes, is characterized by disruptions in both behavioral and molecular circadian rhythms. Supervised machine learning predicts insomnia disorder using clock gene expression data To determine whether the observed differences in clock gene expression could distinguish patients from controls and between insomnia subtypes, we applied a supervised machine learning using the clock gene expression dataset. We trained a linear support vector machine (SVM) with leave-one-out cross-validation, using mean gene expression levels as input features to classify samples as "control" or "patient”. In each iteration, the model was trained on all subjects except one, whose health status (control or patient) was then predicted. This approach achieved robust classification performance, correctly identifying 13 insomnia subjects with an accuracy of 92%, along with 92% precision and 92% recall (Figure 6A and Supplementary Table S3). Using the same SVM approach, we classified insomnia subgroups, achieving an overall accuracy of 92%, with 86% precision and 100% recall (Figure 6B and Supplementary Table S4). These findings highlight a significant distinction between control and patient groups, identified solely from mean core-clock gene expression levels. Additionally, the model’s ability to differentiate chronic insomnia subtypes underscores its potential for detecting nuanced molecular differences. Discussion Chronic insomnia is a complex disorder characterized by physiological and molecular alterations. Understanding its underlying mechanisms and identifying novel biomarkers is essential for improving diagnostic strategies and moving toward more objective approaches. Our findings support growing evidence that chronic insomnia disrupts circadian rhythms, with ISSD patients exhibiting the most pronounced impairments, including sleep disturbances, altered body temperature rhythms, elevated nighttime cortisol, and significant changes in clock gene expression. Additionally, machine learning identified key genes that may serve as biomarkers for diagnosing insomnia and distinguishing between subtypes. Consistent with previous findings 7 , ISSD patients demonstrated the most profound sleep disturbances, including significantly reduced TST, prolonged sleep latency, and lower sleep efficiency, whereas INSD patients exhibited milder impairments. A key limitation of standard PSG is the reliance on single-night recordings, which may not fully capture the severity of insomnia due to night-to-night variability 34 . Additionally, discrepancies between subjective and objective sleep assessments, often termed paradoxical insomnia or sleep state misperception, remain a challenge, as PSG often reveals less severe sleep disturbances than reported by patients 34,35 . These limitations highlight the need for complementary biomarkers to improve diagnostic accuracy. Increased nocturnal activity, as reflected by elevated PIM, TAT, and ZCM values, suggests that chronic insomnia patients experience more frequent or prolonged nighttime movements, potentially linked to physiological hyperarousal 36 . Although actigraphy is increasingly recognized as a valuable tool for evaluating circadian rhythm integrity 37,38 , our actigraphy-derived measures did not reveal significant differences in circadian robustness among chronic insomnia patients. However, a strong negative correlation between ISI scores and circadian robustness (IS and CFI) suggests that greater insomnia severity is associated with weaker circadian function. We observed alterations in body temperature rhythms, with chronic insomnia patients with insomnia patients exhibiting attenuated amplitude and lower absolute values of wrist and axillary temperature. Given that skin temperature serves as both a marker and modulator of circadian rhythms 39,40 , these findings suggest possible circadian misalignment. Prior studies have linked skin temperature dysregulation to insomnia symptoms, including difficulties with sleep onset and maintenance 16 41 . While some research suggests impaired thermoregulation as a contributing factor in certain insomnia subtypes, findings remain inconsistent, warranting further investigation 18,42 . Elevated nighttime cortisol, a well-documented hallmark of hypothalamic-pituitary-adrenal (HPA) axis dysregulation 43–46 , commonly observed in insomnia. In our study, ISSD patients exhibited significantly higher nocturnal cortisol levels compared to controls, but this effect was not observed in INSD patients, reinforcing the notion that hyperarousal is more pronounced in ISSD 7 . Prior research has reported positive correlations between nighttime cortisol and insomnia severity (ISI scores) 47 , further supporting HPA axis dysregulation in insomnia. However, cortisol levels may not always be significantly elevated in insomnia 48,49 , which can be partly explained by the heterogeneity of insomnia subtypes and potential confounding factors such as medication use, highlighting the need for further research into the differential pathophysiology of ISSD and INSD. Our study reveals significant alterations in clock gene expression in chronic insomnia patients, particularly in ISSD. PER1 and PER2 expression peaked during wakefulness and light exposure, while BMAL1 , REV-ERBα , and REV-ERBβ showed modest but significant diurnal expression variations. These findings suggest that insomnia-associated circadian disruptions extend beyond behavioral and hormonal dysregulation to the molecular level. Previous studies have linked polymorphisms in clock genes, particularly in CLOCK and PER3 , to insomnia risk 50–53 . However, research on clock gene expression patterns in chronic insomnia patients is limited. Prior work has demonstrated that acute sleep deprivation can disrupt PER2 and BMAL1 rhythms, indicating that sleep loss influences clock gene dynamics 54 . Additionally, studies in sleep apnea patients have reported reversible alterations in clock gene expression following continuous positive airway pressure (CPAP) treatment, further supporting the role of clock gene expression as a potential biomarker for sleep disorders 24 . Our findings align with these observations, as PER1 and PER2 expression was significantly affected in insomnia patients, particularly in ISSD. This reinforces their potential as sensitive markers of sleep deprivation and circadian misalignment. However, it remains unclear whether these molecular disruptions contribute to insomnia onset or result from chronic sleep disturbances. Longitudinal studies are needed to clarify causality and further explore the molecular mechanisms underlying insomnia subtypes. Our findings highlight the intricate relationship between chronic insomnia and circadian dysregulation across molecular, physiological, and behavioral levels. Further supporting this, we show that ISI is negatively correlated with BMAL , CLOCK , PER2 , CRY1-2 , and cortisol, with more pronounced disruptions in ISSD and INSD subgroups, reinforcing the role of molecular circadian alterations in insomnia pathology. Additionally, ISI positively correlated with axillary temperature and PER1 expression, linking sleep disturbances to thermoregulatory and circadian disruptions. Actigraphy-based circadian parameters also reflected these disruptions, with CFI correlating strongly with CLOCK , PER1 , PER2 , CRY1 , and CRY2 expression. INSD patients exhibited distinct alterations, including negative correlations between axillary temperature and ISI, RA, and CFI, suggesting impaired thermoregulation and weakened circadian stability. The association between TST and nighttime cortisol, as well as PER2 and REV-ERBβ further supports a link between hyperarousal, sleep duration, and molecular rhythms. Additionally, age correlated positively with BMAL , CLOCK , PER1 , CRY1 , and REV-ERBα expression, suggesting that aging may exacerbate circadian misalignment in insomnia. Currently, chronic insomnia disorder diagnosis relies primarily on self-reported symptoms, with objective assessments used only in select cases 9 . Our study suggests that blood-based biomarkers analysis, combined with machine-learning, may provide a cost-effective alternative for identifying chronic insomnia cases and distinguishing between subtypes. Supervised machine-learning of clock gene expression profiles successfully differentiated chronic insomnia patients from control and further distinguished ISSD from INSD patients. Machine learning algorithms identified PER2 , CRY1 and BMAL as key classifiers to distinguishing insomnia patients from controls and CLOCK , BMAL and PER2 for differentiating between insomnia subtypes. These findings highlight the potential of clock genes as a diagnostic biomarker and a tool for patient stratification. In summary, our study provides further evidence that chronic insomnia, particularly ISSD, is associated with circadian dysfunction at behavioral, hormonal, and molecular levels. Altered clock gene expression, disrupted body temperature rhythms, and elevated nighttime cortisol underscore the role of circadian dysregulation in insomnia pathology. While actigraphy showed limited sensitivity in detecting circadian alterations, skin temperature changes and clock gene expression patterns emerged as promising markers of circadian dysfunction in insomnia. Moreover, while other studies have applied machine learning to assess circadian function 24,55 , our study highlights the potential of machine learning approaches based on clock gene expression for identifying chronic insomnia cases and stratifying patients by subtype. Future research should validate these findings in larger cohorts, investigate causal mechanisms, and develop target interventions, particularly for ISSD patients, who exhibit more severe circadian disruptions. Recognizing ISSD and subtyping insomnia is important for personalized therapeutic approaches and should be integrated into clinical practice for more effective management. Study Limitations Although our sample size was relatively small, the robustness of our findings support the application of machine learning methodologies. The rigorous eligibility criteria and the stringent participant selection process help mitigate concerns related to sample size limitations. However, we did not assess the potential impact of individual clinical histories on clock gene expression. Nevertheless, the absence of predominant comorbidities within any study group suggests that this may not represent a significant confounder. Additionally, intrinsic inter-individual variability must be considered, as it may influence molecular and cellular responses to disease and contribute to the observed heterogeneity in outcomes. Despite this, the randomized and counterbalanced number of conditions across participants reduces the potential for systematic bias. We did not specifically analyze the seasonal effects on clock gene expression, as patient recruitment occurred year-round. However, a previous study by our group with a similar design found no significant seasonal influence 24 , suggesting minimal impact. Another consideration is the use of at-home sleep tests in healthy subjects rather than a type II PSG, which was impractical for routine clinical assessments due to cost, time demands, labor-intensive procedures, and long waiting lists 13 . To overcome these issues, we used the WatchPAT, a widely adopted sleep-screening tool, for the healthy subject cohort 56,57 . However, it is important to note that the use of objective sleep measures remain optional in insomnia diagnosis 9 . Our study establishes a foundational framework for understanding the role of clock outputs and clock gene expression in chronic insomnia and its subtypes. However, given the complexity of clock-regulated signaling pathways, further research is needed to elucidate how circadian clocks interact with neuronal circuits and systemic physiology to better understand the pathophysiology of chronic insomnia. Declarations Acknowledgments The authors are thankful to Francieli Ruiz for supplying the actigraphs used in this study and to Gasoxmed, LDA. Portugal for providing the WatchPAT® and respective probes for this study. Funding This study was financed by the European Regional Development Fund (ERDF), through Centro 2020 Regional Operational Programme COMPETE 2020 - Operational Programme for Competitiveness and Internationalization and Portuguese national funds via Fundação para a Ciência e a Tecnologia (FCT; UIDB/04539/2020, UIDP/04539/2020 and LA/P/0058/2020), and the European Social Fund through the Human Capital Operational Programme and Portuguese national funds via FCT under 2020.04499.BD, PD/BD/135497/2018, 2020.04850.BD, 2021.05334.BD and under COVID/BD/152507/2022. This work was supported by Mapfre, with the project "InsomniaO'clock" - reference IT137-24-400. Author contributions CCA, CC and ARA conceived the study plan and design; CC and ARA supervised the work; CCA conducted the study and performed the sample processing, experiments for clock outputs evaluation and gene expression analysis, and PSG and actigraphy-derived data analysis; BS assisted in experiments, LSG and AS assisted in analysis and discussion, RR assisted with actigraphy evaluation; JM and JS recruited the patients; JS evaluated patients and all the necessary clinical data; MF gathered all clinical data from all patients; JA and TG were responsible for the design of the computational analysis. JA carried out the bioinformatics rhythmicity analysis in gene expression datasets and performed machine learning-based analysis with the obtained study datasets; TD analyzed the raw actigraphy data. CCA wrote the first draft of the manuscript. All co-authors reviewed the manuscript and approved the final version to be submitted. Competing interests The authors declare no competing interests. References Morin CM, Jarrin DC. Epidemiology of Insomnia: Prevalence, Course, Risk Factors, and Public Health Burden. Sleep Med Clin . 2022;17(2):173-191. doi:10.1016/J.JSMC.2022.03.003 American Academy of Sleep Medicine. International Classification of Sleep Disorders, 3rd Edition, Text Revision (ICSD-3-TR) .; 2023. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, Text Revision (DSM-5-TR). American Psychiatric Association Publishing; 2022. doi:10.1176/appi.books.9780890425787 Bjorvatn B, Jernelöv S, Pallesen S. Insomnia – a heterogenic disorder often comorbid with psychological and somatic disorders and diseases: A narrative review with focus on diagnostic and treatment challenges. Front Psychol . 2021;12:289. doi:10.3389/fpsyg.2021.639198 Blanken TF, Benjamins JS, Borsboom D, et al. Insomnia disorder subtypes derived from life history and traits of affect and personality. The lancet Psychiatry . 2019;6(2):151-163. doi:10.1016/S2215-0366(18)30464-4 Perlis ML, Giles DE, Mendelson WB, Bootzin RR, Wyatt JK. Psychophysiological insomnia: the behavioural model and a neurocognitive perspective. J Sleep Res . 1997;6(3):179-188. doi:10.1046/j.1365-2869.1997.00045.x Vgontzas A, Fernandez-Mendoza J, Liao D, Bixler E. Insomnia with objective short sleep duration: The most biologically severe phenotype of the disorder. Sleep Med Rev . 2013;17(4):241-254. doi:10.1016/j.smrv.2012.09.005 Rezaie L, Fobian AD, McCall WV, Khazaie H. Paradoxical insomnia and subjective–objective sleep discrepancy: A review. Sleep Med Rev . 2018;40:196-202. doi:10.1016/J.SMRV.2018.01.002 Riemann D, Espie CA, Altena E, et al. The European Insomnia Guideline: An update on the diagnosis and treatment of insomnia 2023. J Sleep Res . 2023;32(6):e14035. doi:10.1111/JSR.14035 Edinger JD, Fins AI, Goeke JM, et al. The empirical identification of insomnia subtypes: a cluster analytic approach. Sleep . 1996;19(5):398-411. Vgontzas A, Liao D, Pejovic S, Calhoun S, Karataraki M, Bixler EO. Insomnia with objective short sleep duration is associated with type 2 diabetes: A population-based study. Diabetes Care . 2009;32(11):1980-1985. doi:10.2337/dc09-0284 Gaines J, Vgontzas A, Fernandez-Mendoza J, Bixler EO. Obstructive sleep apnea and the metabolic syndrome: The road to clinically-meaningful phenotyping, improved prognosis, and personalized treatment. Sleep Med Rev . 2018;42:211-219. doi:10.1016/j.smrv.2018.08.009 Chiner E. Approach To The Cost Of Polysomnography In A Spanish Hospital. Internet J Pulm Med . 2012;2(2). doi:10.5580/1789 Baglioni C, Regen W, Teghen A, et al. Sleep changes in the disorder of insomnia: A meta-analysis of polysomnographic studies. Sleep Med Rev . 2014;18(3):195-213. doi:10.1016/j.smrv.2013.04.001 Lack LC, Micic G, Lovato N. Circadian aspects in the aetiology and pathophysiology of insomnia. J Sleep Res . 2023;32(6). doi:10.1111/jsr.13976 Morris M, Lack L, Dawson D. Sleep-Onset Insomniacs Have Delayed Temperature Rhythms. Sleep . 1990;13(1):1-14. doi:10.1093/sleep/13.1.1 Wright KP, Hull JT, Hughes RJ, Ronda JM, Czeisler CA. Sleep and Wakefulness Out of Phase with Internal Biological Time Impairs Learning in Humans. J Cogn Neurosci . 2006;18(4):508-521. doi:10.1162/jocn.2006.18.4.508 Lack LC, Gradisar M, Van Someren EJW, Wright HR, Lushington K. The relationship between insomnia and body temperatures. Sleep Med Rev . 2008;12(4):307-317. doi:10.1016/j.smrv.2008.02.003 Flynn-Evans EE, Shekleton JA, Miller B, et al. Circadian Phase and Phase Angle Disorders in Primary Insomnia. Sleep . 2017;40(12). doi:10.1093/sleep/zsx163 Takimoto M, Hamada A, Tomoda A, et al. Daily expression of clock genes in whole blood cells in healthy subjects and a patient with circadian rhythm sleep disorder. Am J Physiol Integr Comp Physiol . 2005;289(5):R1273-R1279. doi:10.1152/ajpregu.00126.2005 Burioka N, Fukuoka Y, Takata M, et al. Circadian rhythms in the CNS and peripheral clock disorders: function of clock genes: influence of medication for bronchial asthma on circadian gene. J Pharmacol Sci . 2007;103(2):144-149. doi:10.1254/jphs.FMJ06003X4 Fukuya H, Emoto N, Nonaka H, Yagita K, Okamura H, Yokoyama M. Circadian expression of clock genes in human peripheral leukocytes. Biochem Biophys Res Commun . 2007;354(4):924-928. doi:10.1016/j.bbrc.2007.01.063 Takata M, Burioka N, Ohdo S, et al. Daily Expression of mRNAs for the Mammalian Clock Genes Per2 and Clock in Mouse Suprachiasmatic Nuclei and Liver and Human Peripheral Blood Mononuclear Cells. Jpn J Pharmacol . 2002;90(3):263-269. doi:10.1254/jjp.90.263 Gaspar LS, Hesse J, Yalçin M, et al. Long-term continuous positive airway pressure treatment ameliorates biological clock disruptions in obstructive sleep apnea. EBioMedicine . 2021;65(103248):103248. doi:10.1016/j.ebiom.2021.103248 Zhang H, Sun H, Li J, Fan Y, Jülich ST, Lei X. Subtypes of insomnia revealed by the heterogeneity of neuroanatomical patterns: A structural MRI study. Biol Psychol . 2023;180:108591. doi:10.1016/j.biopsycho.2023.108591 Clemente V, Ruivo Marques D, Miller-Mendes M, Morin CM, Serra J, Allen Gomes A. The European Portuguese version of the insomnia severity index. J Sleep Res . 2021;30(1). doi:10.1111/JSR.13198 Sargento P, Perea V, Ladera V, Lopes P, Oliveira J. The Epworth Sleepiness Scale in Portuguese adults: from classical measurement theory to Rasch model analysis. Sleep Breath . 2015;19(2):693-701. doi:10.1007/s11325-014-1078-6 Reis R, Teixeira F, Martins V, et al. Validation of a Portuguese version of the STOP-Bang questionnaire as a screening tool for obstructive sleep apnea: Analysis in a sleep clinic. Pulmonology . 2015;21(2):61-68. doi:10.1016/J.RPPNEN.2014.04.009 Pais-Ribeiro J, Silva I, Ferreira T, Martins A, Meneses R, Baltar M. Validation study of a Portuguese version of the Hospital Anxiety and Depression Scale. Psychol Health Med . 2007;12(2):225-237. doi:10.1080/13548500500524088 Fekedulegn D, Andrew ME, Shi M, Violanti JM, Knox S, Innes KE. Actigraphy-Based Assessment of Sleep Parameters. Ann Work Expo Heal . 2020;64(4):350-367. doi:10.1093/annweh/wxaa007 Livak KJ, Schmittgen TD. Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2−ΔΔCT Method. Methods . 2001;25(4):402-408. doi:10.1006/METH.2001.1262 Lück S, Thurley K, Thaben PF, Westermark PO. Rhythmic Degradation Explains and Unifies Circadian Transcriptome and Proteome Data. Cell Rep . 2014;9(2):741-751. doi:10.1016/j.celrep.2014.09.021 Taibi DM, Landis CA, Vitiello M V. Concordance of Polysomnographic and Actigraphic Measurement of Sleep and Wake in Older Women with Insomnia. J Clin Sleep Med . 2013;9(3):217-225. doi:10.5664/JCSM.2482 Frase L, Nissen C, Spiegelhalder K, Feige B. The importance and limitations of polysomnography in insomnia disorder—a critical appraisal. J Sleep Res . 2023;32(6). doi:10.1111/jsr.14036 Feige B, Al‐Shajlawi A, Nissen C, et al. Does REM sleep contribute to subjective wake time in primary insomnia? A comparison of polysomnographic and subjective sleep in 100 patients. J Sleep Res . 2008;17(2):180-190. doi:10.1111/j.1365-2869.2008.00651.x Blackwell T, Redline S, Ancoli-Israel S, et al. Comparison of Sleep Parameters from Actigraphy and Polysomnography in Older Women: The SOF Study. Sleep . 2008;31(2):283-291. doi:10.1093/sleep/31.2.283 Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, Pollak CP. The Role of Actigraphy in the Study of Sleep and Circadian Rhythms. Sleep . 2003;26(3):342-392. doi:10.1093/sleep/26.3.342 Fossion R, Rivera AL, Toledo-Roy JC, Ellis J, Angelova M. Multiscale adaptive analysis of circadian rhythms and intradaily variability: Application to actigraphy time series in acute insomnia subjects. PLoS One . 2017;12(7):e0181762. doi:10.1371/JOURNAL.PONE.0181762 Martinez-Nicolas A, Ortiz-Tudela E, Rol MA, Madrid JA. Uncovering Different Masking Factors on Wrist Skin Temperature Rhythm in Free-Living Subjects. PLoS One . 2013;8(4):e61142. doi:10.1371/JOURNAL.PONE.0061142 Cuesta M, Boudreau P, Cermakian N, Boivin DB. Skin Temperature Rhythms in Humans Respond to Changes in the Timing of Sleep and Light. J Biol Rhythms . 2017;32(3):257-273. doi:10.1177/0748730417702974 Lack LC, Mercer JD, Wright HR. Circadian rhythms of early morning awakening insomniacs. J Sleep Res . 1996;5(4):211-219. doi:10.1111/j.1365-2869.1996.00211.x Iijima R, Kadooka A, Sugawara K, Fushimi M, Hosoe M, Aritake-Okada S. Subjective sleep onset latency is influenced by sleep structure and body heat loss in human subjects. J Sleep Res . 2024;33(5):e14122. doi:10.1111/jsr.14122 Rodenbeck A, Huether G, Rüther E, Hajak G. Interactions between evening and nocturnal cortisol secretion and sleep parameters in patients with severe chronic primary insomnia. Neurosci Lett . 2002;324(2):159-163. doi:10.1016/S0304-3940(02)00192-1 Vgontzas A, Bixler EO, Lin H-M, et al. Chronic Insomnia Is Associated with Nyctohemeral Activation of the Hypothalamic-Pituitary-Adrenal Axis: Clinical Implications. J Clin Endocrinol Metab . 2001;86(8):3787-3794. doi:10.1210/jcem.86.8.7778 Dressle RJ, Feige B, Spiegelhalder K, et al. HPA axis activity in patients with chronic insomnia: A systematic review and meta-analysis of case–control studies. Sleep Med Rev . 2022;62:101588. doi:10.1016/j.smrv.2022.101588 Zhang X-X, Sun S-Y, Ma Z-J, et al. Changed nocturnal levels of stress-related hormones couple with sleep-wake states in the patients with chronic insomnia disorder: A clinical pilot study. Sleep Med . 2024;117:177-183. doi:10.1016/j.sleep.2024.03.017 Passos GS, Youngstedt SD, Rozales AARC, et al. Insomnia Severity is Associated with Morning Cortisol and Psychological Health. Sleep Sci . 2023;16(01):092-096. doi:10.1055/s-0043-1767754 Riemann D, Klein T, Rodenbeck A, et al. Nocturnal cortisol and melatonin secretion in primary insomnia. Psychiatry Res . 2002;113(1-2):17-27. doi:10.1016/S0165-1781(02)00249-4 Rösler L, van Kesteren E-J, Leerssen J, et al. Hyperarousal dynamics reveal an overnight increase boosted by insomnia. J Psychiatr Res . 2024;179:279-285. doi:10.1016/j.jpsychires.2024.09.032 Serretti A, Benedetti F, Mandelli L, et al. Genetic dissection of psychopathological symptoms: insomnia in mood disorders and CLOCK gene polymorphism. Am J Med Genet B Neuropsychiatr Genet . 2003;121B(1):35-38. doi:10.1002/AJMG.B.20053 Lee S. Association between CLOCK Gene Polymorphisms and Insomnia Risk According to Food Groups: A KoGES Longitudinal Study. Nutrients . 2023;15(10):2300. doi:10.3390/NU15102300/S1 Semenova N V., Madaeva IM, Bairova TA, Zhambalova RM, Sholokhov LF, Kolesnikova LI. Association of the melatonin circadian rhythms with clock 3111T/C gene polymorphism in Caucasian and Asian menopausal women with insomnia. Chronobiol Int . 2018;35(8):1066-1076. doi:10.1080/07420528.2018.1456447 Semenova N V., Madaeva IM, Bairova TI, et al. 3111T/C Clock Gene Polymorphism in Women with Insomnia. Bull Exp Biol Med . 2017;163(4):461-464. doi:10.1007/s10517-017-3828-5 Kavčič P, Rojc B, Dolenc-Grošelj L, Claustrat B, Fujs K, Poljak M. The impact of sleep deprivation and nighttime light exposure on clock gene expression in humans. Croat Med J . 2011;52(5):594-603. doi:10.3325/cmj.2011.52.594 Hesse J, Malhan D, Yalçin M, Aboumanify O, Basti A, Relógio A. An Optimal Time for Treatment—Predicting Circadian Time by Machine Learning and Mathematical Modelling. Cancers 2020, Vol 12, Page 3103 . 2020;12(11):3103. doi:10.3390/CANCERS12113103 Phua CQ, Jang IJ, Tan KB, et al. Reducing cost and time to diagnosis and treatment of obstructive sleep apnea using ambulatory sleep study: a Singapore sleep centre experience. Sleep Breath . 2021;25(1):281-288. doi:10.1007/s11325-020-02115-z Hamida ST-B, Ahmed B, Cvetkovic D, Jovanov E, Kennedy G, Penzel T. A New Era in Sleep Monitoring: The Application of Mobile Technologies in Insomnia Diagnosis. In: ; 2015:101-127. doi:10.1007/978-3-319-12817-7_5 Tables Tables are available in the Supplementary Files section. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files Supplementalmaterial.docx Supplemental material Tables.docx Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6268489","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":453631030,"identity":"b9a566cc-d8df-415b-98d0-9e8bd88353c1","order_by":0,"name":"Ana Rita Alvaro","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAu0lEQVRIiWNgGAWjYDACZhBRISFHihaQnjM2xnABHqKsYWxLS2wgWot5O//BDz/OHE7vn938gJmnhiHPnpAWmcPMzJI9FYdzZ9w5ZsDMc4yhmKAtEkC/SPCcOZy7QSKHgZm3gSGxhwgtzD//th1ONyBFC5s0b1taAklazKxlztgYzriRZnBwzjGJYp4DhLTwH3x8802FhDz/jOSHD97U2OSxNxCyBhkAzZdIIEUDBJChZRSMglEwCoY7AADloTVMWjWErgAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-2387-374X","institution":"University of Coimbra","correspondingAuthor":true,"prefix":"","firstName":"Ana","middleName":"Rita","lastName":"Alvaro","suffix":""},{"id":453631031,"identity":"2dfdb441-6638-4a8b-9d59-977b4dacf505","order_by":1,"name":"Catarina Carvalhas-Almeida","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Catarina","middleName":"","lastName":"Carvalhas-Almeida","suffix":""},{"id":453631032,"identity":"9a7325f4-8cc3-4eeb-ba38-e65b986af942","order_by":2,"name":"João Alves","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"João","middleName":"","lastName":"Alves","suffix":""},{"id":453631033,"identity":"ee0e33bc-e1f0-4048-8959-ac72609b349b","order_by":3,"name":"Tiago Davi","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Tiago","middleName":"","lastName":"Davi","suffix":""},{"id":453631034,"identity":"fa25edda-b2fd-43dc-98f2-cc436c3ded87","order_by":4,"name":"Bárbara Santos","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Bárbara","middleName":"","lastName":"Santos","suffix":""},{"id":453631035,"identity":"6de77fba-23b9-4f2b-9fa1-31d2e6499d9a","order_by":5,"name":"Laetitia Gaspar","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Laetitia","middleName":"","lastName":"Gaspar","suffix":""},{"id":453631036,"identity":"a8ddb08a-d947-4ec2-b90e-684481b991de","order_by":6,"name":"Rodrigo F.N. Ribeiro","email":"","orcid":"https://orcid.org/0000-0002-5159-4406","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Rodrigo","middleName":"F.N.","lastName":"Ribeiro","suffix":""},{"id":453631037,"identity":"2b84d3a0-3317-4fd1-a5b1-1db91988c25d","order_by":7,"name":"Joana Serra","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Joana","middleName":"","lastName":"Serra","suffix":""},{"id":453631038,"identity":"af5633cf-1a51-4e53-a07b-7ed842564f47","order_by":8,"name":"Mafalda Ferreira","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Mafalda","middleName":"","lastName":"Ferreira","suffix":""},{"id":453631039,"identity":"b776c66f-f78a-4606-b9f2-302081f445ad","order_by":9,"name":"Joaquim Moita","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Joaquim","middleName":"","lastName":"Moita","suffix":""},{"id":453631040,"identity":"e4268f23-a27f-414f-bd89-b4d02b3b5b26","order_by":10,"name":"Amita Sehgal","email":"","orcid":"https://orcid.org/0000-0001-7354-9641","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Amita","middleName":"","lastName":"Sehgal","suffix":""},{"id":453631041,"identity":"2b9af78f-377f-4036-8772-322ae85d796c","order_by":11,"name":"Claudia Cavadas","email":"","orcid":"https://orcid.org/0000-0001-8020-9266","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Claudia","middleName":"","lastName":"Cavadas","suffix":""}],"badges":[],"createdAt":"2025-03-20 10:07:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6268489/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6268489/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85346720,"identity":"5b3c255d-dfa6-4b40-8039-8d72090aaf69","added_by":"auto","created_at":"2025-06-25 02:13:12","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":590055,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStudy flow diagram and schematic of the study design.\u003c/strong\u003e (A) The study flow diagram outlines subject selection and eligibility assessment. Of the 45 patients initially evaluated, 13 healthy control subjects and 14 patients with chronic insomnia disorder met the inclusion criteria. The chronic insomnia disorder group was further divided into two subgroups based on total sleep duration: 8 patients with insomnia with short sleep duration (ISSD) and 6 patients with insomnia with normal sleep duration (INSD). (B) Study design. All participants underwent a one-night polysomnography (PSG) and completed two weeks of actigraphy monitoring before PSG evaluation. Data collection included demographic and anthropometric measures (age, BMI), lifestyle and clinical history information, and sleep-related questionnaires. Additionally, body temperature was monitored, and blood samples were collected at four time points (8:00, 11:00, 15:30, and 21:30 h) for circadian marker analysis. Machine learning algorithms were applied to cluster groups based on clock gene expression profiles.\u003c/p\u003e","description":"","filename":"figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/efd60dbfe1770d7721b696f2.png"},{"id":85346721,"identity":"c2b7ed3f-71f0-481b-8669-82a20474b6a6","added_by":"auto","created_at":"2025-06-25 02:13:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1533619,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRest:activity patterns, daily activity and temperature profiles assessed by actigraphy in\u003c/strong\u003e \u003cstrong\u003ehealthy control subjects and patients with chronic Insomnia, stratified by short (ISSD) and normal (INSD) sleep duration.\u003c/strong\u003e(A) Representative actograms illustrate rest:activity patterns over two weeks in control, ISSD and INSD groups. Double-plotted 24-h actograms generated from participants’ rest:activity patterns reflect the characteristic differences detected between study groups with blue peaks representing activity (PIM), light blue sleeping states, light green resting states, light yellow for light phases, light grey for dark phases, and purple for off-wrist periods. (B) Representative 24-hour actograms of the PIM activity for a healthy control subject, a patient with ISSD, and a patient with INSD. (C-D) Group-averaged 24-hour profiles of PIM activity and temperature rhythms over 14 days. Graphs are plotted as mean ± SEM. N = 6-14 per group. (E) Correlation of Insomnia severity index (ISI) score with circadian parametric and non-parametric measures, indicating a decline in circadian robustness with increasing insomnia severity. Colors and dot sizes represent the strength and direction of the correlations, evaluated using Spearman’s test (*p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001), N = 14.\u003c/p\u003e","description":"","filename":"figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/dc02e10cf61363cda4e4f95d.png"},{"id":85346723,"identity":"315b8e87-7612-4577-baa8-52e24cf6c698","added_by":"auto","created_at":"2025-06-25 02:13:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":317432,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eChronic Insomnia alters axillary temperature, plasma cortisol levels and profiles.\u003c/strong\u003e (A) Axillary temperature and plasma cortisol levels were measured at four time points (8:00, 11:00, 15:30, and 21:30 h) in patients with chronic insomnia and controls (grey line). (B-C) Data were further analyzed in subgroups: insomnia with short sleep duration (ISSD) and insomnia with normal sleep duration (INSD). Graphs represent mean ± SEM. Significant differences relative to the control group at each timepoint were assessed using an unpaired parametric t-test (*p \u0026lt; 0.05, **p \u0026lt; 0.01). N = 6–14 per group.\u003c/p\u003e","description":"","filename":"figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/57beb7b2916595251e470db3.png"},{"id":85346727,"identity":"c21b7ffa-9727-49ba-a9e6-32df48007044","added_by":"auto","created_at":"2025-06-25 02:13:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":364211,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eChronic Insomnia alters the expression of core-clock genes in peripheral blood mononuclear cells (PBMCs), with more pronounced alterations in patients with short sleep duration (ISSD) patients.\u003c/strong\u003e (A) Expression levels of nine clock genes (\u003cem\u003eBMAL1, CLOCK, PER1–3, CRY1–2, REV-ERBα, and REV-ERBβ\u003c/em\u003e) were measured at four time points (8, 11, 15:30, and 21:30 h) in PBMCs of chronic insomnia patients and healthy controls(grey line). Data were further analyzed in subgroups: (B) insomnia with short sleep duration (ISSD) and (C) insomnia with normal sleep duration (INSD). Data are presented as mean ΔΔCT values ± SEM. Significant differences relative to controls at each timepoint were determined using an unpaired parametric t-test (*p \u0026lt; 0.05, **p \u0026lt; 0.01, *** \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001). N = 6–14 per group.\u003c/p\u003e","description":"","filename":"figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/20b480bc07ff11b413cccf65.png"},{"id":85346736,"identity":"b9ca1c81-b524-45d0-89b5-de6b1e90bd80","added_by":"auto","created_at":"2025-06-25 02:13:13","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2804292,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelation analyses of sleep metrics, circadian parameters, clinical symptom scores, clock outputs, and core-clock gene expression in chronic insomnia and its subgroups: insomnia with short sleep duration (ISSD) and insomnia with normal sleep duration (INSD). \u003c/strong\u003eVariables include\u003cstrong\u003e \u003c/strong\u003ebody mass index (BMI), Epworth sleepiness scale, insomnia severity index (ISI), total sleep time (TST), rapid eye movement (REM), wakefulness after sleep onset (WASO), relative amplitude (RA), intradaily variability (IV), intradailty stability (IS) and circadian function index (CFI).\u003cstrong\u003e \u003c/strong\u003e(A-C) Dot plots highlight significant correlations at different time points (8:00, 11:00, 15:30, and 21:30 h) in chronic insomnia patients (A), ISSD (B), and INSD (C). Colors and dot sizes represent correlation strength and direction, assessed using Spearman’s test (*p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001). N = 6–14 per group.\u003c/p\u003e","description":"","filename":"figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/76672155fe1cbd092df63893.png"},{"id":85346741,"identity":"f9f5e680-0ec7-4584-88b7-778472af82e7","added_by":"auto","created_at":"2025-06-25 02:13:14","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":426182,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDifferential expression of core clock genes in peripheral blood mononuclear cells (PBMCs) in chronic insomnia patients.\u003c/strong\u003e Chronic insomnia patients exhibit distinct mean expression profiles of core-clock genes in PBMCs compared to control subjects. Significant differences are also observed between chronic insomnia subgroups with short sleep duration (ISSD) and normal sleep duration (INSD). Subjects were classified based on the mean expression levels of specific clock genes of thefour time points across the day. (A) Classification of control subjects and chronic insomnia patients based on the mean mRNA expression levels of \u003cem\u003ePER1\u003c/em\u003e, \u003cem\u003eCRY2\u003c/em\u003e, and \u003cem\u003eBMAL1\u003c/em\u003e. (B) Classification of ISSD and INSD subgroups based on the mean mRNA expression levels of \u003cem\u003ePER2\u003c/em\u003e, \u003cem\u003eCLOCK\u003c/em\u003e, and \u003cem\u003eBMAL1\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/371370e3f909e0870be6afb7.png"},{"id":85348797,"identity":"5eeeb28c-a39a-4555-a338-3482fba91c1f","added_by":"auto","created_at":"2025-06-25 02:29:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":7308729,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/09b9ea80-8ff3-4652-9352-311e6584f827.pdf"},{"id":85346724,"identity":"6a56ddc2-8f9f-4ea4-9632-d8dabc66f87f","added_by":"auto","created_at":"2025-06-25 02:13:13","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1127251,"visible":true,"origin":"","legend":"Supplemental material","description":"","filename":"Supplementalmaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/8cdd4c47f86ed8048f93639a.docx"},{"id":85346722,"identity":"c12b02e3-4659-478f-9158-010161aa60b5","added_by":"auto","created_at":"2025-06-25 02:13:13","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":20227,"visible":true,"origin":"","legend":"","description":"","filename":"Tables.docx","url":"https://assets-eu.researchsquare.com/files/rs-6268489/v1/7d467093bf4f86629cfc1e88.docx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Clock gene signature predicts Insomnia and links to sleep/circadian parameters","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChronic insomnia disorder is a prevalent condition affecting approximately 10% of the global population \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. It is defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, Text Revision (DSM-5-TR) and the International Classification of Sleep Disorders, 3rd edition, Text Revision (ICSD-3-TR) as persistent nighttime symptoms despite adequate sleep opportunities, accompanied by daytime impairments such as fatigue, occurring at least three nights \u003cem\u003eper\u003c/em\u003e week for over three months \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Diagnosis typically relies on self-reported symptoms, limiting objective assessment and standardization. Over time, varying definitions of insomnia and its subtypes, such as primary insomnia or classifications based on sleep duration or individual traits, have led to inconsistencies in research findings and diagnostic challenges \u003csup\u003e\u003cspan additionalcitationids=\"CR5 CR6 CR7\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, hindering the identification of reliable biomarkers \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eA clinically relevant subtype, insomnia with short sleep duration (ISSD), is characterized by a total sleep time (TST) of less than six hours, as measured by single-night polysomnography (PSG) \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. ISSD is associated with an increased risk of significant medical and psychiatric comorbidities, adverse cardiometabolic outcomes and neurocognitive impairments \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Moreover, individuals with ISSD respond less favorably to cognitive-behavioral therapy compared to those with insomnia with normal sleep duration (INSD, TST\u0026thinsp;\u0026gt;\u0026thinsp;6 hours) \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e, highlighting the need for subtyping insomnia during diagnosis for tailored treatment approaches.\u003c/p\u003e \u003cp\u003eDespite the clinical relevance of ISSD and INSD, objective measures such as PSG and actigraphy are rarely incorporated into routine diagnosis of insomnia due to high costs, logistical challenges \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e, and their inability to capture subjective wakefulness experiences \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. Additionally, European guidelines do not mandate objective sleep measures for diagnosing insomnia \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. To address these challenges, identifying reliable biological markers could enhance diagnostic precision and inform targeted treatment strategies.\u003c/p\u003e \u003cp\u003eEmerging evidence suggests that chronic insomnia may disrupt the circadian clock system, though the extent of these effects remains unclear \u003csup\u003e\u003cspan additionalcitationids=\"CR16 CR17 CR18\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Previous studies have proposed the analysis of clock gene expression profiles in human peripheral blood mononuclear cells (PBMCs) as a promising approach for assessing circadian rhythm \u003csup\u003e\u003cspan additionalcitationids=\"CR21 CR22\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e.. Indeed, our prior study has shown through machine learning analysis, that patients with obstructive sleep apnea (OSA), another highly prevalent sleep disorder, can be distinguished from healthy controls based on altered expression patterns of several clock genes \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, suggesting that clock gene expression may serve as potential biomarker for sleep disorders. In recent years, machine learning approaches have increasingly been utilized to identify biomarkers for different sleep disorders and to monitor treatment responses \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eBuilding on this concept, we investigated the impact of chronic insomnia on the biological clock by analyzing key physiological markers, including plasma cortisol, body temperature, and clock gene expression in PBMCs. We also characterized sleep using PSG, assessed actigraphy-derived circadian parameters, and machine learning-based analysis for identifying biomarkers that enable robust differentiation of insomnia patients from controls and between subtypes.\u003c/p\u003e \u003cp\u003eOur findings reveal significant circadian disruptions and altered clock gene expression in chronic insomnia, particularly in ISSD, underscoring the potential of circadian biomarkers for improved diagnosis and treatment. Machine learning analysis enabled robust differentiation between insomnia patients and controls, as well as between ISSD and INSD subtypes, based on clock gene expression. Overall, this work advances our understanding of insomnia\u0026rsquo;s circadian disruption and highlights the utility of integrating circadian biomarkers with interpretable machine learning techniques to improve insomnia diagnosis, optimize treatment strategies based on subtype, and reduce the societal burden of insomnia, including work absenteeism and healthcare costs.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch3\u003e\u003cem\u003eEthical Approval\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eThis study received approval from the Ethical Committees of the Faculty of Medicine, University of Coimbra (CE-162/2021) and the Coimbra Hospital and University Centre (ULSC, OBS.SF.55-2021) in Coimbra, Portugal. All experimental procedures adhered to the ethical guidelines and regulations established in the 1964 Declaration of Helsinki and its subsequent amendments. The study also complied with the provisions of Regulation (EU) 2016/679 of the European Parliament and Council, governing the General Data Protection Regulation (GDPR), as well as Portuguese Law n.\u0026ordm; 12/2005 of January 26 and its implementing regulations detailed in Decree-Law n.\u0026ordm; 131/2014 of August 29, 2014. Informed consent was obtained from all participants prior to their inclusion in the study.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003e\u003cem\u003eStudy design\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eFemale and male adult volunteers (age \u0026ge; 18 years) suspected of having\u0026nbsp;chronic\u0026nbsp;insomnia\u0026nbsp;were recruited for a sleep study at the Sleep Medicine Unit of CHUC, Coimbra, Portugal, between January 2022 and September 2023. All selected subjects underwent a detailed screening process, which included a clinical interview conducted by an experienced psychiatrist, validated questionnaires and continuous wrist actigraphy monitoring over two-weeks (ActTrust\u0026reg; 2, Condor Instruments Ltda, SP, Brazil).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eParticipants identified as potential chronic insomnia cases were asked to undergo a single-night, home-based polysomnography (PSG) study (type II) to determine sleep duration subtype and were categorized into two groups: insomnia patients with short sleep duration (ISSD) (TST \u0026lt; 6 hours) and insomnia patients with normal sleep duration (INSD) (TST \u0026ge; 6 hours) \u003csup\u003e7\u003c/sup\u003e. PSG was also used to rule out other sleep disorders. Sleep studies were staged and scored according to the American Academy of Sleep Medicine (AASM) scoring manual v2.4, 2012.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHealthy volunteers were invited to participate in a one-night, home-based sleep study using the WatchPAT\u0026trade; 300 device (Itmar Medical, Gasoxmed, Portugal) to record their respiratory events and sleep parameters. Control subjects of similar age (\u0026plusmn; 5 years) and including both sexes, were classified as disease-free if met the following criteria: no lifetime history of significant insomnia-associated symptoms, TST \u0026ge; 6 hours, absence of medical or psychiatric comorbidities that could interfere with sleep (e.g. major depressive disorder) verified through clinical history and a structured clinical interview using the SCID-IV; no history of other sleep disorders as confirmed by clinical evaluation and screening PSG; no engagement in shiftwork within the preceding 6 months; no current use of prescription medications of over-the-counter products that could affect sleep; not being pregnant or lactating within the past 6 months.\u003c/p\u003e\n\u003cp\u003eData from the PSG test (PSG report) and WatchPAT\u0026trade; 300 (report) were obtained from each participant. An ID was assigned to each subject to ensure data confidentiality and traceability.\u003c/p\u003e\n\u003cp\u003eOn the morning following PSG or WatchPAT\u0026trade; 300 examination, participants completed the following validated questionnaires to assess various sleep and mental health parameters: Insomnia Severity Index (ISI) \u003csup\u003e26\u003c/sup\u003e, Epworth sleepiness scale \u003csup\u003e27\u003c/sup\u003e, STOP-BANG \u003csup\u003e28\u003c/sup\u003e and Hospital Anxiety and Depression Scale (HADS) \u003csup\u003e29\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDuring this period, individual data was collected from all participants, including sex, age, body mass index (BMI), lifestyle (sleep/wake routine, diet, meal schedules, physical exercise), and clinical history (comorbidities and medication).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBlood samples were collected from all subjects following their sleep studies during hospital visits. Samples were collected at four different times: in the morning (8:00 and 11:00 h), afternoon (15:30 h) and night (21:30 h). Blood was collected for the analysis of cortisol plasma levels, and clock genes expression in PBMCs. Axillary temperature was monitored at the same four time points. Between sampling sessions, participants were allowed to leave the Sleep Unit to continue their usual daily activities, ensuring minimal disruption to their routine. The study design is summarized in Figure\u0026nbsp;1.\u003c/p\u003e\n\u003ch3\u003eMeasuring rest:activity and skin temperature rhythms\u003c/h3\u003e\n\u003cp\u003eThe structure and timing of daily rest:activity patterns were assessed using wrist-worn ActTrust 2 actigraphs (Condor Instruments, SP, Brazil). Participants wore the devices continuously on their non-dominant wrist for 14 consecutive days under normal ambulatory conditions. During this period, participants were instructed to follow their usual routines and record their sleep patterns in a sleep diary. They were advised to remove the actigraphs during activities involving water (e.g. bathing, dishwashing or swimming). The actigraphs were configured with a sampling frequency of 25 Hz to optimize battery life over the recording period. The devices continuously recorded motor activity (using the Proportional Integral Mode, PIM), skin temperature (\u0026deg;C), and light exposure (lux) in one-minute epochs. PIM measures movement intensity by calculating the area under the signal curve for each epoch \u003csup\u003e30\u003c/sup\u003e. The raw actigraphy data was processed using the ActStudio software version 2.1.2 (Condor Instruments, SP, Brazil), which automatically detected sleep-wake cycles, off-wrist periods, sleep parameters, parametric (COSINOR) and non-parametric circadian rhythm analysis (NPCRA) calculations, and export of the study graphs. Off-wrist periods were automatically excluded from the analysis by the software.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSleep parameters included: sleep efficiency (the percentage of time spent asleep while in bed), total sleep time (the cumulative amount of sleep), number of awakenings and wakefulness after sleep onset (WASO). COSINOR analysis included measures of amplitude (the difference between the peak and the mean value of the cosine function), midline estimating statistic of rhythm\u0026nbsp;(MESOR, mean value of a rhythmic variable over a 24-hour period) and acrophase (timing of the peak activity in a circadian rhythm). NPCRA was conducted to assess the circadian structure of rest:activity rhythms, including intradaily variability (IV), interdaily stability (IS), and relative amplitude (RA). The analysis also examined average activity levels during the least active 5-hour period (L5) and the most active 10-hour period (M10). IV reflects rest:activity fragmentation, with higher values denoting greater consistency. IS measures rhythm stability across days, with higher values denoting greater consistency. RA quantifies rhythm strength by comparing activity levels between M10 and L5. Within-subject variability in L5, M10 and RA was assessed using the standard deviation of daily measures.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor generating the actigraphy plots for wrist temperature, PIM, TAT, and ZCM, actigraphy data was grouped by the mean per minute, for each participant and displayed on a 24-hour horizontal axis. Data transformations were performed using the Numpy and Pandas Python packages, within the Jupyter environment. The visualizations were generated using the Vega-Lite visualization grammar, with the Altair Python binding. The layered chart has two components: an area chart for the SEM bounds and a line chart for the mean values of the data. \u003c/p\u003e\n\u003ch3\u003eAxillary temperature\u003c/h3\u003e\n\u003cp\u003eAxillary temperature\u0026nbsp;was measured twice at each of the four time points within 24\u0026nbsp;h (8:00, 11:00, 15:30 and 21:30\u0026nbsp;h) in both chronic insomnia patients and control subjects, following the sleep unit routine, through an electronic thermometer. The average between the two temperature measurements was used.\u003c/p\u003e\n\u003ch3\u003eBlood samples collection and processing\u003c/h3\u003e\n\u003cp\u003eBlood samples (15\u0026ndash;20 mL) were collected from chronic insomnia patients and control subjects at four time points within 24 hours (8:00, 11:00, 15:30, and 21:30 h) via the antecubital vein using a 21-gauge needle into K2EDTA tubes (#367839, BD Vacutainer) and subsequently processed as previously described \u003csup\u003e24\u003c/sup\u003e. \u0026nbsp;Plasma and PBMCs aliquots were stored at - 80 ⁰C until further analysis.\u003c/p\u003e\n\u003ch3\u003eCortisol plasma levels assessment\u003c/h3\u003e\n\u003cp\u003ePlasma cortisol levels were quantified using human-specific commercial ELISA assays (EH0641 respectively, Fine Test), according to the manufacturer\u0026apos;s protocol. Absorbance was measured at 450\u0026nbsp;nm using a SpectraMax Plus 384 Microplate Reader (Molecular Devices). The concentrations were then calculated by interpolating the results from standard curves generated by plotting the concentration of the standards provided by each kit against their absorbance using a sigmoidal 4-parameter logistic (4-PL) curve, with concentration expressed logarithmically.\u003c/p\u003e\n\u003ch3\u003eClock gene expression analysis\u003c/h3\u003e\n\u003cp\u003eTotal\u0026nbsp;RNA\u0026nbsp;was extracted from PBMCs using the miRCURY\u0026nbsp;RNA Isolation\u0026nbsp;kit \u0026ndash; Cell and Plant (#300,110, Exiqon) according to manufacturer\u0026apos;s protocol, with DNAse digestion to prevent\u0026nbsp;genomic DNA\u0026nbsp;contamination. Total RNA was quantified by optical density (OD) using the ND-1000 Nanodrop Spectrophotometer (Thermo Scientific), and purity was assessed by the OD ratio at 260 and 280 nm. RNA samples were converted into cDNA, using the iScript cDNA Synthesis Kit (Bio-Rad), according to manufacturer\u0026apos;s instructions and as previously described \u003csup\u003e24\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe mRNA levels of \u003cem\u003eBMAL1,\u0026nbsp;\u003c/em\u003e\u003cem\u003eCLOCK\u003cem\u003e, PER1\u0026ndash;3, CRY1\u0026ndash;2, REV-ERB\u0026alpha;, and REV-ERB\u0026beta;\u003c/em\u003e\u003c/em\u003e of all the enrolled subjects were assessed by real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR), using the iCycler iQ\u0026trade; Real-Time PCR Detection System (Bio-Rad). Primer sequences and qPCR conditions are listed in Table S1. Relative gene expression was calculated according to the \u0026Delta;CT method \u003csup\u003e31\u003c/sup\u003e. Each assay included a non-template control (NTC), a no reverse transcription (NRT) control, and a standard curve for each target gene that was used in all plates to normalize for interplate variability. Hypoxanthine-guanine phosphoribosyltransferase (\u003cem\u003eHPRT\u003c/em\u003e), Glyceraldehyde 3-phosphate dehydrogenase (\u003cem\u003eGAPDH\u003c/em\u003e) and \u0026beta;\u0026minus;2-microglobulin (\u003cem\u003e\u0026beta;\u0026minus;2\u003c/em\u003e \u003cem\u003eM\u003c/em\u003e) were used as housekeeping genes, as validated previously \u003csup\u003e24\u003c/sup\u003e.\u0026nbsp;Bio-Rad CFX Maestro software (Bio-Rad) was used to automatically determine amplification efficiencies and threshold cycles (CT). mRNA expression data\u0026nbsp;is presented as \u0026Delta;\u0026Delta;CT values, relative to the average of the \u0026Delta;Ct of all time points of all enrolled subjects, except for machine learning analysis, where the \u0026Delta;CT values were used.\u003c/p\u003e\n\u003ch3\u003eStatistical analysis\u003c/h3\u003e\n\u003cp\u003eIf not stated otherwise, statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, Inc). Data distribution was assessed using the Shapiro\u0026ndash;Wilk test and the quartile\u0026ndash;quartile plot (QQ-plot) was visually inspected. Based on data distribution, unpaired t-tests were performed to detect statistically significant differences between control and chronic insomnia group, at each time of the day. For correlation dot plot, pairwise Spearman correlation\u0026nbsp;coefficients were calculated\u0026nbsp;using\u0026nbsp;Origin\u0026nbsp;2025\u0026nbsp;(OriginLab,\u0026nbsp;New\u0026nbsp;York,\u0026nbsp;USA). All statistical tests performed were two-sided, with statistical significance set at 0.05. *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001. No values were excluded. Specific statistical details are described in the figure legends.\u003c/p\u003e\n\u003ch3\u003eRhythmicity analysis\u003c/h3\u003e\n\u003cp\u003eFor the detection of circadian oscillating genes, harmonic regression fitting was used as implemented in the \u003cstrong\u003egeoTS\u003c/strong\u003e R package using the haRmonics function \u003csup\u003e32\u003c/sup\u003e. The dataset was categorized into three groups: (1) controls, (2) ISSD and (3) INSD. Missing values were imputed using the respective group\u0026apos;s mean \u0026Delta;\u0026Delta;CT values normalized to the average of all participants. We then computed the mean gene expression \u0026Delta;\u0026Delta;CT values normalized to the average of all participants values across four time points (8:00, 11:00, 15:30, and 21:30 hours) for each group and each gene. The summary plot using data from each participant of the study is depicted using the standard error of mean and the figures plotted in R programming software using ggplot2 package (v.3.3.2). A smooth harmonic regression curve was used for connecting the summary plot.\u003c/p\u003e\n\u003cp\u003eFor each gene and each group, we fitted the harmonic regression model y(t) =\u0026nbsp;\u003cem\u003em\u003c/em\u003e +\u0026nbsp;\u003cem\u003ea\u003c/em\u003e \u0026times; sin (2 \u0026times; \u0026pi; \u0026times; t/\u0026omega;) +\u0026nbsp;\u003cem\u003eb\u003c/em\u003e \u0026times; cos (2 \u0026times; \u0026pi; \u0026times; t/\u0026omega;), with an \u0026omega; being a period of 24 hours. The model parameters (\u003cem\u003em\u003c/em\u003e,\u0026nbsp;\u003cem\u003ea\u003c/em\u003e,\u0026nbsp;\u003cem\u003eb\u003c/em\u003e) were estimated using haRmonics function. The amplitude\u0026nbsp;\u003cem\u003eA\u003c/em\u003e and acrophase\u0026nbsp;\u0026phi;\u0026nbsp;were derived using:\u0026nbsp;\u003cem\u003eA\u003c/em\u003e= \u0026radic; (\u003cem\u003ea\u003c/em\u003e\u003csup\u003e\u0026nbsp;2\u003c/sup\u003e + \u003cem\u003eb\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e), \u0026phi;\u0026nbsp;=\u0026nbsp;arctan (\u003cem\u003eb\u003c/em\u003e, \u003cem\u003ea\u003c/em\u003e).\u0026nbsp;To visualize the fitted curves, a second-degree polynomial model of the form:\u0026nbsp;\u003cem\u003ey(x) =\u0026nbsp;\u003c/em\u003e\u003cem\u003eb\u003c/em\u003e\u003cem\u003e\u003csub\u003e0\u003c/sub\u003e\u003c/em\u003e\u003cem\u003e\u0026nbsp;+\u0026nbsp;\u003c/em\u003e\u003cem\u003eb\u003c/em\u003e\u003cem\u003e\u003csub\u003e1\u003c/sub\u003e\u003c/em\u003e\u003cem\u003ex +\u0026nbsp;\u003c/em\u003e\u003cem\u003eb\u003c/em\u003e\u003cem\u003e\u003csub\u003e2\u003c/sub\u003e\u003c/em\u003e\u003cem\u003e\u0026nbsp;x\u003csup\u003e2\u0026nbsp;\u003c/sup\u003e\u003c/em\u003ewas fitted to the harmonic regression outputs at the four sampling time points. We then computed the associated p-values via the F-test, which was obtained using the F-statistic components provided by the summary function in R. Error bars (standard errors) were obtained by dividing the sample standard deviation by the number of participants in each group.\u003c/p\u003e\n\u003cp\u003eFor each gene, we generated separate plots illustrating the group-wise mean gene-expression values \u0026Delta;\u0026Delta;CT at each sampling time, with error bars and the corresponding smoothed polynomial curves for each group.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eMachine learning\u003c/h3\u003e\n\u003cp\u003eTo develop and validate a predictive model for classifying insomnia status based on clock gene expression, we used gene expression data from PBMCs collected at multiple time points throughout the day (8:00, 11:00, 15:30, and 21:30 h). For each participant, we derived a single feature vector by averaging gene expression levels across all time points for each gene.\u003c/p\u003e\n\u003cp\u003eWe implemented a Support Vector Machine (SVM) using the Scikit-learn library in Python. In each training iteration, we selected a subset of the data containing three genes, representing each participant as a 3D data point with a feature vector of mean expression values for the selected genes. To identify the most predictive genes, we conducted multiple training iterations, testing all possible three-gene combinations from a set of nine genes.\u003c/p\u003e\n\u003cp\u003eThe dataset was split into training, validation, and test sets. Hyperparameter tuning was performed using the training and validation sets, varying parameters such as kernel type and regularization through grid search to optimize performance. The ten models achieving the highest validation accuracy were selected.\u003c/p\u003e\n\u003cp\u003eFinally, we evaluated generalization performance by testing the best-performing SVM configuration on the test set. This approach ensured unbiased performance estimates on unseen data.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eCohort profile: demographic and clinical characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDuring recruitment,\u0026nbsp;30 potential insomnia patients and 15 potential controls completed the baseline visit. Of these, 14 patients with insomnia and 13 control subjects met the eligibility criteria and were included in the study (Figure 1A and B).\u0026nbsp;The demographic and clinical characteristics are summarized in Table 1. The insomnia group comprised 1 male and 13 females, while the control group included 4 males and 9 females. All participants were Caucasian. There were no statistically significant differences between the groups in terms of age, BMI, STOP-BANG scores, or daytime sleepiness (Table 1). The ISI score was significantly higher in the insomnia group compared to controls. Similarly, Hospital Anxiety and Depression Scale (HADS) scores were significantly elevated in the insomnia group, relative to controls. It should be noted, none of the participants met the clinical criteria for anxiety or depression, as those with diagnosed psychiatric conditions were excluded. Insomnia patients also had higher rates of comorbid conditions, including hypertension, cardiovascular diseases, metabolic disorders (such as diabetes and dyslipidemia), and gastrointestinal conditions. Details of medication use are provided in Table 1.\u003c/p\u003e\n\u003cp\u003eAccurate assessment of sleep duration and quality is essential for understanding insomnia, which is characterized by subjective complaints of poor sleep despite variations in objective measurements. While PSG remains the gold standard for evaluating sleep architecture and diagnosing sleep disorders, actigraphy provides a non-invasive, long-term alternative that also captures circadian parameters\u0026nbsp;\u003csup\u003e33\u003c/sup\u003e. However, the degree to which these methods detect sleep disturbances in chronic insomnia remains a key question.\u003c/p\u003e\n\u003cp\u003eThis study aimed to characterize sleep parameters and stratify patients by objective sleep duration based on a single-night PSG and to determine if patients with chronic insomnia exhibited circadian parameter alterations by actigraphy (Figure 1B).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBased on TST duration, 8 patients were classified as ISSD, and 6 were categorized with INSD (Figure 1A and Table 2). PSG analysis shows that patients with chronic insomnia and the ISSD subgroup, but not the INSD, show significantly reduced TST compared to controls. Sleep latency was significantly prolonged in the ISSD subgroup compared to controls, while sleep efficiency was significantly reduced in the insomnia group and the ISSD subgroup compared to controls. Additionally, REM percentage (%) was significantly reduced and the number of awakenings was markedly higher in all insomnia groups compared to controls. Importantly, no significant differences were identified in apnea-hypopnea index (AHI) values across groups, confirming that sleep-disordered breathing was not a confounding factor in our sample (Table 2).\u003c/p\u003e\n\u003cp\u003eOur findings confirm that ISSD patients experience the most profound sleep disturbances, including significantly reduced TST, prolonged sleep latency, and lower sleep efficiency, while INSD patients show milder impairments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLower body temperatures suggest circadian dysfunction in patients with chronic insomnia\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further investigate sleep-wake patterns and circadian rhythm stability, we examined actograms, which provide a visual representation of rest-activity cycles over multiple days\u0026nbsp;\u003csup\u003e33\u003c/sup\u003e. These actograms allowed us to assess differences in sleep organization between groups, offering complementary insights to the PSG and actigraphy-derived sleep parameters. Representative actograms revealed marked differences between control participants and patients with insomnia. As expected, control group actograms showed well-organized entrainment to the day/night cycle, with stable phase and amplitude across days, aside from slight weekend phase shifts. In contrast, actograms from ISSD and INSD patients demonstrated increased sleep-onset latency, disrupted sleep maintenance and reduced sleep duration on most nights. Both controls and insomnia patients (ISSD and INSD) exhibited slightly increased sleep periods during weekends (Figure 2A).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eQuantitative analysis of actigraphy-derived rest-activity rhythms further supported these findings. Insomnia patients demonstrated higher overall activity levels, as indicated by increased average daily activity (PIM) graphs (Figure 2B and 2C), time-above-threshold (TAT) and Zero Crossing Method (ZCM) values (Supplementary Figure S1A and B), suggesting prolonged wakefulness of chronic insomnia patients.\u003c/p\u003e\n\u003cp\u003eWe next assessed circadian rhythm parameters, including parametric and non-parametric measures, wrist temperature measurements. However, most parametric and non-parametric measures, including M10, L5, RA, IS, circadian function index (CFI), MESOR, and amplitude did not significantly differ between groups. The only significant circadian rhythm alteration was lower intradaily variability in insomnia patients (Supplementary Figure S1 C-J). Skin temperature rhythms were shown to be disrupted in chronic insomnia, with patients exhibiting lower wrist temperatures compared to controls (Figure 2D), which could be indicative of circadian dysfunction. Despite these minimal differences in circadian rhythm robustness, a strong negative correlation was observed between CFI, IS and ISI scores (Figure 3 E), indicating that worse insomnia symptoms were associated with weaker circadian function.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGiven the reduced wrist temperature observed in patients with chronic insomnia and the negative correlation observed between actigraphy-derived circadian measures (CFI and IS) and ISI scores, we hypothesized that chronic insomnia might be associated with circadian disruption in biological outputs. To further investigate this, we also assessed the impact of insomnia on the biological clock by measuring axillary temperature and plasma levels of cortisol at four time points across the day (08:00, 11:00, 15:30, and 21:30 h) in both insomnia patients (ISSD and INSD) and control subjects (Figure 3).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSignificant differences in axillary temperature were identified between insomnia patients and controls. The insomnia group exhibited significantly lower axillary temperatures during the morning time points (08:00 and 11:00 h). Notably, the INSD subgroup mirrored this pattern, while the ISSD subgroup demonstrated significantly lower axillary temperature only at 08:00 h (Figure 3A-C). These diurnal differences in axillary temperature were consistent with wrist temperature data obtained via actigraphy (Figure 2D).\u003c/p\u003e\n\u003cp\u003eWe also observed that cortisol levels had distinct circadian pattern of dysregulation in insomnia patients. The overall insomnia group showed significantly elevated cortisol levels at night (21:30 h). Further subgroup analysis indicated that the ISSD subgroup displayed heightened cortisol levels at both 15:30 and 21:30 h, while no significant cortisol alterations were observed in the INSD subgroup (Figure 3A-C). These results show that chronic insomnia patients, specially the ISSD group, show cortisol dysregulation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThese data suggest modest circadian dysfunction through actigraphy, as reflected by largely unchanged parametric and non-parametric measures and a significant correlation between circadian metrics and insomnia severity. Additionally, lower wrist and axillary temperatures, accompanied by elevated nighttime cortisol in insomnia patients suggest further circadian misalignment, warranting further investigation.\u003c/p\u003e\n\u003cp\u003eChronic Insomnia, in particular ISSD, promotes alterations in expression levels and the temporal expression profiles of several clock genes\u003c/p\u003e\n\u003cp\u003eGiven the overall impact of chronic insomnia on temperature and cortisol rhythms, we aimed to investigate whether clock genes were affected at the transcriptional level in peripheral blood samples, given their importance in circadian rhythm regulation. To this end, we analyzed the expression levels of nine core-clock genes (\u003cem\u003eBMAL1, CLOCK, PER1-3, CRY1-2, REV-ERB\u0026alpha;,\u0026nbsp;\u003c/em\u003eand\u003cem\u003e\u0026nbsp;REV-ERB\u0026beta;\u003c/em\u003e) at the four time points of the study (8:00, 11:00, 15:30 and 21:30 h) in PBMCs of patients with chronic insomnia and its subgroups, compared to control subjects.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePatients with insomnia showed higher expression levels of \u003cem\u003ePER1\u0026nbsp;\u003c/em\u003e(8:00 and 15:30 h)\u003cem\u003e, PER2\u0026nbsp;\u003c/em\u003e(8:00 and 15:30 h)\u003cem\u003e, REV-ERB\u0026alpha;\u003c/em\u003e (11:00 h), and \u003cem\u003eREV-ERB\u0026beta;\u003c/em\u003e (15:30 h), relative to control subjects. No significant changes were detected in the expression levels of \u003cem\u003eBMAL\u003c/em\u003e, \u003cem\u003eCLOCK, PER1, CRY1 and CRY2\u0026nbsp;\u003c/em\u003ein PBMCs of insomnia patients (Figure 4A). Subgroup analysis revealed that the ISSD subgroup exhibited similar alterations in clock gene expression to the total chronic insomnia group. By contrast, the INSD subgroup showed fewer differences, when compared with the insomnia group, with significantly higher expression levels of \u003cem\u003eBMAL\u003c/em\u003e, \u003cem\u003ePER1\u003c/em\u003e, and \u003cem\u003eCRY2\u003c/em\u003e only at the morning timepoint (08:00 h) (Figure 4B and C).\u003c/p\u003e\n\u003cp\u003eTo better characterize the clock genes expression along the day and part of the night, we then analyzed the alignment of the datasets with a 24-hour rhythmic oscillation using harmonic regression.\u0026nbsp;Specifically, we applied a harmonic regression model with a fixed 24-hour period to the average data from all patients with insomnia, as well as to its subgroups (Supplementary Figure S2 A-C). The results indicate that \u003cem\u003ePER1\u0026nbsp;\u003c/em\u003eshows a marginally significant result (p = 0.048) in the INSD group, suggesting possible differences in circadian rhythms compared to the control group (Supplementary Table S2). Although trends in amplitude and acrophase are apparent, particularly the generally higher amplitudes in insomnia-related groups, the \u003cem\u003ep\u003c/em\u003e-values indicated that these differences, except for \u003cem\u003ePER1\u0026nbsp;\u003c/em\u003ein the INSD group, are not statistically significant. Using this methodology of analysis, the results seem to suggest that circadian rhythms of clock genes are relatively stable across the groups in this dataset.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eExplorative correlation analysis shows significant correlations between insomnia severity and circadian markers\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further investigate the relationship between insomnia severity and circadian regulation, we conducted an exploratory correlation analysis to assess the associations between insomnia sleep-related parameters (derived from PSG), actigraphy-derived circadian metrics, and biological circadian markers.\u003c/p\u003e\n\u003cp\u003eInterestingly, age emerged as a significant factor, positively correlating with the expression levels of \u003cem\u003eBMAL\u003c/em\u003e, \u003cem\u003eCLOCK\u003c/em\u003e, \u003cem\u003ePER1\u003c/em\u003e, \u003cem\u003eCRY1\u003c/em\u003e, and \u003cem\u003eREV-ERB\u0026alpha;\u0026nbsp;\u003c/em\u003ein patients with insomnia and its subgroups, suggesting that clock gene expression may change with aging in individuals with chronic insomnia (Figure 5A).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur analysis identified strong negative correlations between ISI and expression levels of \u003cem\u003eBMAL\u003c/em\u003e, \u003cem\u003eCLOCK\u003c/em\u003e, \u003cem\u003ePER2\u003c/em\u003e, \u003cem\u003eCRY1-2,\u0026nbsp;\u003c/em\u003eand cortisol\u003cem\u003e.\u0026nbsp;\u003c/em\u003eThese correlations were even more pronounced when examining the two insomnia subgroups, ISSD and INSD, separately. Additionally, ISI showed significant positive correlations with axillary temperature and \u003cem\u003ePER1\u003c/em\u003e expression, further supporting the involvement of circadian dysfunction in insomnia pathology.\u003c/p\u003e\n\u003cp\u003eTST was significantly negatively correlated with nighttime cortisol levels across all insomnia groups, reinforcing the well-established link between short sleep duration and increased hyperarousal. Moreover, modest associations were also observed between TST and \u003cem\u003ePER2\u003c/em\u003e and \u003cem\u003eREV-ERB\u0026beta;\u003c/em\u003e expression, indicating potential links between sleep duration and molecular circadian rhythms (Figure 5 A-C).\u003c/p\u003e\n\u003cp\u003eAmong the actigraphy-derived circadian parameters, the strongest associations were observed between CFI and the expression levels of \u003cem\u003eCLOCK, PER1, PER2, CRY1\u003c/em\u003e, and \u003cem\u003eCRY2\u003c/em\u003e in the insomnia group. These findings suggest that molecular clock disruptions could be reflected in circadian activity patterns, reinforcing the role of intrinsic circadian misalignment in insomnia.\u003c/p\u003e\n\u003cp\u003eSubgroup analysis further highlighted distinct associations in INSD patients. Specifically, axillary temperature was negatively correlated with age, ISI, relative amplitude (RA), and CFI. This suggests that reduced circadian stability (as reflected by lower CFI and RA values) is linked to both insomnia severity and impaired thermoregulation, particularly in INSD patients. Interestingly, the negative correlation between axillary temperature and CFI further supports the role of altered thermoregulation in insomnia-related circadian disruption (Figure 5C).\u003c/p\u003e\n\u003cp\u003eOverall, these findings highlight the complex interplay between insomnia severity, circadian markers, and molecular clock regulation. The observed associations suggest that chronic insomnia, particularly its subtypes, is characterized by disruptions in both behavioral and molecular circadian rhythms.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupervised machine learning predicts insomnia disorder using clock gene expression data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo determine whether the observed differences in clock gene expression could distinguish patients from controls and between insomnia subtypes, we applied a supervised machine learning using the clock gene expression dataset. We trained a linear support vector machine (SVM) with leave-one-out cross-validation, using mean gene expression levels as input features to classify samples as \u0026quot;control\u0026quot; or \u0026quot;patient\u0026rdquo;. In each iteration, the model was trained on all subjects except one, whose health status (control or patient) was then predicted. This approach achieved robust classification performance, correctly identifying 13 insomnia subjects with an accuracy of 92%, along with 92% precision and 92% recall (Figure 6A and Supplementary Table S3).\u003c/p\u003e\n\u003cp\u003eUsing the same SVM approach, we classified insomnia subgroups, achieving an overall accuracy of 92%, with 86% precision and 100% recall (Figure 6B and Supplementary Table S4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThese findings highlight a significant distinction between control and patient groups, identified solely from mean core-clock gene expression levels. Additionally, the model\u0026rsquo;s ability to differentiate chronic insomnia subtypes underscores its potential for detecting nuanced molecular differences.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eChronic insomnia is a complex disorder characterized by physiological and molecular alterations. Understanding its underlying mechanisms and identifying novel biomarkers is essential for improving diagnostic strategies and moving toward more objective approaches. Our findings support\u0026nbsp;growing evidence that chronic insomnia\u0026nbsp;disrupts\u0026nbsp;circadian\u0026nbsp;rhythms, with ISSD\u0026nbsp;patients\u0026nbsp;exhibiting the most pronounced impairments, including\u0026nbsp;sleep\u0026nbsp;disturbances, altered body temperature rhythms, elevated nighttime cortisol, and significant changes in clock gene expression. Additionally, machine learning identified key genes that may serve as biomarkers for diagnosing insomnia and\u0026nbsp;distinguishing\u0026nbsp;between subtypes.\u003c/p\u003e\n\u003cp\u003eConsistent with previous findings \u003csup\u003e7\u003c/sup\u003e, ISSD patients\u0026nbsp;demonstrated the most profound sleep disturbances, including significantly reduced TST, prolonged sleep latency, and lower sleep efficiency, whereas INSD patients exhibited milder impairments. A key limitation of standard PSG is the reliance on single-night recordings, which may not fully capture the severity of insomnia due to night-to-night variability\u0026nbsp;\u003csup\u003e34\u003c/sup\u003e.\u0026nbsp;Additionally, discrepancies between subjective and objective sleep assessments, often termed paradoxical insomnia or sleep state misperception, remain a challenge, as PSG\u0026nbsp;often reveals less severe sleep disturbances than reported by patients \u003csup\u003e34,35\u003c/sup\u003e.\u0026nbsp;These limitations highlight the need for complementary biomarkers to improve diagnostic accuracy.\u003c/p\u003e\n\u003cp\u003eIncreased nocturnal activity, as reflected by elevated PIM, TAT, and ZCM values, suggests that chronic insomnia patients experience more frequent or prolonged nighttime movements, potentially linked to physiological hyperarousal \u003csup\u003e36\u003c/sup\u003e. Although actigraphy is increasingly recognized as a valuable tool for evaluating circadian rhythm integrity \u003csup\u003e37,38\u003c/sup\u003e, our actigraphy-derived measures did not reveal significant differences in\u0026nbsp;circadian robustness among chronic insomnia patients. However, a strong negative correlation between ISI scores and circadian robustness (IS and CFI) suggests that greater insomnia severity is associated with weaker circadian function.\u003c/p\u003e\n\u003cp\u003eWe observed alterations in body temperature rhythms, with chronic insomnia patients with insomnia patients exhibiting attenuated amplitude and lower absolute values of wrist and axillary temperature. Given that skin temperature serves as both a marker and modulator of circadian rhythms\u0026nbsp;\u003csup\u003e39,40\u003c/sup\u003e, these findings\u0026nbsp;suggest possible circadian misalignment. Prior studies have linked skin temperature dysregulation to insomnia symptoms, including difficulties with sleep onset and maintenance\u0026nbsp;\u003csup\u003e16\u003c/sup\u003e \u003csup\u003e41\u003c/sup\u003e.\u0026nbsp;While some research suggests impaired thermoregulation as a contributing factor in certain insomnia subtypes, findings remain inconsistent, warranting further investigation\u0026nbsp;\u003csup\u003e18,42\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eElevated nighttime cortisol, a well-documented hallmark of hypothalamic-pituitary-adrenal (HPA) axis dysregulation \u003csup\u003e43\u0026ndash;46\u003c/sup\u003e,\u0026nbsp;commonly observed in insomnia. In our study, ISSD patients exhibited significantly higher nocturnal cortisol levels compared to controls, but this effect was not observed in INSD patients, reinforcing the notion that hyperarousal is more pronounced in ISSD \u003csup\u003e7\u003c/sup\u003e.\u0026nbsp;Prior research has reported positive correlations between nighttime cortisol and insomnia severity (ISI scores)\u0026nbsp;\u003csup\u003e47\u003c/sup\u003e,\u0026nbsp;further supporting HPA axis dysregulation in\u0026nbsp;insomnia. However, cortisol levels may not always be significantly elevated in insomnia \u003csup\u003e48,49\u003c/sup\u003e,\u0026nbsp;which can be partly explained by the heterogeneity of insomnia subtypes and potential confounding factors such as medication use, highlighting the need for further research into the differential pathophysiology of ISSD and INSD.\u003c/p\u003e\n\u003cp\u003eOur study reveals significant alterations in clock gene expression in chronic insomnia patients, particularly in ISSD.\u0026nbsp;\u003cem\u003ePER1\u003c/em\u003e and\u0026nbsp;\u003cem\u003ePER2\u003c/em\u003e expression peaked during wakefulness and light exposure, while\u0026nbsp;\u003cem\u003eBMAL1\u003c/em\u003e,\u0026nbsp;\u003cem\u003eREV-ERB\u0026alpha;\u003c/em\u003e, and\u0026nbsp;\u003cem\u003eREV-ERB\u0026beta;\u003c/em\u003e showed modest but significant diurnal expression variations.\u0026nbsp;These findings suggest that\u0026nbsp;insomnia-associated\u0026nbsp;circadian disruptions\u0026nbsp;extend\u0026nbsp;beyond behavioral and hormonal dysregulation to the molecular level.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePrevious studies have linked polymorphisms in clock genes, particularly in \u003cem\u003eCLOCK\u003c/em\u003e and \u003cem\u003ePER3\u003c/em\u003e, to insomnia risk \u003csup\u003e50\u0026ndash;53\u003c/sup\u003e. However, research on\u0026nbsp;clock gene expression patterns in chronic insomnia patients is limited. Prior work has demonstrated that acute sleep deprivation can disrupt \u003cem\u003ePER2\u003c/em\u003e and \u003cem\u003eBMAL1\u003c/em\u003e rhythms, indicating that sleep loss influences clock gene dynamics\u0026nbsp;\u003csup\u003e54\u003c/sup\u003e.\u0026nbsp;Additionally, studies in sleep apnea patients have reported reversible alterations in clock gene expression following continuous positive airway pressure (CPAP) treatment, further supporting the role of clock gene expression as a potential biomarker for sleep disorders\u0026nbsp;\u003csup\u003e24\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eOur findings align with these\u0026nbsp;observations, as \u003cem\u003ePER1\u003c/em\u003e and \u003cem\u003ePER2\u003c/em\u003e expression was significantly affected in insomnia patients, particularly in ISSD. This reinforces their potential as sensitive markers of sleep deprivation and circadian misalignment. However, it remains unclear whether these molecular disruptions contribute to insomnia onset or result from chronic sleep disturbances. Longitudinal studies are needed to clarify causality and further explore the molecular mechanisms underlying insomnia subtypes.\u003c/p\u003e\n\u003cp\u003eOur findings highlight the intricate relationship between chronic insomnia and circadian dysregulation across molecular, physiological, and behavioral levels. Further supporting this, we show that ISI is negatively correlated with \u003cem\u003eBMAL\u003c/em\u003e, \u003cem\u003eCLOCK\u003c/em\u003e, \u003cem\u003ePER2\u003c/em\u003e, \u003cem\u003eCRY1-2\u003c/em\u003e, and cortisol, with more pronounced disruptions in ISSD and INSD subgroups, reinforcing the role of molecular circadian alterations in insomnia pathology. Additionally, ISI positively correlated with axillary temperature and \u003cem\u003ePER1\u003c/em\u003e expression, linking sleep disturbances to thermoregulatory and circadian disruptions.\u003c/p\u003e\n\u003cp\u003eActigraphy-based circadian parameters also reflected these disruptions, with CFI correlating strongly with \u003cem\u003eCLOCK\u003c/em\u003e, \u003cem\u003ePER1\u003c/em\u003e, \u003cem\u003ePER2\u003c/em\u003e, \u003cem\u003eCRY1\u003c/em\u003e, and \u003cem\u003eCRY2\u003c/em\u003e expression. INSD patients exhibited distinct alterations, including negative correlations between axillary temperature and ISI, RA, and CFI, suggesting impaired thermoregulation and weakened circadian stability. The association between TST and nighttime cortisol, as well as\u0026nbsp;\u003cem\u003ePER2\u003c/em\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eand \u003cem\u003eREV-ERB\u0026beta;\u003c/em\u003e further supports a link between hyperarousal, sleep duration, and molecular rhythms. Additionally, age correlated positively with \u003cem\u003eBMAL\u003c/em\u003e, \u003cem\u003eCLOCK\u003c/em\u003e, \u003cem\u003ePER1\u003c/em\u003e, \u003cem\u003eCRY1\u003c/em\u003e, and \u003cem\u003eREV-ERB\u0026alpha;\u0026nbsp;\u003c/em\u003eexpression, suggesting that aging may exacerbate circadian misalignment in insomnia.\u003c/p\u003e\n\u003cp\u003eCurrently, chronic insomnia disorder diagnosis relies primarily on self-reported symptoms, with objective assessments used only in select cases \u003csup\u003e9\u003c/sup\u003e. Our study suggests that blood-based biomarkers analysis, combined with machine-learning, may provide a cost-effective alternative for identifying chronic insomnia cases and distinguishing between subtypes. Supervised machine-learning of clock gene expression profiles successfully differentiated chronic insomnia patients from control and further distinguished ISSD from INSD patients. Machine learning algorithms identified \u003cem\u003ePER2\u003c/em\u003e, \u003cem\u003eCRY1\u003c/em\u003e and \u003cem\u003eBMAL\u003c/em\u003e as key classifiers to distinguishing insomnia patients from controls and \u003cem\u003eCLOCK\u003c/em\u003e, \u003cem\u003eBMAL\u003c/em\u003e and \u003cem\u003ePER2\u003c/em\u003e for differentiating between insomnia subtypes. These findings highlight the potential of clock genes as a diagnostic biomarker and a tool for patient stratification.\u003c/p\u003e\n\u003cp\u003eIn summary, our study provides further evidence that chronic insomnia, particularly ISSD, is associated with circadian dysfunction at behavioral, hormonal, and molecular levels. Altered clock gene expression, disrupted body temperature rhythms, and elevated nighttime cortisol underscore the role of circadian dysregulation in insomnia pathology. While actigraphy showed limited sensitivity in detecting circadian alterations, skin temperature changes and clock gene expression patterns emerged as promising markers of circadian dysfunction in insomnia. Moreover, while other studies have applied machine learning to assess circadian function \u003csup\u003e24,55\u003c/sup\u003e, our study highlights the potential of machine learning approaches based on clock gene expression for identifying chronic insomnia cases and stratifying patients by subtype. Future research should validate these findings in larger cohorts, investigate causal mechanisms, and develop target interventions, particularly for ISSD patients, who exhibit more severe circadian disruptions. Recognizing ISSD and subtyping insomnia is important for personalized therapeutic approaches and should be integrated into clinical practice for more effective management.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStudy Limitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlthough our sample size was relatively small, the robustness of our findings support the application of machine learning methodologies. The rigorous eligibility criteria and the stringent participant selection process help mitigate concerns related to sample size limitations. However, we did not assess the potential impact of individual clinical histories on clock gene expression. Nevertheless, the absence of predominant comorbidities within any study group suggests that this may not represent a significant confounder. Additionally, intrinsic inter-individual variability must be considered, as it may influence molecular and cellular responses to disease and contribute to the observed heterogeneity in outcomes. Despite this, the randomized and counterbalanced number of conditions across participants reduces the potential for systematic bias.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe did not specifically analyze the seasonal effects on clock gene expression, as patient recruitment occurred year-round. However, a previous study by our group with a similar design found no significant seasonal influence \u003csup\u003e24\u003c/sup\u003e, suggesting minimal impact.\u0026nbsp;Another consideration is the use of at-home sleep tests in healthy subjects rather than a type II PSG, which was impractical for routine clinical assessments due to cost, time demands, labor-intensive procedures, and long waiting lists\u0026nbsp;\u003csup\u003e13\u003c/sup\u003e. To overcome these issues, we used the WatchPAT, a widely adopted sleep-screening tool, for the healthy subject cohort\u0026nbsp;\u003csup\u003e56,57\u003c/sup\u003e. However, it is important to note that the use of objective sleep measures remain optional in insomnia diagnosis\u0026nbsp;\u003csup\u003e9\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eOur study establishes a foundational framework for understanding the role of clock outputs and clock gene expression in chronic insomnia and its subtypes. However, given the complexity of clock-regulated signaling pathways, further research is needed to elucidate how circadian clocks interact with neuronal circuits and systemic physiology to better understand the pathophysiology of chronic insomnia.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors are thankful to Francieli Ruiz for supplying the actigraphs used in this study and to Gasoxmed, LDA. Portugal for providing the WatchPAT\u0026reg; and respective probes for this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was financed by the European Regional Development Fund (ERDF), through Centro 2020 Regional Operational Programme COMPETE 2020 - Operational Programme for Competitiveness and Internationalization and Portuguese national funds via Funda\u0026ccedil;\u0026atilde;o para a Ci\u0026ecirc;ncia e a Tecnologia (FCT; \u0026nbsp;UIDB/04539/2020, UIDP/04539/2020 and LA/P/0058/2020), and the European Social Fund through the Human Capital Operational Programme and Portuguese national funds via FCT under 2020.04499.BD, PD/BD/135497/2018, 2020.04850.BD, 2021.05334.BD and under COVID/BD/152507/2022. This work was supported by Mapfre, with the project \u0026quot;InsomniaO\u0026apos;clock\u0026quot; - reference IT137-24-400.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCCA, CC and ARA conceived the study plan and design; CC and ARA supervised the work; CCA conducted the study and performed the sample processing, experiments for clock outputs evaluation and\u0026nbsp;gene expression analysis, and PSG and actigraphy-derived data analysis; BS assisted in experiments, LSG and AS assisted in analysis and discussion, RR assisted with actigraphy evaluation; JM and JS recruited the patients; JS evaluated patients and all the necessary clinical data; MF gathered all clinical data from all patients; JA and TG were responsible for the design of the computational analysis. JA carried out the bioinformatics rhythmicity analysis in gene expression datasets and performed machine learning-based analysis with the obtained study datasets; TD analyzed the raw actigraphy data. CCA wrote the first draft of the manuscript. All co-authors reviewed the manuscript and approved the final version to be submitted.\u003c/p\u003e\n\u003ch3\u003eCompeting interests The authors declare no competing interests.\u003c/h3\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMorin CM, Jarrin DC. Epidemiology of Insomnia: Prevalence, Course, Risk Factors, and Public Health Burden. \u003cem\u003eSleep Med Clin\u003c/em\u003e. 2022;17(2):173-191. doi:10.1016/J.JSMC.2022.03.003\u003c/li\u003e\n\u003cli\u003eAmerican Academy of Sleep Medicine. \u003cem\u003eInternational Classification of Sleep Disorders, 3rd Edition, Text Revision (ICSD-3-TR)\u003c/em\u003e.; 2023.\u003c/li\u003e\n\u003cli\u003eAmerican Psychiatric Association. \u003cem\u003eDiagnostic and Statistical Manual of Mental Disorders, 5th Edition, Text Revision (DSM-5-TR).\u003c/em\u003e American Psychiatric Association Publishing; 2022. doi:10.1176/appi.books.9780890425787\u003c/li\u003e\n\u003cli\u003eBjorvatn B, Jernel\u0026ouml;v S, Pallesen S. Insomnia \u0026ndash; a heterogenic disorder often comorbid with psychological and somatic disorders and diseases: A narrative review with focus on diagnostic and treatment challenges. \u003cem\u003eFront Psychol\u003c/em\u003e. 2021;12:289. doi:10.3389/fpsyg.2021.639198\u003c/li\u003e\n\u003cli\u003eBlanken TF, Benjamins JS, Borsboom D, et al. Insomnia disorder subtypes derived from life history and traits of affect and personality. \u003cem\u003eThe lancet Psychiatry\u003c/em\u003e. 2019;6(2):151-163. doi:10.1016/S2215-0366(18)30464-4\u003c/li\u003e\n\u003cli\u003ePerlis ML, Giles DE, Mendelson WB, Bootzin RR, Wyatt JK. Psychophysiological insomnia: the behavioural model and a neurocognitive perspective. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 1997;6(3):179-188. doi:10.1046/j.1365-2869.1997.00045.x\u003c/li\u003e\n\u003cli\u003eVgontzas A, Fernandez-Mendoza J, Liao D, Bixler E. Insomnia with objective short sleep duration: The most biologically severe phenotype of the disorder. \u003cem\u003eSleep Med Rev\u003c/em\u003e. 2013;17(4):241-254. doi:10.1016/j.smrv.2012.09.005\u003c/li\u003e\n\u003cli\u003eRezaie L, Fobian AD, McCall WV, Khazaie H. Paradoxical insomnia and subjective\u0026ndash;objective sleep discrepancy: A review. \u003cem\u003eSleep Med Rev\u003c/em\u003e. 2018;40:196-202. doi:10.1016/J.SMRV.2018.01.002\u003c/li\u003e\n\u003cli\u003eRiemann D, Espie CA, Altena E, et al. The European Insomnia Guideline: An update on the diagnosis and treatment of insomnia 2023. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 2023;32(6):e14035. doi:10.1111/JSR.14035\u003c/li\u003e\n\u003cli\u003eEdinger JD, Fins AI, Goeke JM, et al. The empirical identification of insomnia subtypes: a cluster analytic approach. \u003cem\u003eSleep\u003c/em\u003e. 1996;19(5):398-411.\u003c/li\u003e\n\u003cli\u003eVgontzas A, Liao D, Pejovic S, Calhoun S, Karataraki M, Bixler EO. Insomnia with objective short sleep duration is associated with type 2 diabetes: A population-based study. \u003cem\u003eDiabetes Care\u003c/em\u003e. 2009;32(11):1980-1985. doi:10.2337/dc09-0284\u003c/li\u003e\n\u003cli\u003eGaines J, Vgontzas A, Fernandez-Mendoza J, Bixler EO. Obstructive sleep apnea and the metabolic syndrome: The road to clinically-meaningful phenotyping, improved prognosis, and personalized treatment. \u003cem\u003eSleep Med Rev\u003c/em\u003e. 2018;42:211-219. doi:10.1016/j.smrv.2018.08.009\u003c/li\u003e\n\u003cli\u003eChiner E. Approach To The Cost Of Polysomnography In A Spanish Hospital. \u003cem\u003eInternet J Pulm Med\u003c/em\u003e. 2012;2(2). doi:10.5580/1789\u003c/li\u003e\n\u003cli\u003eBaglioni C, Regen W, Teghen A, et al. Sleep changes in the disorder of insomnia: A meta-analysis of polysomnographic studies. \u003cem\u003eSleep Med Rev\u003c/em\u003e. 2014;18(3):195-213. doi:10.1016/j.smrv.2013.04.001\u003c/li\u003e\n\u003cli\u003eLack LC, Micic G, Lovato N. Circadian aspects in the aetiology and pathophysiology of insomnia. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 2023;32(6). doi:10.1111/jsr.13976\u003c/li\u003e\n\u003cli\u003eMorris M, Lack L, Dawson D. Sleep-Onset Insomniacs Have Delayed Temperature Rhythms. \u003cem\u003eSleep\u003c/em\u003e. 1990;13(1):1-14. doi:10.1093/sleep/13.1.1\u003c/li\u003e\n\u003cli\u003eWright KP, Hull JT, Hughes RJ, Ronda JM, Czeisler CA. Sleep and Wakefulness Out of Phase with Internal Biological Time Impairs Learning in Humans. \u003cem\u003eJ Cogn Neurosci\u003c/em\u003e. 2006;18(4):508-521. doi:10.1162/jocn.2006.18.4.508\u003c/li\u003e\n\u003cli\u003eLack LC, Gradisar M, Van Someren EJW, Wright HR, Lushington K. The relationship between insomnia and body temperatures. \u003cem\u003eSleep Med Rev\u003c/em\u003e. 2008;12(4):307-317. doi:10.1016/j.smrv.2008.02.003\u003c/li\u003e\n\u003cli\u003eFlynn-Evans EE, Shekleton JA, Miller B, et al. Circadian Phase and Phase Angle Disorders in Primary Insomnia. \u003cem\u003eSleep\u003c/em\u003e. 2017;40(12). doi:10.1093/sleep/zsx163\u003c/li\u003e\n\u003cli\u003eTakimoto M, Hamada A, Tomoda A, et al. Daily expression of clock genes in whole blood cells in healthy subjects and a patient with circadian rhythm sleep disorder. \u003cem\u003eAm J Physiol Integr Comp Physiol\u003c/em\u003e. 2005;289(5):R1273-R1279. doi:10.1152/ajpregu.00126.2005\u003c/li\u003e\n\u003cli\u003eBurioka N, Fukuoka Y, Takata M, et al. Circadian rhythms in the CNS and peripheral clock disorders: function of clock genes: influence of medication for bronchial asthma on circadian gene. \u003cem\u003eJ Pharmacol Sci\u003c/em\u003e. 2007;103(2):144-149. doi:10.1254/jphs.FMJ06003X4\u003c/li\u003e\n\u003cli\u003eFukuya H, Emoto N, Nonaka H, Yagita K, Okamura H, Yokoyama M. Circadian expression of clock genes in human peripheral leukocytes. \u003cem\u003eBiochem Biophys Res Commun\u003c/em\u003e. 2007;354(4):924-928. doi:10.1016/j.bbrc.2007.01.063\u003c/li\u003e\n\u003cli\u003eTakata M, Burioka N, Ohdo S, et al. Daily Expression of mRNAs for the Mammalian Clock Genes Per2 and Clock in Mouse Suprachiasmatic Nuclei and Liver and Human Peripheral Blood Mononuclear Cells. \u003cem\u003eJpn J Pharmacol\u003c/em\u003e. 2002;90(3):263-269. doi:10.1254/jjp.90.263\u003c/li\u003e\n\u003cli\u003eGaspar LS, Hesse J, Yal\u0026ccedil;in M, et al. Long-term continuous positive airway pressure treatment ameliorates biological clock disruptions in obstructive sleep apnea. \u003cem\u003eEBioMedicine\u003c/em\u003e. 2021;65(103248):103248. doi:10.1016/j.ebiom.2021.103248\u003c/li\u003e\n\u003cli\u003eZhang H, Sun H, Li J, Fan Y, J\u0026uuml;lich ST, Lei X. Subtypes of insomnia revealed by the heterogeneity of neuroanatomical patterns: A structural MRI study. \u003cem\u003eBiol Psychol\u003c/em\u003e. 2023;180:108591. doi:10.1016/j.biopsycho.2023.108591\u003c/li\u003e\n\u003cli\u003eClemente V, Ruivo Marques D, Miller-Mendes M, Morin CM, Serra J, Allen Gomes A. The European Portuguese version of the insomnia severity index. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 2021;30(1). doi:10.1111/JSR.13198\u003c/li\u003e\n\u003cli\u003eSargento P, Perea V, Ladera V, Lopes P, Oliveira J. The Epworth Sleepiness Scale in Portuguese adults: from classical measurement theory to Rasch model analysis. \u003cem\u003eSleep Breath\u003c/em\u003e. 2015;19(2):693-701. doi:10.1007/s11325-014-1078-6\u003c/li\u003e\n\u003cli\u003eReis R, Teixeira F, Martins V, et al. Validation of a Portuguese version of the STOP-Bang questionnaire as a screening tool for obstructive sleep apnea: Analysis in a sleep clinic. \u003cem\u003ePulmonology\u003c/em\u003e. 2015;21(2):61-68. doi:10.1016/J.RPPNEN.2014.04.009\u003c/li\u003e\n\u003cli\u003ePais-Ribeiro J, Silva I, Ferreira T, Martins A, Meneses R, Baltar M. Validation study of a Portuguese version of the Hospital Anxiety and Depression Scale. \u003cem\u003ePsychol Health Med\u003c/em\u003e. 2007;12(2):225-237. doi:10.1080/13548500500524088\u003c/li\u003e\n\u003cli\u003eFekedulegn D, Andrew ME, Shi M, Violanti JM, Knox S, Innes KE. Actigraphy-Based Assessment of Sleep Parameters. \u003cem\u003eAnn Work Expo Heal\u003c/em\u003e. 2020;64(4):350-367. doi:10.1093/annweh/wxaa007\u003c/li\u003e\n\u003cli\u003eLivak KJ, Schmittgen TD. Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2\u0026minus;\u0026Delta;\u0026Delta;CT Method. \u003cem\u003eMethods\u003c/em\u003e. 2001;25(4):402-408. doi:10.1006/METH.2001.1262\u003c/li\u003e\n\u003cli\u003eL\u0026uuml;ck S, Thurley K, Thaben PF, Westermark PO. Rhythmic Degradation Explains and Unifies Circadian Transcriptome and Proteome Data. \u003cem\u003eCell Rep\u003c/em\u003e. 2014;9(2):741-751. doi:10.1016/j.celrep.2014.09.021\u003c/li\u003e\n\u003cli\u003eTaibi DM, Landis CA, Vitiello M V. Concordance of Polysomnographic and Actigraphic Measurement of Sleep and Wake in Older Women with Insomnia. \u003cem\u003eJ Clin Sleep Med\u003c/em\u003e. 2013;9(3):217-225. doi:10.5664/JCSM.2482\u003c/li\u003e\n\u003cli\u003eFrase L, Nissen C, Spiegelhalder K, Feige B. The importance and limitations of polysomnography in insomnia disorder\u0026mdash;a critical appraisal. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 2023;32(6). doi:10.1111/jsr.14036\u003c/li\u003e\n\u003cli\u003eFeige B, Al‐Shajlawi A, Nissen C, et al. Does REM sleep contribute to subjective wake time in primary insomnia? A comparison of polysomnographic and subjective sleep in 100 patients. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 2008;17(2):180-190. doi:10.1111/j.1365-2869.2008.00651.x\u003c/li\u003e\n\u003cli\u003eBlackwell T, Redline S, Ancoli-Israel S, et al. Comparison of Sleep Parameters from Actigraphy and Polysomnography in Older Women: The SOF Study. \u003cem\u003eSleep\u003c/em\u003e. 2008;31(2):283-291. doi:10.1093/sleep/31.2.283\u003c/li\u003e\n\u003cli\u003eAncoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, Pollak CP. The Role of Actigraphy in the Study of Sleep and Circadian Rhythms. \u003cem\u003eSleep\u003c/em\u003e. 2003;26(3):342-392. doi:10.1093/sleep/26.3.342\u003c/li\u003e\n\u003cli\u003eFossion R, Rivera AL, Toledo-Roy JC, Ellis J, Angelova M. Multiscale adaptive analysis of circadian rhythms and intradaily variability: Application to actigraphy time series in acute insomnia subjects. \u003cem\u003ePLoS One\u003c/em\u003e. 2017;12(7):e0181762. doi:10.1371/JOURNAL.PONE.0181762\u003c/li\u003e\n\u003cli\u003eMartinez-Nicolas A, Ortiz-Tudela E, Rol MA, Madrid JA. Uncovering Different Masking Factors on Wrist Skin Temperature Rhythm in Free-Living Subjects. \u003cem\u003ePLoS One\u003c/em\u003e. 2013;8(4):e61142. doi:10.1371/JOURNAL.PONE.0061142\u003c/li\u003e\n\u003cli\u003eCuesta M, Boudreau P, Cermakian N, Boivin DB. Skin Temperature Rhythms in Humans Respond to Changes in the Timing of Sleep and Light. \u003cem\u003eJ Biol Rhythms\u003c/em\u003e. 2017;32(3):257-273. doi:10.1177/0748730417702974\u003c/li\u003e\n\u003cli\u003eLack LC, Mercer JD, Wright HR. Circadian rhythms of early morning awakening insomniacs. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 1996;5(4):211-219. doi:10.1111/j.1365-2869.1996.00211.x\u003c/li\u003e\n\u003cli\u003eIijima R, Kadooka A, Sugawara K, Fushimi M, Hosoe M, Aritake-Okada S. Subjective sleep onset latency is influenced by sleep structure and body heat loss in human subjects. \u003cem\u003eJ Sleep Res\u003c/em\u003e. 2024;33(5):e14122. doi:10.1111/jsr.14122\u003c/li\u003e\n\u003cli\u003eRodenbeck A, Huether G, R\u0026uuml;ther E, Hajak G. Interactions between evening and nocturnal cortisol secretion and sleep parameters in patients with severe chronic primary insomnia. \u003cem\u003eNeurosci Lett\u003c/em\u003e. 2002;324(2):159-163. doi:10.1016/S0304-3940(02)00192-1\u003c/li\u003e\n\u003cli\u003eVgontzas A, Bixler EO, Lin H-M, et al. Chronic Insomnia Is Associated with Nyctohemeral Activation of the Hypothalamic-Pituitary-Adrenal Axis: Clinical Implications. \u003cem\u003eJ Clin Endocrinol Metab\u003c/em\u003e. 2001;86(8):3787-3794. doi:10.1210/jcem.86.8.7778\u003c/li\u003e\n\u003cli\u003eDressle RJ, Feige B, Spiegelhalder K, et al. HPA axis activity in patients with chronic insomnia: A systematic review and meta-analysis of case\u0026ndash;control studies. \u003cem\u003eSleep Med Rev\u003c/em\u003e. 2022;62:101588. doi:10.1016/j.smrv.2022.101588\u003c/li\u003e\n\u003cli\u003eZhang X-X, Sun S-Y, Ma Z-J, et al. Changed nocturnal levels of stress-related hormones couple with sleep-wake states in the patients with chronic insomnia disorder: A clinical pilot study. \u003cem\u003eSleep Med\u003c/em\u003e. 2024;117:177-183. doi:10.1016/j.sleep.2024.03.017\u003c/li\u003e\n\u003cli\u003ePassos GS, Youngstedt SD, Rozales AARC, et al. Insomnia Severity is Associated with Morning Cortisol and Psychological Health. \u003cem\u003eSleep Sci\u003c/em\u003e. 2023;16(01):092-096. doi:10.1055/s-0043-1767754\u003c/li\u003e\n\u003cli\u003eRiemann D, Klein T, Rodenbeck A, et al. Nocturnal cortisol and melatonin secretion in primary insomnia. \u003cem\u003ePsychiatry Res\u003c/em\u003e. 2002;113(1-2):17-27. doi:10.1016/S0165-1781(02)00249-4\u003c/li\u003e\n\u003cli\u003eR\u0026ouml;sler L, van Kesteren E-J, Leerssen J, et al. Hyperarousal dynamics reveal an overnight increase boosted by insomnia. \u003cem\u003eJ Psychiatr Res\u003c/em\u003e. 2024;179:279-285. doi:10.1016/j.jpsychires.2024.09.032\u003c/li\u003e\n\u003cli\u003eSerretti A, Benedetti F, Mandelli L, et al. Genetic dissection of psychopathological symptoms: insomnia in mood disorders and CLOCK gene polymorphism. \u003cem\u003eAm J Med Genet B Neuropsychiatr Genet\u003c/em\u003e. 2003;121B(1):35-38. doi:10.1002/AJMG.B.20053\u003c/li\u003e\n\u003cli\u003eLee S. Association between CLOCK Gene Polymorphisms and Insomnia Risk According to Food Groups: A KoGES Longitudinal Study. \u003cem\u003eNutrients\u003c/em\u003e. 2023;15(10):2300. doi:10.3390/NU15102300/S1\u003c/li\u003e\n\u003cli\u003eSemenova N V., Madaeva IM, Bairova TA, Zhambalova RM, Sholokhov LF, Kolesnikova LI. Association of the melatonin circadian rhythms with clock 3111T/C gene polymorphism in Caucasian and Asian menopausal women with insomnia. \u003cem\u003eChronobiol Int\u003c/em\u003e. 2018;35(8):1066-1076. doi:10.1080/07420528.2018.1456447\u003c/li\u003e\n\u003cli\u003eSemenova N V., Madaeva IM, Bairova TI, et al. 3111T/C Clock Gene Polymorphism in Women with Insomnia. \u003cem\u003eBull Exp Biol Med\u003c/em\u003e. 2017;163(4):461-464. doi:10.1007/s10517-017-3828-5\u003c/li\u003e\n\u003cli\u003eKavčič P, Rojc B, Dolenc-Gro\u0026scaron;elj L, Claustrat B, Fujs K, Poljak M. The impact of sleep deprivation and nighttime light exposure on clock gene expression in humans. \u003cem\u003eCroat Med J\u003c/em\u003e. 2011;52(5):594-603. doi:10.3325/cmj.2011.52.594\u003c/li\u003e\n\u003cli\u003eHesse J, Malhan D, Yal\u0026ccedil;in M, Aboumanify O, Basti A, Rel\u0026oacute;gio A. An Optimal Time for Treatment\u0026mdash;Predicting Circadian Time by Machine Learning and Mathematical Modelling. \u003cem\u003eCancers 2020, Vol 12, Page 3103\u003c/em\u003e. 2020;12(11):3103. doi:10.3390/CANCERS12113103\u003c/li\u003e\n\u003cli\u003ePhua CQ, Jang IJ, Tan KB, et al. Reducing cost and time to diagnosis and treatment of obstructive sleep apnea using ambulatory sleep study: a Singapore sleep centre experience. \u003cem\u003eSleep Breath\u003c/em\u003e. 2021;25(1):281-288. doi:10.1007/s11325-020-02115-z\u003c/li\u003e\n\u003cli\u003eHamida ST-B, Ahmed B, Cvetkovic D, Jovanov E, Kennedy G, Penzel T. A New Era in Sleep Monitoring: The Application of Mobile Technologies in Insomnia Diagnosis. In: ; 2015:101-127. doi:10.1007/978-3-319-12817-7_5\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Insomnia, machine learning, actigraphy, circadian clock, polysomnography","lastPublishedDoi":"10.21203/rs.3.rs-6268489/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6268489/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eChronic Insomnia is a prevalent sleep disorder that remains difficult to diagnose due to subjective symptoms and heterogeneous presentations. The most severe form, insomnia with short sleep duration (ISSD), is defined by a total sleep time of less than six hours on polysomnography. However, objective assessments are rarely recommended in diagnostic guidelines, highlighting the need for alternative biomarkers. Disruptions in the circadian clock system may contribute to chronic insomnia, though the extent of these effects remains unclear. In this study, we investigate sleep and circadian rhythm-related alterations in chronic insomnia and its subtypes, ISSD and insomnia with normal sleep duration (INSD), by assessing plasma cortisol, wrist and axillary temperature, and clock gene expression in peripheral blood mononuclear cells (PBMCs). Additionally, we use machine learning to identify the most relevant clock genes for detecting insomnia and classifying its subtypes. Chronic insomnia patients exhibited reduced body temperature rhythms, elevated nighttime cortisol levels, and significant alterations in clock genes expression, including in \u003cem\u003eBMAL1\u003c/em\u003e, \u003cem\u003ePER1-2\u003c/em\u003e, \u003cem\u003eREV-ERBα\u003c/em\u003e, and \u003cem\u003eREV-ERBβ\u003c/em\u003e, compared to controls. Most alterations were more significant in the ISSD. Moreover, associations between clock gene expression, sleep-related parameters and insomnia severity index (ISI) scores were identified. Using machine learning, we identified three genes as sensitive biomarkers distinguishing chronic insomnia from controls and differentiating between ISSD and INSD subtypes. Our findings suggest that circadian markers and machine learning could improve understanding of chronic insomnia and aid biomarker discovery for diagnosis.\u003c/p\u003e","manuscriptTitle":"Clock gene signature predicts Insomnia and links to sleep/circadian parameters","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-25 02:13:07","doi":"10.21203/rs.3.rs-6268489/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0fb33f72-b3d5-4875-994c-4e19bb0f9e0e","owner":[],"postedDate":"June 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":48239339,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":48239340,"name":"Biological sciences/Molecular biology"}],"tags":[],"updatedAt":"2025-06-25T02:13:08+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-25 02:13:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6268489","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6268489","identity":"rs-6268489","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}