Disorder-specific alterations of transient oscillatory dynamics during sleep across cortical and subcortical networks

Disorder-specific alterations of transient oscillatory dynamics during sleep across cortical and subcortical networks | 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 Disorder-specific alterations of transient oscillatory dynamics during sleep across cortical and subcortical networks Nazanin Biabani, Olga Ivanenko, Adam Birdseye, Sean Higgins, Katarina Ilic, and 12 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7660761/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 20 Jan, 2026 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract Transient sleep oscillations reflect the dynamic coordination of cortical and subcortical circuits, modulated by slow oscillatory activity. However, the disorder-specific signatures of these events across neurological, pain, and sleep disorders remain poorly characterized. In this exploratory study, we analyzed transient oscillatory dynamics in 99 individuals, including healthy controls and patients with narcolepsy type 1, non-REM parasomnia, idiopathic REM sleep behavior disorder, and fibromyalgia syndrome. Using slow oscillatory referenced time-frequency peak histograms, we applied principal and independent component analysis to uncover spectral and phase-coupling patterns across non-REM and REM stages. We identified reproducible, trait-like oscillatory structures in controls and disorder-specific deviations in patient groups, particularly during NREM sleep. Specifically, patients with narcolepsy type 1 and non-REM parasomnia exhibited altered fast sigma coupling and phase dispersion, while idiopathic REM sleep behavior disorder patients showed reduced fast sigma density and diminished phase synchrony, despite retention of spindle-like spectral structure. In internal cross-validation, slow oscillatory-power features supported robust group-level discrimination in select EEG derivations; however, broader validation in independent samples is required. These findings highlight distinctive, stage-specific microstructural alterations in sleep and pain pathologies and support the future potential of time-frequency peak analysis as a non-invasive tool for phenotyping thalamocortical and subcortical circuit function. Health sciences/Neurology Biological sciences/Neuroscience sleep microstructure slow oscillations sleep spindles time-frequency analysis sleep disorders Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Sleep is increasingly recognised not as a sequence of discrete stages, but as a continuum of evolving neural dynamics spanning multiple spatial and temporal scales 1,2 . Central to this organization are transient oscillatory events, brief, frequency-specific patterns that index thalamocortical synchrony and modulate cognition, arousal, and memory consolidation 3-5 . Canonical sleep scoring, based on amplitude and morphology thresholds, often fails to capture the fine-grained structure and timing of these oscillations, particularly their modulation by cortical slow oscillations (SOs) 6-8 . Emerging frameworks now conceptualise oscillations not as discrete events but as field-like fluctuations shaped by local SO phase and power 9-13 . Among these, the time-frequency peak (TF-peak) method enables fine-resolution mapping of oscillatory distributions across frequency, time, and state 14 . This approach has revealed stable, individual-specific traits in healthy sleep, yet its applicability to clinical populations remains unexplored. Recent methodological advances have begun to reconceptualize sleep oscillations as field-like, rather than event-like, entities. In particular, the time-frequency peak (TF-peak) framework introduced by Stokes and colleagues 14 allows for the unbiased characterization of transient events across the 4–25 Hz range, parametrized by their coupling to SO power and phase. This approach has demonstrated that transient oscillatory distributions exhibit high trait stability, topographical specificity, and inter-individual variability, suggesting that they may serve as robust electrophysiological phenotypes. Complementary modelling by Chen and colleagues 15 further revealed that spindle timing is more strongly governed by intrinsic history-dependent processes than by external SO phase alone, challenging assumptions of exogenous pacemaking and inviting a systems-level perspective on oscillatory coordination. These findings raise critical questions about the nature and organization of transient oscillations in sleep disorders. To what extent do alterations in SO-coupled dynamics reflect disorder-specific pathophysiological processes? Are deviations stage-specific, or do they generalize across NREM and REM architecture? Can TF-peak distributions provide sensitive and specific markers of neural circuit dysfunction? To address these questions, we examined transient oscillatory dynamics across five well-characterized groups: healthy individuals and patients with fibromyalgia syndrome (FM) 16 , narcolepsy type 1 (NT1), non-REM parasomnia (NREMP), and idiopathic REM sleep behavior disorder (iRBD) 17 . These disorders were selected to span distinct etiological mechanisms, ranging from sensory amplification (FM) 16 and neuromodulatory disruption (NT1), to cortical hyperexcitability (NREMP) 18 and progressive neurodegeneration (iRBD) 19 . Each represents a unique perturbation of the broader thalamocortical and cortico-subcortical systems subserving sleep. An overview of the methods is presented in Figure 1. Using histograms of TF-peak occurrences parametrized by SO power and SO phase, we decomposed oscillatory structure via principal and independent component analysis (PCA and ICA), enabling dimensionality reduction and cross-group comparison. Importantly, we analyzed NREM and REM stages both separately and jointly, permitting the identification of stage-specific alterations and their relation to underlying network states. We hypothesized that NT1 would show reduced coupling of fast sigma activity to high SO-power states and altered SO-phase preference, reflecting impaired spindle synchronization secondary to orexin deficiency. NREMP was expected to exhibit increased TF-peak density and broader phase dispersion, indicative of impaired inhibitory gating during transitions between cortical states. In iRBD, we anticipated changes in phase alignment, consistent with early brainstem-cortical desynchronization. FM, by contrast, was hypothesized to show modest, spatially restricted deviations reflecting localized thalamocortical dysregulation. By situating transient oscillatory events within a continuous, SO-referenced framework, this exploratory study sought to advance the neurophysiological characterization of sleep, establish mechanistic signatures of disease-specific dysfunction, and lay the groundwork for non-invasive biomarker development grounded in sleep microstructure. Results Demographic and sleep architecture data for all five cohorts are summarized in SI Appendix, Table S1 . SO-Power Histograms Group-Level Distributions In combined NREM + REM stages, healthy controls showed distinct frequency–power coupling: fast sigma (12–15 Hz) peaks aligned with high SO power (≥ 50%), while slower sigma (10–12 Hz) TF peaks were associated with higher SO power (≥ 75%) (Fig. 2 ). Clinical groups exhibited disorder-specific deviations. NT and iRBD showed reduced fast sigma density at high SO power levels. NREMP showed alterations in slow and fast sigma activity during high SO power, though these effects varied by frequency band and derivation. FM displayed elevated theta and low-alpha (8–10 Hz) activity, especially in NREM. Difference maps highlighted these shifts, with NT and iRBD showing decreased density in 12–15 Hz/high SO-power bins, and FM showing increases in low-frequency bins. Please refer to Fig. 2 and SI Appendix Fig S1 -S4. Component Decomposition and Statistical Comparisons For SO power histograms, PCA and ICA resulted in nine recurring component patterns (Pow1–Pow9; Fig. 3 and SI Appendix Fig. S5 ) capturing distinct frequency-power motifs. Kruskal–Wallis tests revealed significant group differences across multiple components. For instance, in RBD-specific pattern Pow1, there was a decrease in fast, and an increase in slow sigma band density. In Pow3, fast sigma was elevated in NT and NREMP. Finally, in Pow6a, high-frequency bins > 15 Hz/50%+ SO-power had higher values in RBD and NREMP versus Control. Effect sizes (η²) ranged from 0.12 to 0.35, strongest in frontal (F4) and central (C3) leads. Detailed post hoc results are in SI Appendix, Tables S2–S16 . Stage-Specific Patterns Disorder-specific alterations were primarily driven by NREM (Fig. 3 , SI Appendix Fig. S5 ). NT and NREMP retained significant deviations in Pow3 during NREM + REM and NREM-only analyses. FM did not significantly differ from controls in NREM. In REM, only iRBD and NREMP showed significant deviations; specifically, both disorders were noted, for example, for a decrease in frequency bands below 9 Hz. These findings underscore NREM as the dominant stage for sleep-spindle related microstructural alterations. Reliability of Component Structure To evaluate the stability of extracted spectral features, we performed a split-half reliability analysis across all EEG derivations ( SI Appendix, Fig. S12 ). Principal component structures derived from the SO-power histograms demonstrated strong internal consistency. Specifically, Spearman correlation coefficients between components extracted from each subset and those from the full dataset exceeded |0.75| for the first five components in most of the cases. These results indicate that the identified frequency–power patterns reflect reproducible signal structure rather than noise or sample-specific variance, supporting their utility in comparative analyses and downstream modelling. Exploratory Classification Analyses We next evaluated whether SO-power features could support group-level discrimination using exploratory logistic regression classifiers ( SI Appendix, Figure S14 ). Models were trained on principal component projection scores and assessed using internal five-fold cross-validation. Performance, quantified via the area under the receiver operating characteristic curve (AUC), was highest in frontal and central EEG derivations. Classifiers’ ability to distinguish each of the groups from others was measured by the area under the receiver operating characteristic curve metric (ROC-AUC) with the following results in the channel F4: 0.917, 0.840, 0.815, 0.910, and 0.890 for Control, NREMP, NT, RBD, and FM, respectively. Please refer to Figs. 5 and 6 . To assess robustness, we conducted permutation tests using 1,000 label shuffles ( SI Appendix, Fig. S16 ). In all cases, the true model performance exceeded the null distribution, yielding p-values below 0.05. SO-Phase Histograms Group-Level Distributions In a complementary analysis, we examined the phase alignment of transient oscillatory events relative to the slow oscillatory cycle. Among healthy controls (please see Fig. 2 ), fast sigma (12–16 Hz) TF-peaks reliably clustered around the SO trough (0 radians), a pattern consistent with previously reported 14 SO-spindle coupling. As expected, this organization was evident in NREM and combined NREM + REM stages, but absent during REM, where SO amplitude and rhythmicity are diminished. Patient groups displayed distinctive alterations in this phase-coupling architecture. In NT and NREMP, fast sigma (12–16 Hz) TF-peak distributions were broader and showed reduced clustering around the SO trough (0 radians), with relative increases in earlier SO phases. This pattern reflects phase dispersion and altered temporal alignment relative to the slow oscillatory cycle. iRBD participants showed preserved spectral profiles but a marked attenuation in phase-locking, alongside reduced fast sigma density, suggesting disrupted temporal coordination in the context of partially preserved oscillatory structure. FM participants, by contrast, did not exhibit a consistent group-level shift, although their phase histograms were more granular and less structured, pointing to increased variability. Please refer to Fig. 2 and SI Appendix, Fig. S6-S9 . Component Patterns and Statistical Tests Dimensionality reduction via PCA and ICA identified seven consistent phase-coupling components, labelled Pha1 through Pha7 ( SI Appendix, Fig. S10-S11 ). For instance, the Pha1 component, which characterized tight sigma coupling to the SO trough, was significantly diminished in RBD, reflecting the loss of coherent phase alignment. The Pha2 component, showing an inverted phase preference, was elevated in NT and indicative of early-phase shifts. Pha6 captured phase dispersion across broader frequency ranges (mainly 12–15 Hz) and was specific for NREMP. These deviations were supported by Kruskal–Wallis and Mann–Whitney tests with appropriate correction for multiple comparisons (please see SI Appendix, Tables S17-S28 ). Effect size analyses reinforced these findings and align with our hypothesis that phase-based coupling metrics are sensitive to circuit-level disruption, even when spectral power remains preserved. Stage-Specific Analyses The specificity of phase-coupling alterations to sleep stage was further evaluated ( SI Appendix, Fig. S10-S11) . Across all derivations, significant group differences were confined to NREM and NREM + REM conditions; no components reached statistical significance in REM-only analyses. This asymmetry supports the notion that SO-phase coupling, and its pathological disruption, are most robust during NREM sleep, when cortical synchronization is highest and thalamocortical gating mechanisms are most active. Although REM may still carry relevant oscillatory information, its reduced SO amplitude likely limits the reliability of phase-based metrics during this stage. Reliability and Classification Split-half analysis confirmed the reproducibility of phase-based PCA components, with the first five components again exhibiting correlation coefficients above |0.7| in most of the cases ( SI Appendix, Fig. S13) . However, discriminative modelling using these features yielded more variable performance: AUC values ranged from 0.717 to 0.902 in frontal channels, but permutation tests did not consistently achieve statistical significance ( SI Appendix, Fig. S17 ). Compared to SO-power features, SO-phase features demonstrated lower signal-to-noise ratios and higher inter-subject variability, particularly in clinical groups. These findings suggest that while SO-phase patterns offer mechanistic insight, their standalone discriminative power may be more limited in practice. Discussion This study characterizes the temporal and spectral architecture of transient oscillatory activity in sleep, revealing stage-specific and disorder-specific deviations in TF-peak distributions across four clinical and one control groups. By anchoring TF-peak dynamics to continuous slow oscillatory (SO) power and phase metrics, we move beyond traditional event-based approaches and provide a framework for identifying subtle disruptions in thalamocortical and cortico-subcortical coordination. Our results suggest that transient sleep oscillations may encode robust inter-individual signatures in health, while exhibiting systematic and physiologically interpretable alterations in neurological and sleep disorders. In NT1, SO-power histograms (Fig. 2 ) and component projections (Figs. 3 and 4 ) revealed significant changes in fast sigma (12–15 Hz) activity at high SO-power levels, accompanied by dispersed SO-phase coupling (Fig. 3 and Supplement, Fig. S10 and S11 ). These effects suggest impaired spindle recruitment during periods of cortical synchrony. The altered coupling strength, rather than spindle density, highlights a disruption in the temporal precision of thalamocortical feedback, likely attributable to orexinergic dysfunction 20 . Given that precise spindle–SO coupling plays a key role in memory consolidation and emotional regulation 21 , 22 , this temporal desynchrony, if proven in future larger and longitudinal studies, may help explain cognitive and affective disturbances commonly observed in NT1 23 . These findings highlight the translational potential of TF-peak–based metrics as non-invasive indicators of thalamocortical instability in sleep–wake disorders. In NREMP, we identified altered TF-peak density within both slow and fast sigma bands during periods of elevated SO power, though the direction and magnitude of these effects varied across cortical derivations and component topographies (Fig. 2 a; Supplementary Figure S3 ). These disruptions were most evident in components such as Pow2 and Pow6a, which exhibited significant group-level deviations alongside substantial inter-individual variability (Fig. 3 ; Supplementary Tables S6–S7 ). Complementing these spectral shifts, NREMP patients also displayed broadened SO-phase distributions in sigma frequencies ( Supplement, Figure S10 , Pha6), possibly indicative of reduced temporal precision in thalamocortical coordination. Such phase dispersion and density instability are consistent with impaired inhibitory gating and cortical disinhibition, mechanisms thought to underlie the arousal-prone transitions characteristic of parasomnias 24 . The localization of these alterations to NREM sleep, the stage most vulnerable to state dissociation, further argues their pathophysiological relevance 24 . Although behavioral arousals were not quantified here, the observed microstructural instabilities suggest that TF-peak dynamics may encode a latent susceptibility to sleep-wake fragmentation 25 , 26 . Future studies integrating high-resolution electroclinical data could determine whether these spectral-phase features anticipate abnormal motor behaviors or dissociative transitions, thereby offering a mechanistic bridge between neural dynamics and clinical expression 24 , 27 . iRBD was distinguished by a specific profile characterised by reduced fast sigma density and markedly diminished phase coupling (Figs. 3 and 4 ). This decoupling was evident during both NREM and REM sleep, with the latter showing the most pronounced divergence from controls. Importantly, although the spatial distribution of sigma activity was broadly preserved, the combination of reduced event density and weakened alignment to the SO trough indicates a disruption in the temporal coordination of thalamocortical circuits, despite the presence of morphologically spindle-like oscillations. This dissociation between preserved structure and impaired timing aligns with evidence for early degeneration of brainstem nuclei and their cortical projections in α-synucleinopathies 28 – 30 . Conventional spindle metrics, which prioritise amplitude or count, may therefore overlook early dysfunction. The disruption of SO-coupled temporal structure in iRBD may serve as a physiological marker of subcortical-cortical disintegration, potentially refining early detection strategies in at-risk populations 31 . Of note, FM exhibited the most spatially and spectrally circumscribed alterations among the clinical groups 32 . While spindle density and phase–amplitude coupling metrics did not differ significantly from controls, increased TF-peak activity was observed in the theta (4–6 Hz) and low-alpha (8–10 Hz) bands during NREM + REM combined stages (Figs. 2 and 3 ), with frontal and occipital derivations most affected 33 . This enhancement of low-frequency activity may reflect increased cortical excitability or sensory gain, possibly consistent with prior reports of alpha intrusions and heightened arousability in FM 33,34 . The absence of fast sigma or phase-coupling disruptions suggests that thalamocortical rhythm generators remain functionally intact, despite localised spectral deviations. These findings align with the interpretation of FM as a condition of altered perceptual filtering rather than global network dysfunction, and they highlight the sensitivity of TF-peak methods to subtle, non-structural alterations in sleep microarchitecture. Stage-specific analyses confirmed that the most robust deviations occurred during NREM sleep. Across all disorders, NREM histograms and components revealed clearer separation from controls than REM sleep-only modes (Figs. 3 and 4 , SI Appendix Fig. S5, S10-S11 ). REM sleep-stage data, although more variable, proved informative in RBD, where phase decoupling remained detectable. This asymmetry underscores the differential vulnerability of NREM and REM sleep to circuit-level disruption. While NREM sleep remains the dominant substrate for large-scale oscillatory synchrony, REM sleep may still expose more subtle instabilities, particularly in conditions involving degeneration or dysregulation of subcortical modulatory pathways. Clinical Implications These findings indicate that transient oscillatory dynamics, referenced to slow oscillation (SO) power and phase, may offer a principled means of probing thalamocortical and subcortical function. The observed alterations, disorder-specific in topology and stage sensitivity, support a view of sleep microstructure as a physiologically meaningful readout of underlying circuit dynamics 11 , 14 . Importantly, such features are not only mechanistically interpretable but also clinically accessible, requiring minimal instrumentation and offering millisecond-level temporal precision 14 . This approach may, thus, carry translational utility across a range of diagnostic contexts. For instance, the attenuation of SO-coupled phase synchrony in iRBD could serve as an early physiological indicator of brainstem-cortical disintegration, potentially anticipating neurodegenerative trajectories 35 . In NT1, reduced spindle-SO alignment may reflect latent thalamocortical dysregulation, with relevance to cognitive and affective symptoms. More broadly, TF-peak–derived signatures may aid in differential classification where behavioural phenotypes are ambiguous, and in refining diagnostic boundaries across syndromic spectra such as parasomnia and hypersomnia. Beyond diagnosis, the high dimensionality and temporal specificity of this method may support longitudinal tracking in clinical trials, where changes in oscillatory coordination could precede overt symptomatology. Compared to conventional sleep metrics, such as spindle density or stage proportions, TF-peak histograms may arguably provide a richer, state-dependent fingerprint of network behaviour(14). Future studies incorporating longitudinal designs, cognitive phenotyping, and pharmacological modulation may clarify whether these microstructural features anticipate clinical progression, therapeutic response, or phenotypic conversion. Limitations and Future Directions Several important limitations must be acknowledged. First, this study is exploratory in nature and draws on modestly sized clinical samples, which limits both generalizability and the statistical power for detailed subgroup analyses. While effect sizes were moderate to large, replication in independent, larger cohorts is essential to confirm these findings. In particular, the fibromyalgia (FM) and non-REM parasomnia (NREMP) groups had modest sample sizes (n = 11 and n = 16, respectively), which, while typical in exploratory EEG studies of sleep pathology, limit statistical power for detecting subtler or regionally restricted effects. As such, the observed deviations in these cohorts, especially those limited to a single derivation or frequency band, should be interpreted cautiously. Larger replication samples are required to robustly confirm the presence, spatial extent, and clinical specificity of these alterations, especially given the known heterogeneity within both FM and NREMP populations. Second, the use of standard six-channel EEG constrains spatial resolution and precludes precise topographical or source-level inferences. Future studies employing high-density EEG or source-localized magnetoencephalography (MEG) may yield more anatomically specific insights into the spatial organization of transient oscillatory dynamics. Third, while PCA and ICA provided interpretable decompositions of SO-coupled features, both methods are sensitive to preprocessing decisions and inter-dataset variability. ICA, in particular, lacks component ordering and can be less stable across subsamples; thus, our interpretations emphasized PCA components, which demonstrated a strong reproducibility in split-half analyses. Alternative dimensionality reduction approaches, such as Uniform Manifold Approximation and Projection (UMAP) or supervised embedding, may reveal additional structure not captured here. Fourth, phase metrics in scalp EEG are inherently noisier, particularly during REM sleep, than power-based measures. Consequently, observed phase effects should be interpreted with caution, especially where effect sizes are small or trends did not reach statistical significance. Fifth, although classification performance was strong, particularly for NT1, iRBD, and NREMP, these results were derived from cross-validation within the same dataset used to construct PCA features. Without independent test sets, there is a risk of model overfitting. Moreover, the control and patient groups were diagnostically well-defined and non-overlapping, which may have accentuated group separability. As such, our models should not be interpreted as clinically diagnostic tools at this stage. Broader validation in heterogeneous, real-world populations with overlapping symptoms and comorbidities will be necessary to determine translational applicability. The development of a publicly available benchmarking dataset for TF-peak–based phenotyping could also accelerate reproducibility and standardization across laboratories. Finally, although we standardized referencing during preprocessing, control and patient datasets were recorded using different reference schemes, which may introduce residual variance. While all analyses were conducted on re-referenced and standardized data, and statistical tests were channel-specific, future work using uniformly acquired EEG across cohorts would help mitigate this potential confound. The observed component patterns and group differences were robust across derivations and not restricted to any single site or channel, suggesting minimal influence from data acquisition variability. Importantly, this study does not establish causal or longitudinal relationships between the observed oscillatory features and clinical outcomes. Longitudinal designs, particularly in at-risk populations such as idiopathic RBD, will be essential to determine whether the observed deviations in transient dynamics have prognostic utility. Similarly, the behavioural correlates of these electrophysiological features, such as memory consolidation, arousal thresholds, or dream phenomenology, remain important avenues for future investigation. Conclusion By mapping transient sleep oscillations onto continuous measures of SO power and phase, this study reveals reproducible alterations across several clinical populations. These deviations, whether reductions in synchronized spindling, phase dispersion, or increased sub-sigma activity, reflect distinct pathophysiological signatures of network disruption. Together, our findings argue for a redefinition of sleep microstructure not in terms of discrete events, but as a spectrum of dynamic, state-dependent oscillatory activity. This paradigm may inform the development of electrophysiological phenotypes in sleep medicine, with future applications in diagnosis, monitoring, and mechanistic understanding. Materials and Methods A retrospective exploratory cross-sectional study was conducted on 99 adults (≥18 years) across five diagnostic categories: idiopathic REM sleep behavior disorder (iRBD; n=17), narcolepsy type 1 (NT1; n=16), non-REM parasomnia (NREMP; n=16), fibromyalgia syndrome (FM; n=11), and healthy controls (n=39; from The Montreal Archive of Sleep Studies) 36 . All clinical diagnoses were made in accordance with the International Classification of Sleep Disorders – Third Edition (ICSD-3) and confirmed by board-certified sleep specialists 16,17 . Participants included were ≥18 years of age and had no major psychiatric or neurological comorbidities, substance dependence, or use of medications known to alter sleep architecture. Ethical approval for the study was granted by the institutional Research Ethics Committee (Project No. 12436) 37,38 . The analysis was conducted on fully anonymized retrospective data, in compliance with the UK Data Protection Act and the General Data Protection Regulation (GDPR) (Regulation (EU) 2016/679). Informed consent was not required due to the retrospective design and the use of non-identifiable data 37,38 . The study was carried out in accordance with the Declaration of Helsinki(WMA, 2013). All overnight polysomnographic (PSG) recordings included standard six-channel EEG using a 10–20 montage (F3, F4, C3, C4, O1, O2). Referencing schemes were harmonized during preprocessing to ensure consistency across datasets. Sleep staging was performed manually in accordance with American Academy of Sleep Medicine (AASM) criteria 39 . Summary socio-demographic and sleep architecture metrics are provided in SI Appendix, Table S1. EEG Preprocessing and TF-Peak Extraction EEG preprocessing was conducted using MNE-Python (v1.5.0) 40 and the DYNAM-O toolbox (version 1.0) 14 . Signals were down-sampled to 100 Hz, bandpass filtered (0.1–40 Hz), and re-referenced. Channels with persistent artefacts were excluded per DYNAM-O's artefact rejection pipeline. TF-peaks were extracted using a watershed-based algorithm that identifies local maxima across the 4–25 Hz frequency range 14 . Each TF-peak was annotated by the concurrent SO power (0.3–1.5 Hz) and SO phase at its time of occurrence, yielding two-dimensional histograms per channel for both power- and phase-anchored events. Histograms were generated for three conditions: NREM+REM combined, NREM-only, and REM-only. For REM, only the interquartile range (25–75%) of SO power was included to reduce noise, reflecting the lower SO amplitude in this stage. Statistical Decomposition and Group Comparisons Each histogram (per stage and channel) was standardized 41 and submitted to PCA 42 and ICA 43 using Scikit-learn (v1.3.2) 44 . The first 10 PCA components (capturing ~70% of total variance) were retained. ICA was performed with 5 components for SO-power and 3–4 for SO-phase histograms, based on convergence criteria. Group comparisons were conducted using Kruskal–Wallis tests 45 across all five groups, with false discovery rate (FDR) control via the Benjamini–Yekutieli method 46 (α = 0.1). When significant, post hoc Mann–Whitney U tests 47 were performed between controls and each patient group, with Bonferroni-adjusted 48 p-values (p ≤ 0.05) reported in SI Appendix, Tables S2-S28 . η2 effect size 49 was calculated for all the cases ( Tables S2-S28 ), and Cliff’s delta metric 50 was additionally computed for key components. Reliability and Classification Analyses Split-half reliability was assessed by randomly dividing subjects into balanced subsets, performing PCA separately, and correlating eigenvectors between subsets and the full sample. Components with Spearman 51 r > |0.75| (first five PCs) were deemed highly reliable; those with r > |0.5| moderately reliable. To evaluate the discriminative potential of histogram-derived PCA features, we trained logistic regression models 41,52 using five-fold cross-validation 53,54 . Each subject served once in the validation fold. Performance was measured by AUC-ROC 55 . Statistical significance of classification was tested via 1,000 permutation tests 56 . Visualization and Software Data visualizations were generated using DYNAM-O (v1.0) 14 , Matplotlib (v3.8.0) 57 , and Seaborn (v0.13.0) 58 . Statistical calculations were performed by SciPy (v1.11.4) 59 and Statsmodels (v0.14.1) 46 . Declarations Acknowledgments The authors gratefully acknowledge current and former members of the Sleep and Brain Plasticity Laboratory, as well as colleagues from the Sleep Disorders Centre at Guy’s Hospital, for their contributions, support, and collaborative input throughout the development of this work. We extend our sincere thanks to the patients and patient advocacy groups whose participation and insights helped shape the aims and direction of this study. Author Contributions Conceptualization: Ivana Rosenzweig, Robert Leech Methodology and Analysis: Olga Ivanenko, Nazanin Biabani, Zoran Cvetkovic, Robert Leech Writing – Original Draft: Nazanin Biabani, Olga Ivanenko, Ivana Rosenzweig Writing – Review & Editing: All authors Supervision: Ivana Rosenzweig, Robert Leech Data and Code Availability Statement The custom analysis code used in this study will be made publicly available via GitHub upon publication [link]. Due to ethical and legal restrictions, the raw clinical EEG data cannot be shared. These data include sensitive health information and are governed by the data protection policies of Guy’s and St Thomas’ NHS Foundation Trust and King’s College London. Requests for access to derived data or anonymized summary metrics may be considered on a case-by-case basis by the corresponding author and are subject to review by the Trust’s Research and Development Office and the institutional Data Protection Officer, in accordance with GDPR and NHS research governance frameworks. Competing Interest Statement The authors declare no competing interests. Funding This research was funded in whole, or in part, by the Wellcome Trust (103952/Z/14/Z). For the purpose of open access, the author IR has applied a CC BY public copyright license to any author accepted manuscript version arising from this submission. This article represents independent research in part funded by the NIHR Maudsley Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London. References Andrillon, T. & Oudiette, D. What is sleep exactly? Global and local modulations of sleep oscillations all around the clock. Neuroscience & Biobehavioral Reviews 155 , 105465 (2023). Peyrache, A. et al. Spatiotemporal dynamics of neocortical excitation and inhibition during human sleep. Proceedings of the national academy of sciences 109 , 1731-1736 (2012). Oudiette, D., Antony, J. W., Creery, J. D. & Paller, K. A. The role of memory reactivation during wakefulness and sleep in determining which memories endure. Journal of Neuroscience 33 , 6672-6678 (2013). Steriade, M. & Deschenes, M. The thalamus as a neuronal oscillator. Brain Research Reviews 8 , 1-63 (1984). Buzsaki, G. & Draguhn, A. Neuronal oscillations in cortical networks. science 304 , 1926-1929 (2004). Horie, K. et al. Automated sleep stage scoring employing a reasoning mechanism and evaluation of its explainability. Scientific reports 12 , 12799 (2022). Himanen, S.-L. & Hasan, J. Limitations of rechtschaffen and kales. Sleep medicine reviews 4 , 149-167 (2000). Moser, D. et al. Sleep classification according to AASM and Rechtschaffen & Kales: effects on sleep scoring parameters. Sleep 32 , 139-149 (2009). Klinzing, J. G. et al. Spindle activity phase-locked to sleep slow oscillations. Neuroimage 134 , 607-616 (2016). Cakan, C. et al. Spatiotemporal patterns of adaptation-induced slow oscillations in a whole-brain model of slow-wave sleep. Frontiers in computational neuroscience 15 , 800101 (2022). Olbrich, E. & Achermann, P. Analysis of oscillatory patterns in the human sleep EEG using a novel detection algorithm. Journal of sleep research 14 , 337-346 (2005). Dimitrov, T., He, M., Stickgold, R. & Prerau, M. J. Sleep spindles comprise a subset of a broader class of electroencephalogram events. Sleep 44 , doi:10.1093/sleep/zsab099 (2021). Mylonas, D. et al. A Novel Approach to Estimating the Cortical Sources of Sleep Spindles Using Simultaneous EEG/MEG. Frontiers in Neurology Volume 13 - 2022 , doi:10.3389/fneur.2022.871166 (2022). Stokes, P. A. et al. Transient oscillation dynamics during sleep provide a robust basis for electroencephalographic phenotyping and biomarker identification. Sleep 46 , zsac223, doi:10.1093/sleep/zsac223 (2023). Chen, S., He, M., Brown, R. E., Eden, U. T. & Prerau, M. J. Individualized temporal patterns drive human sleep spindle timing. Proc Natl Acad Sci U S A 122 , e2405276121, doi:10.1073/pnas.2405276121 (2025). Treede, R. D. et al. Chronic pain as a symptom or a disease: the IASP Classification of Chronic Pain for the International Classification of Diseases (ICD-11). Pain 160 , 19-27, doi:10.1097/j.pain.0000000000001384 (2019). Sateia, M. J. International classification of sleep disorders-third edition: highlights and modifications. Chest 146 , 1387-1394, doi:10.1378/chest.14-0970 (2014). Gibbs, S. A. et al. Sleep-related epileptic behaviors and non-REM-related parasomnias: Insights from stereo-EEG. Sleep Medicine Reviews 25 , 4-20, doi:https://doi.org/10.1016/j.smrv.2015.05.002 (2016). Högl, B., Stefani, A. & Videnovic, A. Idiopathic REM sleep behaviour disorder and neurodegeneration — an update. Nature Reviews Neurology 14 , 40-55, doi:10.1038/nrneurol.2017.157 (2018). Thannickal, T. C. et al. Reduced number of hypocretin neurons in human narcolepsy. Neuron 27 , 469-474 (2000). Helfrich, R. F., Mander, B. A., Jagust, W. J., Knight, R. T. & Walker, M. P. Old brains come uncoupled in sleep: slow wave-spindle synchrony, brain atrophy, and forgetting. Neuron 97 , 221-230. e224 (2018). Denis, D. et al. Slow oscillation‐spindle coupling is negatively associated with emotional memory formation following stress. European Journal of Neuroscience 55 , 2632-2650 (2022). Harel, B. T. et al. The nature and magnitude of cognitive impairment in narcolepsy type 1, narcolepsy type 2, and idiopathic hypersomnia: a meta-analysis. SLEEP Advances 5 , doi:10.1093/sleepadvances/zpae043 (2024). Castelnovo, A., Lopez, R., Proserpio, P., Nobili, L. & Dauvilliers, Y. NREM sleep parasomnias as disorders of sleep-state dissociation. Nature Reviews Neurology 14 , 470-481 (2018). Lecci, S. et al. Coordinated infraslow neural and cardiac oscillations mark fragility and offline periods in mammalian sleep. Science Advances 3 , e1602026, doi:doi:10.1126/sciadv.1602026 (2017). Castelnovo, A. et al. Conscious experiences during non-rapid eye movement sleep parasomnias. Neuroscience & Biobehavioral Reviews 167 , 105919, doi:https://doi.org/10.1016/j.neubiorev.2024.105919 (2024). Mutti, C. et al. Commonalities and Differences in NREM Parasomnias and Sleep-Related Epilepsy: Is There a Continuum Between the Two Conditions? Front Neurol 11 , 600026, doi:10.3389/fneur.2020.600026 (2020). Barone, D. A. & Henchcliffe, C. Rapid eye movement sleep behavior disorder and the link to alpha-synucleinopathies. Clinical Neurophysiology 129 , 1551-1564 (2018). Grinberg, L. T., Rueb, U. & Heinsen, H. Brainstem: Neglected Locus in Neurodegenerative Diseases. Frontiers in Neurology volume 2 - 2011 , doi:10.3389/fneur.2011.00042 (2011). Gelegen, C. et al. Relevance of sleep and associated structural changes in GBA1 mouse to human rapid eye movement behavior disorder. Scientific Reports 12 , 7973, doi:10.1038/s41598-022-11516-x (2022). Iranzo, A. et al. Rapid-eye-movement sleep behaviour disorder as an early marker for a neurodegenerative disorder: a descriptive study. The Lancet Neurology 5 , 572-577 (2006). Bigatti, S. M., Hernandez, A. M., Cronan, T. A. & Rand, K. L. Sleep disturbances in fibromyalgia syndrome: relationship to pain and depression. Arthritis and rheumatism 59 , 961-967, doi:10.1002/art.23828 (2008). Vijayan, S., Klerman, E. B., Adler, G. K. & Kopell, N. J. Thalamic mechanisms underlying alpha-delta sleep with implications for fibromyalgia. Journal of neurophysiology 114 , 1923-1930, doi:10.1152/jn.00280.2015 (2015). Drewes, A. M., Nielsen, K. D., Jennum, P. & Andreasen, A. Alpha intrusion in fibromyalgia. Journal of Musculoskeletal Pain 1 , 223-228 (1993). Boeve, B. F., Silber, M. H., Ferman, T. J., Lucas, J. A. & Parisi, J. E. Association of REM sleep behavior disorder and neurodegenerative disease may reflect an underlying synucleinopathy. Movement disorders: official journal of the Movement Disorder Society 16 , 622-630 (2001). O'Reilly, C., Gosselin, N., Carrier, J. & Nielsen, T. Montreal Archive of Sleep Studies: an open-access resource for instrument benchmarking and exploratory research. J Sleep Res 23 , 628-635, doi:10.1111/jsr.12169 (2014). World Medical, A. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA 310 , 2191-2194, doi:10.1001/jama.2013.281053 (2013). Council, E. P. a. o. t. Vol. Regulation (EU) 2016/679 (ed EU) (Official Journal of the European Union, 2016). Iber, C. The AASM manual for the scoring of sleep and associated events: rules, terminology, and technical specification. (No Title) (2007). Gramfort, A. et al. MEG and EEG data analysis with MNE-Python. Front Neurosci 7 , 267, doi:10.3389/fnins.2013.00267 (2013). Bishop, C. M. & Nasrabadi, N. M. Pattern recognition and machine learning . Vol. 4 (Springer, 2006). Tipping, M. E. & Bishop, C. M. Probabilistic principal component analysis. Journal of the Royal Statistical Society Series B: Statistical Methodology 61 , 611-622 (1999). Hyvarinen, A. & Oja, E. Independent component analysis: algorithms and applications. Neural Netw 13 , 411-430, doi:10.1016/s0893-6080(00)00026-5 (2000). Pedregosa, F. et al. Scikit-learn: Machine learning in Python. the Journal of machine Learning research 12 , 2825-2830 (2011). Kruskal, W. H. & Wallis, W. A. Use of ranks in one-criterion variance analysis. Journal of the American statistical Association 47 , 583-621 (1952). Benjamini, Y. & Yekutieli, D. The control of the false discovery rate in multiple testing under dependency. Annals of statistics , 1165-1188 (2001). Mann, H. B. & Whitney, D. R. On a test of whether one of two random variables is stochastically larger than the other. The annals of mathematical statistics , 50-60 (1947). Bonferroni, C. E. Teoria statistica delle classi e calcolo delle probabilità . (Seeber, 1936). Cohen, J. Statistical Power Analysis for the Behavioral Sciences . (Academic Press, 1969). Cliff, N. Dominance statistics: Ordinal analyses to answer ordinal questions. Psychological Bulletin 114 , 494-509, doi:10.1037/0033-2909.114.3.494 (1993). Spearman, C. The proof and measurement of association between two things. (1961). Fan, R.-E., Chang, K.-W., Hsieh, C.-J., Wang, X.-R. & Lin, C.-J. LIBLINEAR: A library for large linear classification. the Journal of machine Learning research 9 , 1871-1874 (2008). Stone, M. Cross‐validatory choice and assessment of statistical predictions. Journal of the royal statistical society: Series B (Methodological) 36 , 111-133 (1974). Hastie, T., Tibshirani, R., Friedman, J. H. & Friedman, J. H. The elements of statistical learning: data mining, inference, and prediction . Vol. 2 (Springer, 2009). Hanley, J. A. & McNeil, B. J. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 143 , 29-36, doi:10.1148/radiology.143.1.7063747 (1982). Ojala, M. & Garriga, G. C. Permutation tests for studying classifier performance. Journal of machine learning research 11 (2010). Hunter, J. D. Matplotlib: A 2D graphics environment. Computing in science & engineering 9 , 90-95 (2007). Waskom, M. L. Seaborn: statistical data visualization. Journal of Open Source Software 6 , 3021 (2021). Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nature methods 17 , 261-272 (2020). Additional Declarations No competing interests reported. Supplementary Files Biabanietalsupplement.pdf Cite Share Download PDF Status: Published Journal Publication published 20 Jan, 2026 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 27 Oct, 2025 Reviews received at journal 06 Oct, 2025 Reviews received at journal 02 Oct, 2025 Reviewers agreed at journal 29 Sep, 2025 Reviewers agreed at journal 26 Sep, 2025 Reviewers invited by journal 24 Sep, 2025 Editor invited by journal 24 Sep, 2025 Editor assigned by journal 22 Sep, 2025 Submission checks completed at journal 20 Sep, 2025 First submitted to journal 19 Sep, 2025 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-7660761","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":524602657,"identity":"c5e66915-3d2a-4b3e-89a5-cf759cdde480","order_by":0,"name":"Nazanin Biabani","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Nazanin","middleName":"","lastName":"Biabani","suffix":""},{"id":524602658,"identity":"ffcedc6c-0ffd-487c-8eae-f98020516af9","order_by":1,"name":"Olga Ivanenko","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Olga","middleName":"","lastName":"Ivanenko","suffix":""},{"id":524602659,"identity":"e47ef2be-2974-47a6-8ff3-30774248de1d","order_by":2,"name":"Adam Birdseye","email":"","orcid":"","institution":"Guy's and St Thomas’ NHS Foundation Trust","correspondingAuthor":false,"prefix":"","firstName":"Adam","middleName":"","lastName":"Birdseye","suffix":""},{"id":524602660,"identity":"80833e4a-09f3-4658-88d7-4e55767f0744","order_by":3,"name":"Sean Higgins","email":"","orcid":"","institution":"Guy's and St Thomas’ NHS Foundation Trust","correspondingAuthor":false,"prefix":"","firstName":"Sean","middleName":"","lastName":"Higgins","suffix":""},{"id":524602661,"identity":"f997c696-f68a-4f97-831f-0b2e988b51a8","order_by":4,"name":"Katarina Ilic","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Katarina","middleName":"","lastName":"Ilic","suffix":""},{"id":524602662,"identity":"7b691df4-691a-486f-be7e-0049ccd96b73","order_by":5,"name":"Jan Rosenzweig","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Jan","middleName":"","lastName":"Rosenzweig","suffix":""},{"id":524602663,"identity":"88adf925-aed3-47ac-a7f1-5dc7d4ade4a4","order_by":6,"name":"Zoran Cvetkovic","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Zoran","middleName":"","lastName":"Cvetkovic","suffix":""},{"id":524602664,"identity":"1cbca634-2375-49f9-a306-5e3955a618b7","order_by":7,"name":"Alexander D Nesbitt","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"D","lastName":"Nesbitt","suffix":""},{"id":524602665,"identity":"6b2d43c9-76ef-44d1-bef8-245bb211cafd","order_by":8,"name":"Carlotta Mutti","email":"","orcid":"","institution":"Parma University Hospital, University of Parma","correspondingAuthor":false,"prefix":"","firstName":"Carlotta","middleName":"","lastName":"Mutti","suffix":""},{"id":524602666,"identity":"d5df3ea2-98e3-4007-bfa3-d90e45651cc4","order_by":9,"name":"Liborio Parrino","email":"","orcid":"","institution":"Parma University Hospital, University of Parma","correspondingAuthor":false,"prefix":"","firstName":"Liborio","middleName":"","lastName":"Parrino","suffix":""},{"id":524602667,"identity":"24ea925f-4a16-483a-ab5a-17686f6b7fa2","order_by":10,"name":"Sharon Naismith","email":"","orcid":"","institution":"University of Sydney","correspondingAuthor":false,"prefix":"","firstName":"Sharon","middleName":"","lastName":"Naismith","suffix":""},{"id":524602668,"identity":"4bb21ae2-7e9d-46b8-9b50-875a9df6422c","order_by":11,"name":"Panagis Drakatos","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Panagis","middleName":"","lastName":"Drakatos","suffix":""},{"id":524602669,"identity":"9a1ee0e0-6e68-433f-ad77-de7381fb5db5","order_by":12,"name":"David O’Regan","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"","lastName":"O’Regan","suffix":""},{"id":524602670,"identity":"06415cd5-d871-44bf-a48c-3c53aa0f05a0","order_by":13,"name":"Karl Morten","email":"","orcid":"","institution":"University of Oxford","correspondingAuthor":false,"prefix":"","firstName":"Karl","middleName":"","lastName":"Morten","suffix":""},{"id":524602672,"identity":"d2a5cd35-7304-4d80-abcb-6bb0f6d492f9","order_by":14,"name":"Peter J Goadsby","email":"","orcid":"","institution":"NIHR-Wellcome Trust King’s Clinical Research Facility, King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Peter","middleName":"J","lastName":"Goadsby","suffix":""},{"id":524602674,"identity":"9fa18fbf-151e-4cfc-b767-8aed739a7159","order_by":15,"name":"Robert Leech","email":"","orcid":"","institution":"King’s College London","correspondingAuthor":false,"prefix":"","firstName":"Robert","middleName":"","lastName":"Leech","suffix":""},{"id":524602676,"identity":"3519d072-b76b-4b95-acc2-0f260061ed85","order_by":16,"name":"Ivana Rosenzweig","email":"data:image/png;base64,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","orcid":"","institution":"King’s College London","correspondingAuthor":true,"prefix":"","firstName":"Ivana","middleName":"","lastName":"Rosenzweig","suffix":""}],"badges":[],"createdAt":"2025-09-19 18:08:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7660761/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7660761/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-33669-1","type":"published","date":"2026-01-20T15:58:27+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":93009923,"identity":"b6b88010-65ce-4c21-8de1-1a4f374d1db8","added_by":"auto","created_at":"2025-10-08 07:09:06","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3083261,"visible":true,"origin":"","legend":"","description":"","filename":"Biabanietal.docx","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/fb2249fa15d067712b892691.docx"},{"id":93009898,"identity":"eeeab7d4-785f-4a76-b96e-e407838c2d2d","added_by":"auto","created_at":"2025-10-08 07:09:04","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17325,"visible":true,"origin":"","legend":"","description":"","filename":"b17b723d890a42d68c1351114ae176cc.json","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/32f867f7bda267629fa3d40a.json"},{"id":93011134,"identity":"473b145b-141a-4f8d-90cc-7a5f0bae6f41","added_by":"auto","created_at":"2025-10-08 07:17:05","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16292640,"visible":true,"origin":"","legend":"","description":"","filename":"Biabanietalsupplement.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/c97ca40c23b00440e7040bcc.pdf"},{"id":93009913,"identity":"d8596c43-9501-41d7-a039-4d53d5ec5a58","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":139275,"visible":true,"origin":"","legend":"","description":"","filename":"b17b723d890a42d68c1351114ae176cc1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/c78e5385bdf68b3ea2044318.xml"},{"id":93009912,"identity":"a881864e-bc54-4533-81d6-e391b1038443","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":573142,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/b2371418eb057e16eb829b17.jpeg"},{"id":93009921,"identity":"7bda1fbf-4ef7-4a3d-b78e-898fa8b3a753","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1022057,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/1a84b6be464aebef99f90e2c.png"},{"id":93009905,"identity":"9f5c19f6-b1e5-469b-a851-8367abaeb822","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":647334,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/6256b26876144944694e98a7.png"},{"id":93009907,"identity":"7ce05347-c472-47e4-b48a-9ed09425c8bb","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":700854,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/339b990576488a363d1e2394.png"},{"id":93011136,"identity":"b9e86717-36ae-4163-9761-d1b0aa1df877","added_by":"auto","created_at":"2025-10-08 07:17:05","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":254540,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/f6f9e8f3165fc8959ac7adf9.png"},{"id":93009900,"identity":"2e99bb77-c7e0-455d-90a7-e758e231e10a","added_by":"auto","created_at":"2025-10-08 07:09:04","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45314,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/3d8b7e6f544fa5b709ebfc91.png"},{"id":93011135,"identity":"6b4a6436-7647-48ab-b512-d5f8e4843371","added_by":"auto","created_at":"2025-10-08 07:17:05","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":109339,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/37109655f622206446d2ce51.png"},{"id":93009924,"identity":"14ccb717-de4c-4cd9-96fe-54adb5e2d99f","added_by":"auto","created_at":"2025-10-08 07:09:06","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":131916,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/6f436b7a815a65bac65e8ce2.png"},{"id":93011133,"identity":"3055b80d-f26d-4b59-95f3-d8427d156dcb","added_by":"auto","created_at":"2025-10-08 07:17:04","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":67025,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/5a94f9ecb2c77f28d6512332.png"},{"id":93009916,"identity":"c817a3ce-21cc-4479-968a-64b04f9a73b9","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":73687,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/8b68439f3d57ed89fc4e1a64.png"},{"id":93009922,"identity":"357c0c46-f641-4078-bd4e-1a8fff8d8880","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47615,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/d26f38fbb2efba1433bf4b53.png"},{"id":93009925,"identity":"43808d34-3329-49f1-ab41-4742546c1477","added_by":"auto","created_at":"2025-10-08 07:09:06","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12072,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/c60caaaaa368c2e828942066.png"},{"id":93009920,"identity":"02c8e5a3-a475-4a1b-ab42-e5ce9d17c765","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":138724,"visible":true,"origin":"","legend":"","description":"","filename":"b17b723d890a42d68c1351114ae176cc1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/7b1351c125ce19fb81d41896.xml"},{"id":93009903,"identity":"c08452ff-dfc3-440e-9785-4c93e884bf84","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":156952,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/9324e0b68efa5d4eba51a636.html"},{"id":93009910,"identity":"de30f490-f69c-4f38-80b9-40d7813feec1","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":286968,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverview of the data processing and analysis pipeline.\u003c/strong\u003e \u003cstrong\u003e(A)\u003c/strong\u003e Six EEG channels were extracted and histograms were computed for TF-peak density relative to SO power and phase across NREM and REM sleep. These were decomposed via PCA and ICA, and subjected to group-wise statistical testing. \u003cstrong\u003e(B)\u003c/strong\u003e Split-half reliability analysis evaluated PCA component reproducibility. \u003cstrong\u003e(C)\u003c/strong\u003e Predictive modelling used PCA projections for classification, with performance tested via permutation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations:\u003c/strong\u003e \u003cstrong\u003eAUC,\u003c/strong\u003e area under the ROC curve; \u003cstrong\u003e[C3, C4, F3, F4, O1, O2],\u003c/strong\u003e channel’s code in the standard 10-20% electroencephalography montage; \u003cstrong\u003eEEG,\u003c/strong\u003e electroencephalography; \u003cstrong\u003eICA,\u003c/strong\u003e independent component analysis; \u003cstrong\u003eNREM,\u003c/strong\u003e non-rapid eye movement stages; \u003cstrong\u003ePCA,\u003c/strong\u003e principal component analysis; \u003cstrong\u003ePSG,\u003c/strong\u003e polysomnography; \u003cstrong\u003eREM,\u003c/strong\u003e rapid eye movement stage; \u003cstrong\u003eROC,\u003c/strong\u003e receiver operating characteristic curve; \u003cstrong\u003eSO,\u003c/strong\u003e slow oscillations.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/c1419807350a4f23270b0c8e.png"},{"id":93009902,"identity":"52890969-528c-4da3-953a-a865f52cad47","added_by":"auto","created_at":"2025-10-08 07:09:04","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":494457,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSO-Power and SO-Phase Histograms for Channel C3 (NREM + REM). \u003c/strong\u003e\u003cem\u003eTop row\u003c/em\u003e: average histograms for SO-power\u003cstrong\u003e (A) \u003c/strong\u003eand SO-phase\u003cstrong\u003e (B) \u003c/strong\u003eacross all groups. \u003cem\u003eBottom row\u003c/em\u003e: arithmetic differences between each patient group and Control. Consistent sigma-band structure is observed in controls, with reduced fast sigma peaks in NT and RBD and elevated low-alpha/theta in FM. Phase histograms show altered phase-locking and dispersion in NT and NREMP. Please refer to \u003cstrong\u003eSI Appendix, Table S1\u003c/strong\u003e for detailed channel numbers.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations: Hz,\u003c/strong\u003e hertz; \u003cstrong\u003erad,\u003c/strong\u003e radians; \u003cstrong\u003emin,\u003c/strong\u003eminutes; \u003cstrong\u003eSO,\u003c/strong\u003e slow oscillations.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/1c1353d626f692c641c23570.png"},{"id":93009904,"identity":"596b97d6-b7e3-416c-b8e0-9bf97e2c1695","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":326938,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSummary of SO-Power Component Patterns and Group Differences. (A) \u003c/strong\u003eSpectral–SO-power profiles of the nine canonical component patterns (Pow1–Pow9), shown as heatmaps. Each map illustrates frequency–SO-power coupling structure derived via PCA and ICA across all participants. \u003cstrong\u003e(B) \u003c/strong\u003eMatrix indicating which components showed significant differences (* p≤.050 Bonferroni-adjusted, Mann-Whitney pairwise tests) from Controls across patient groups (FM, NT, NREMP, RBD) and stages (NREM + REM, NREM-only, REM-only). Colorful squares and empty cells indicate the presence and absence of significant differences, respectively. This figure provides a visual summary of component–group–stage associations, highlighting both shared and distinct oscillatory alterations across disorders. Please refer to \u003cstrong\u003eSI Appendix Tables S2-S16\u003c/strong\u003e and \u003cstrong\u003eTable S1\u003c/strong\u003e, for detailed statistical metrics and channel numbers, respectively.\u003c/p\u003e\n\u003cp\u003e** red (increase) and blue (decrease) colors indicate these changes in patient groups with significant differences versus Control; all the patterns are oriented in such a way as to unify the interpretation approaches.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations:\u003c/strong\u003e [\u003cstrong\u003eC3, C4, F3, F4, O1, O2], \u003c/strong\u003echannel’s code in the standard 10-20% electroencephalography montage; \u003cstrong\u003eCI,\u003c/strong\u003econfidence intervals; \u003cstrong\u003eHz\u003c/strong\u003e, hertz; \u003cstrong\u003eFM,\u003c/strong\u003e fibromyalgia;\u003cstrong\u003eICA, \u003c/strong\u003eindependent component analysis; \u003cstrong\u003eICA x\u003c/strong\u003e, ICA component x; \u003cstrong\u003eNREM\u003c/strong\u003e, non-rapid eye movement stages; \u003cstrong\u003eNREMP\u003c/strong\u003e, non-REM parasomnia\u003cstrong\u003e; NT\u003c/strong\u003e, narcolepsy; \u003cstrong\u003ep,\u003c/strong\u003ep-value; \u003cstrong\u003ePCA\u003c/strong\u003e, principal component analysis; \u003cstrong\u003ePCA x,\u003c/strong\u003e PCA component x; \u003cstrong\u003ePow1-Pow9\u003c/strong\u003e, SO-power patterns 1-9; \u003cstrong\u003eRBD,\u003c/strong\u003e rapid eye movement sleep behavior disorder; \u003cstrong\u003eREM,\u003c/strong\u003erapid eye movement sleep stage \u003cstrong\u003eSO,\u003c/strong\u003e slow oscillations; \u003cstrong\u003e%,\u003c/strong\u003e percent.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/1a8fbd015a9426eb42ff7e81.png"},{"id":93009919,"identity":"bff1a996-f933-4d70-a37a-facd5fe8155c","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":359582,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCore SO-Power Pattern Differences Across Groups.\u003c/strong\u003eRepresentative PCA and ICA components (\u003cstrong\u003ePow1, Pow3, Pow6a, Pow7, Pow9\u003c/strong\u003e) that exhibited significant differences between Control and at least one patient group. These components capture frequency–power structures that varied by condition and stage. \u003cstrong\u003ePow1\u003c/strong\u003e and \u003cstrong\u003ePow9\u003c/strong\u003e differentiated iRBD; \u003cstrong\u003ePow3\u003c/strong\u003ewas altered in NT and NREMP; \u003cstrong\u003ePow6a\u003c/strong\u003e and \u003cstrong\u003ePow7\u003c/strong\u003e showed changed in iRBD and NREMP. Highlights: differences are most pronounced in the sigma range (10–15 Hz). Please refer to \u003cstrong\u003eFigure 3\u003c/strong\u003e, \u003cstrong\u003eSI Appendix Tables S2-S16,\u003c/strong\u003e and \u003cstrong\u003eSI Appendix Table S1\u003c/strong\u003e for a full summary of SO-power datasets results, statistical metrics, and details about the groups, respectively.\u003c/p\u003e\n\u003cp\u003e* p≤.05 Bonferroni-adjusted, Mann-Whitney pairwise tests versus Control.\u003c/p\u003e\n\u003cp\u003e** red (increase) and blue (decrease) colors indicate these changes in patient groups with significant differences versus Control; all the patterns are oriented in such a way as to unify the interpretation approaches.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations: [C3, C4, F3],\u003c/strong\u003e channel’s code in the standard 10-20% electroencephalography montage; \u003cstrong\u003eCI,\u003c/strong\u003econfidence intervals; \u003cstrong\u003eHz,\u003c/strong\u003e hertz; \u003cstrong\u003eFM,\u003c/strong\u003e fibromyalgia; \u003cstrong\u003eICA,\u003c/strong\u003e independent component analysis; \u003cstrong\u003eICA x\u003c/strong\u003e, ICA component x; \u003cstrong\u003eNREM\u003c/strong\u003e, non-rapid eye movement stages; \u003cstrong\u003eNREMP,\u003c/strong\u003enon-REM parasomnia; \u003cstrong\u003eNT\u003c/strong\u003e, narcolepsy; \u003cstrong\u003ep,\u003c/strong\u003e p-value; \u003cstrong\u003ePCA,\u003c/strong\u003e principal component analysis; \u003cstrong\u003ePCA x,\u003c/strong\u003e PCA component x; \u003cstrong\u003ePow1-Pow9,\u003c/strong\u003eSO-power patterns 1-9; \u003cstrong\u003eRBD\u003c/strong\u003e, rapid eye movement sleep behavior disorder; \u003cstrong\u003eREM,\u003c/strong\u003erapid eye movement sleep stage; \u003cstrong\u003eSO,\u003c/strong\u003eslow oscillations; \u003cstrong\u003e%, \u003c/strong\u003epercent.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/ad7a8929d3cb2b2eac52c76d.png"},{"id":93013344,"identity":"b12dbdb6-0603-4c2a-b339-485503108dba","added_by":"auto","created_at":"2025-10-08 07:25:05","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":335077,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePCA and ICA Projections in Channel F4 (NREM + REM). (A) \u003c/strong\u003eBoxplots of PCA projection scores for components with significant between-group differences. PCA Component 1 separates NT from controls; Component 2 differentiates NREMP and iRBD. \u003cstrong\u003e(B) \u003c/strong\u003eBoxplots of ICA projection scores, with ICA Component 1 capturing shared variance across all clinical groups. \u003cstrong\u003e(C)\u003c/strong\u003e PCA biplot of individual subjects in the space of Components 1 and 2, colored by group. \u003cstrong\u003e(D)\u003c/strong\u003e ICA biplot of individual subjects in the space of the two leading ICA components. These projections demonstrate diagnostic group separability based on TF-peak–SO-power structure in channel F4. Please refer to \u003cstrong\u003eFigure 3\u003c/strong\u003e, \u003cstrong\u003eSI Appendix Tables S2-S16,\u003c/strong\u003e and \u003cstrong\u003eSI Appendix Table S1\u003c/strong\u003e for a full summary of SO-power datasets results, statistical metrics, and channel numbers, respectively.\u003c/p\u003e\n\u003cp\u003e* p≤.05 Bonferroni-adjusted, Mann-Whitney pairwise tests versus Control.\u003c/p\u003e\n\u003cp\u003e** red (increase) and blue (decrease) colors indicate these changes in patient groups with significant differences versus Control; all the patterns are oriented in such a way as to unify the interpretation approaches.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003e:\u003c/strong\u003e\u003c/em\u003e \u003cstrong\u003eCI,\u003c/strong\u003e 95% confidence intervals; [\u003cstrong\u003eC3, C4, F3, F4, O1, O2], \u003c/strong\u003echannel’s code in the standard 10-20% electroencephalography montage; \u003cstrong\u003eHz\u003c/strong\u003e, hertz; \u003cstrong\u003eFM,\u003c/strong\u003efibromyalgia;\u003cstrong\u003e ICA, \u003c/strong\u003eindependent component analysis; \u003cstrong\u003eNREM,\u003c/strong\u003e non-rapid eye movement stages; \u003cstrong\u003eNREMP\u003c/strong\u003e, non-REM parasomnia\u003cstrong\u003e; NT\u003c/strong\u003e, narcolepsy; \u003cstrong\u003ePCA\u003c/strong\u003e, principal component analysis; \u003cstrong\u003eRBD, \u003c/strong\u003erapid eye movement sleep behavior disorder; \u003cstrong\u003eREM, \u003c/strong\u003erapid eye movement sleep stage; \u003cstrong\u003eSO,\u003c/strong\u003e slow oscillations; \u003cstrong\u003e%\u003c/strong\u003e, percent.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/15ab772e6b2e2eddf60c7a4a.png"},{"id":93009899,"identity":"3181fc87-3562-4c8f-b3cb-c32577f95c2d","added_by":"auto","created_at":"2025-10-08 07:09:04","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":78888,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGroup-level discrimination based on PCA-derived features in channel F4.\u003c/strong\u003e\u003cbr\u003e\n \u003cem\u003e(A)\u003c/em\u003e Receiver operating characteristic (ROC) curves for logistic regression classifiers trained on PCA projection scores. Each curve illustrates the discriminative performance for one group versus all others in internal cross-validation. High discriminability is observed for NT, RBD, and NREMP (AUC \u0026gt; 0.8). \u003cstrong\u003e(B) \u003c/strong\u003eHistograms of accuracy values from 1,000 permutation tests and true model accuracy (dashed line). These results demonstrate the diagnostic potential of TF-peak–SO-power features for non-invasive group differentiation. Please refer to \u003cstrong\u003eSI Appendix Fig. S14-S17 \u003c/strong\u003efor other discriminative modelling results. Please refer to \u003cstrong\u003eSI Appendix Table S1\u003c/strong\u003e for detailed information about the groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAbbreviations:\u003c/strong\u003e \u003cstrong\u003eAUC,\u003c/strong\u003e area under the ROC curve; \u003cstrong\u003eF4, \u003c/strong\u003echannel’s code in the standard 10-20% electroencephalography montage; \u003cstrong\u003eFM,\u003c/strong\u003e fibromyalgia;\u003cstrong\u003eNREM, \u003c/strong\u003enon-rapid eye movement stages;\u003cstrong\u003e NREMP\u003c/strong\u003e, non-REM parasomnia\u003cstrong\u003e; NT\u003c/strong\u003e, narcolepsy; \u003cstrong\u003ep,\u003c/strong\u003ep-value; \u003cstrong\u003ePCA\u003c/strong\u003e, principal component analysis; \u003cstrong\u003eRBD\u003c/strong\u003e, REM behavior disorder; \u003cstrong\u003eREM,\u003c/strong\u003e rapid eye movement sleep stage; \u003cstrong\u003eROC,\u003c/strong\u003e receiver operating characteristic curve; \u003cstrong\u003eSO,\u003c/strong\u003e slow oscillations; \u003cstrong\u003eTF\u003c/strong\u003e \u003cstrong\u003epeaks\u003c/strong\u003e, time-frequency peaks.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/67deed4c1bac65c7f3158a33.png"},{"id":101153123,"identity":"6b79c897-44cd-406a-8001-c787c74f49eb","added_by":"auto","created_at":"2026-01-26 16:14:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3497120,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/959da671-6bbb-4e7b-80d0-cb8270f5d3dc.pdf"},{"id":93009915,"identity":"e3893f88-8ea1-4330-a94a-591ffd6ae5e7","added_by":"auto","created_at":"2025-10-08 07:09:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":16292640,"visible":true,"origin":"","legend":"","description":"","filename":"Biabanietalsupplement.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7660761/v1/671e08c025b0bdb6e6b2010f.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Disorder-specific alterations of transient oscillatory dynamics during sleep across cortical and subcortical networks","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSleep is increasingly recognised not as a sequence of discrete stages, but as a continuum of evolving neural dynamics spanning multiple spatial and temporal scales\u003csup\u003e1,2\u003c/sup\u003e. Central to this organization are transient oscillatory events, brief, frequency-specific patterns that index thalamocortical synchrony and modulate cognition, arousal, and memory consolidation\u003csup\u003e3-5\u003c/sup\u003e. Canonical sleep scoring, based on amplitude and morphology thresholds, often fails to capture the fine-grained structure and timing of these oscillations, particularly their modulation by cortical slow oscillations (SOs)\u003csup\u003e6-8\u003c/sup\u003e. Emerging frameworks now conceptualise oscillations not as discrete events but as field-like fluctuations shaped by local SO phase and power\u003csup\u003e9-13\u003c/sup\u003e. Among these, the time-frequency peak (TF-peak) method enables fine-resolution mapping of oscillatory distributions across frequency, time, and state\u003csup\u003e14\u003c/sup\u003e. This approach has revealed stable, individual-specific traits in healthy sleep, yet its applicability to clinical populations remains unexplored.\u003c/p\u003e\n\u003cp\u003eRecent methodological advances have begun to reconceptualize sleep oscillations as field-like, rather than event-like, entities. In particular, the time-frequency peak (TF-peak) framework introduced by Stokes and colleagues\u003csup\u003e14\u003c/sup\u003e allows for the unbiased characterization of transient events across the 4–25 Hz range, parametrized by their coupling to SO power and phase. This approach has demonstrated that transient oscillatory distributions exhibit high trait stability, topographical specificity, and inter-individual variability, suggesting that they may serve as robust electrophysiological phenotypes. Complementary modelling by Chen and colleagues\u003csup\u003e15\u003c/sup\u003e further revealed that spindle timing is more strongly governed by intrinsic history-dependent processes than by external SO phase alone, challenging assumptions of exogenous pacemaking and inviting a systems-level perspective on oscillatory coordination.\u003c/p\u003e\n\u003cp\u003eThese findings raise critical questions about the nature and organization of transient oscillations in sleep disorders. To what extent do alterations in SO-coupled dynamics reflect disorder-specific pathophysiological processes? Are deviations stage-specific, or do they generalize across NREM and REM architecture? Can TF-peak distributions provide sensitive and specific markers of neural circuit dysfunction? To address these questions, we examined transient oscillatory dynamics across five well-characterized groups: healthy individuals and patients with fibromyalgia syndrome (FM)\u003csup\u003e16\u003c/sup\u003e, narcolepsy type 1 (NT1), non-REM parasomnia (NREMP), and idiopathic REM sleep behavior disorder (iRBD)\u003csup\u003e17\u003c/sup\u003e. These disorders were selected to span distinct etiological mechanisms, ranging from sensory amplification (FM)\u003csup\u003e16\u003c/sup\u003e and neuromodulatory disruption (NT1), to cortical hyperexcitability (NREMP)\u003csup\u003e18\u003c/sup\u003e and progressive neurodegeneration (iRBD)\u003csup\u003e19\u003c/sup\u003e. Each represents a unique perturbation of the broader thalamocortical and cortico-subcortical systems subserving sleep. An overview of the methods is presented in \u003cstrong\u003eFigure 1.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing histograms of TF-peak occurrences parametrized by SO power and SO phase, we decomposed oscillatory structure via principal and independent component analysis (PCA and ICA), enabling dimensionality reduction and cross-group comparison. Importantly, we analyzed NREM and REM stages both separately and jointly, permitting the identification of stage-specific alterations and their relation to underlying network states. We hypothesized that NT1 would show reduced coupling of fast sigma activity to high SO-power states and altered SO-phase preference, reflecting impaired spindle synchronization secondary to orexin deficiency. NREMP was expected to exhibit increased TF-peak density and broader phase dispersion, indicative of impaired inhibitory gating during transitions between cortical states. In iRBD, we anticipated changes in phase alignment, consistent with early brainstem-cortical desynchronization. FM, by contrast, was hypothesized to show modest, spatially restricted deviations reflecting localized thalamocortical dysregulation.\u003c/p\u003e\n\u003cp\u003eBy situating transient oscillatory events within a continuous, SO-referenced framework, this exploratory study sought to advance the neurophysiological characterization of sleep, establish mechanistic signatures of disease-specific dysfunction, and lay the groundwork for non-invasive biomarker development grounded in sleep microstructure.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eDemographic and sleep architecture data for all five cohorts are summarized in SI \u003cb\u003eAppendix, Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e.\u003c/p\u003e\n\u003ch3\u003eSO-Power Histograms\u003c/h3\u003e\n\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eGroup-Level Distributions\u003c/h2\u003e\u003cp\u003eIn combined NREM\u0026thinsp;+\u0026thinsp;REM stages, healthy controls showed distinct frequency\u0026ndash;power coupling: fast sigma (12\u0026ndash;15 Hz) peaks aligned with high SO power (\u0026ge;\u0026thinsp;50%), while slower sigma (10\u0026ndash;12 Hz) TF peaks were associated with higher SO power (\u0026ge;\u0026thinsp;75%) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eClinical groups exhibited disorder-specific deviations. NT and iRBD showed reduced fast sigma density at high SO power levels. NREMP showed alterations in slow and fast sigma activity during high SO power, though these effects varied by frequency band and derivation. FM displayed elevated theta and low-alpha (8\u0026ndash;10 Hz) activity, especially in NREM. Difference maps highlighted these shifts, with NT and iRBD showing decreased density in 12\u0026ndash;15 Hz/high SO-power bins, and FM showing increases in low-frequency bins. Please refer to Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cb\u003eSI Appendix Fig \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e-S4.\u003c/b\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eComponent Decomposition and Statistical Comparisons\u003c/h3\u003e\n\u003cp\u003eFor SO power histograms, PCA and ICA resulted in nine recurring component patterns (Pow1\u0026ndash;Pow9; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cb\u003eSI Appendix Fig. S5\u003c/b\u003e) capturing distinct frequency-power motifs. Kruskal\u0026ndash;Wallis tests revealed significant group differences across multiple components. For instance, in RBD-specific pattern Pow1, there was a decrease in fast, and an increase in slow sigma band density. In Pow3, fast sigma was elevated in NT and NREMP. Finally, in Pow6a, high-frequency bins\u0026thinsp;\u0026gt;\u0026thinsp;15 Hz/50%+ SO-power had higher values in RBD and NREMP versus Control.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eEffect sizes (η\u0026sup2;) ranged from 0.12 to 0.35, strongest in frontal (F4) and central (C3) leads. Detailed post hoc results are in \u003cb\u003eSI Appendix, Tables S2\u0026ndash;S16\u003c/b\u003e.\u003c/p\u003e\n\u003ch3\u003eStage-Specific Patterns\u003c/h3\u003e\n\u003cp\u003eDisorder-specific alterations were primarily driven by NREM (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, \u003cb\u003eSI Appendix Fig. S5\u003c/b\u003e). NT and NREMP retained significant deviations in Pow3 during NREM\u0026thinsp;+\u0026thinsp;REM and NREM-only analyses. FM did not significantly differ from controls in NREM. In REM, only iRBD and NREMP showed significant deviations; specifically, both disorders were noted, for example, for a decrease in frequency bands below 9 Hz. These findings underscore NREM as the dominant stage for sleep-spindle related microstructural alterations.\u003c/p\u003e\n\u003ch3\u003eReliability of Component Structure\u003c/h3\u003e\n\u003cp\u003eTo evaluate the stability of extracted spectral features, we performed a split-half reliability analysis across all EEG derivations (\u003cb\u003eSI Appendix, Fig. S12\u003c/b\u003e). Principal component structures derived from the SO-power histograms demonstrated strong internal consistency. Specifically, Spearman correlation coefficients between components extracted from each subset and those from the full dataset exceeded |0.75| for the first five components in most of the cases. These results indicate that the identified frequency\u0026ndash;power patterns reflect reproducible signal structure rather than noise or sample-specific variance, supporting their utility in comparative analyses and downstream modelling.\u003c/p\u003e\n\u003ch3\u003eExploratory Classification Analyses\u003c/h3\u003e\n\u003cp\u003eWe next evaluated whether SO-power features could support group-level discrimination using exploratory logistic regression classifiers (\u003cb\u003eSI Appendix, Figure S14\u003c/b\u003e). Models were trained on principal component projection scores and assessed using internal five-fold cross-validation. Performance, quantified via the area under the receiver operating characteristic curve (AUC), was highest in frontal and central EEG derivations. Classifiers\u0026rsquo; ability to distinguish each of the groups from others was measured by the area under the receiver operating characteristic curve metric (ROC-AUC) with the following results in the channel F4: 0.917, 0.840, 0.815, 0.910, and 0.890 for Control, NREMP, NT, RBD, and FM, respectively. Please refer to Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e6\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo assess robustness, we conducted permutation tests using 1,000 label shuffles (\u003cb\u003eSI Appendix, Fig. S16\u003c/b\u003e). In all cases, the true model performance exceeded the null distribution, yielding p-values below 0.05.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eSO-Phase Histograms\u003c/h2\u003e\u003cdiv id=\"Sec9\" class=\"Section3\"\u003e\u003ch2\u003eGroup-Level Distributions\u003c/h2\u003e\u003cp\u003eIn a complementary analysis, we examined the phase alignment of transient oscillatory events relative to the slow oscillatory cycle. Among healthy controls (please see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), fast sigma (12\u0026ndash;16 Hz) TF-peaks reliably clustered around the SO trough (0 radians), a pattern consistent with previously reported \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e SO-spindle coupling. As expected, this organization was evident in NREM and combined NREM\u0026thinsp;+\u0026thinsp;REM stages, but absent during REM, where SO amplitude and rhythmicity are diminished.\u003c/p\u003e\u003cp\u003ePatient groups displayed distinctive alterations in this phase-coupling architecture. In NT and NREMP, fast sigma (12\u0026ndash;16 Hz) TF-peak distributions were broader and showed reduced clustering around the SO trough (0 radians), with relative increases in earlier SO phases. This pattern reflects phase dispersion and altered temporal alignment relative to the slow oscillatory cycle. iRBD participants showed preserved spectral profiles but a marked attenuation in phase-locking, alongside reduced fast sigma density, suggesting disrupted temporal coordination in the context of partially preserved oscillatory structure. FM participants, by contrast, did not exhibit a consistent group-level shift, although their phase histograms were more granular and less structured, pointing to increased variability. Please refer to Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cb\u003eSI Appendix, Fig. S6-S9\u003c/b\u003e.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\n\u003ch3\u003eComponent Patterns and Statistical Tests\u003c/h3\u003e\n\u003cp\u003eDimensionality reduction via PCA and ICA identified seven consistent phase-coupling components, labelled Pha1 through Pha7 (\u003cb\u003eSI Appendix, Fig. S10-S11\u003c/b\u003e). For instance, the Pha1 component, which characterized tight sigma coupling to the SO trough, was significantly diminished in RBD, reflecting the loss of coherent phase alignment. The Pha2 component, showing an inverted phase preference, was elevated in NT and indicative of early-phase shifts. Pha6 captured phase dispersion across broader frequency ranges (mainly 12\u0026ndash;15 Hz) and was specific for NREMP. These deviations were supported by Kruskal\u0026ndash;Wallis and Mann\u0026ndash;Whitney tests with appropriate correction for multiple comparisons (please see \u003cb\u003eSI Appendix, Tables S17-S28\u003c/b\u003e).\u003c/p\u003e\u003cp\u003eEffect size analyses reinforced these findings and align with our hypothesis that phase-based coupling metrics are sensitive to circuit-level disruption, even when spectral power remains preserved.\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eStage-Specific Analyses\u003c/h2\u003e\u003cp\u003eThe specificity of phase-coupling alterations to sleep stage was further evaluated (\u003cb\u003eSI Appendix, Fig. S10-S11)\u003c/b\u003e. Across all derivations, significant group differences were confined to NREM and NREM\u0026thinsp;+\u0026thinsp;REM conditions; no components reached statistical significance in REM-only analyses. This asymmetry supports the notion that SO-phase coupling, and its pathological disruption, are most robust during NREM sleep, when cortical synchronization is highest and thalamocortical gating mechanisms are most active. Although REM may still carry relevant oscillatory information, its reduced SO amplitude likely limits the reliability of phase-based metrics during this stage.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eReliability and Classification\u003c/h2\u003e\u003cp\u003eSplit-half analysis confirmed the reproducibility of phase-based PCA components, with the first five components again exhibiting correlation coefficients above |0.7| in most of the cases (\u003cb\u003eSI Appendix, Fig. S13)\u003c/b\u003e. However, discriminative modelling using these features yielded more variable performance: AUC values ranged from 0.717 to 0.902 in frontal channels, but permutation tests did not consistently achieve statistical significance (\u003cb\u003eSI Appendix, Fig. S17\u003c/b\u003e). Compared to SO-power features, SO-phase features demonstrated lower signal-to-noise ratios and higher inter-subject variability, particularly in clinical groups. These findings suggest that while SO-phase patterns offer mechanistic insight, their standalone discriminative power may be more limited in practice.\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study characterizes the temporal and spectral architecture of transient oscillatory activity in sleep, revealing stage-specific and disorder-specific deviations in TF-peak distributions across four clinical and one control groups. By anchoring TF-peak dynamics to continuous slow oscillatory (SO) power and phase metrics, we move beyond traditional event-based approaches and provide a framework for identifying subtle disruptions in thalamocortical and cortico-subcortical coordination. Our results suggest that transient sleep oscillations may encode robust inter-individual signatures in health, while exhibiting systematic and physiologically interpretable alterations in neurological and sleep disorders.\u003c/p\u003e\u003cp\u003eIn NT1, SO-power histograms (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) and component projections (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003e) revealed significant changes in fast sigma (12\u0026ndash;15 Hz) activity at high SO-power levels, accompanied by dispersed SO-phase coupling (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cb\u003eSupplement, Fig. S10 and S11\u003c/b\u003e). These effects suggest impaired spindle recruitment during periods of cortical synchrony. The altered coupling strength, rather than spindle density, highlights a disruption in the temporal precision of thalamocortical feedback, likely attributable to orexinergic dysfunction\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Given that precise spindle\u0026ndash;SO coupling plays a key role in memory consolidation and emotional regulation\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, this temporal desynchrony, if proven in future larger and longitudinal studies, may help explain cognitive and affective disturbances commonly observed in NT1\u003csup\u003e23\u003c/sup\u003e. These findings highlight the translational potential of TF-peak\u0026ndash;based metrics as non-invasive indicators of thalamocortical instability in sleep\u0026ndash;wake disorders.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eIn NREMP, we identified altered TF-peak density within both slow and fast sigma bands during periods of elevated SO power, though the direction and magnitude of these effects varied across cortical derivations and component topographies (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea; \u003cb\u003eSupplementary Figure S3\u003c/b\u003e). These disruptions were most evident in components such as Pow2 and Pow6a, which exhibited significant group-level deviations alongside substantial inter-individual variability (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e; \u003cb\u003eSupplementary Tables S6\u0026ndash;S7\u003c/b\u003e). Complementing these spectral shifts, NREMP patients also displayed broadened SO-phase distributions in sigma frequencies (\u003cb\u003eSupplement, Figure S10\u003c/b\u003e, Pha6), possibly indicative of reduced temporal precision in thalamocortical coordination. Such phase dispersion and density instability are consistent with impaired inhibitory gating and cortical disinhibition, mechanisms thought to underlie the arousal-prone transitions characteristic of parasomnias\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. The localization of these alterations to NREM sleep, the stage most vulnerable to state dissociation, further argues their pathophysiological relevance\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Although behavioral arousals were not quantified here, the observed microstructural instabilities suggest that TF-peak dynamics may encode a latent susceptibility to sleep-wake fragmentation\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. Future studies integrating high-resolution electroclinical data could determine whether these spectral-phase features anticipate abnormal motor behaviors or dissociative transitions, thereby offering a mechanistic bridge between neural dynamics and clinical expression\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eiRBD was distinguished by a specific profile characterised by reduced fast sigma density and markedly diminished phase coupling (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003e). This decoupling was evident during both NREM and REM sleep, with the latter showing the most pronounced divergence from controls. Importantly, although the spatial distribution of sigma activity was broadly preserved, the combination of reduced event density and weakened alignment to the SO trough indicates a disruption in the temporal coordination of thalamocortical circuits, despite the presence of morphologically spindle-like oscillations. This dissociation between preserved structure and impaired timing aligns with evidence for early degeneration of brainstem nuclei and their cortical projections in α-synucleinopathies\u003csup\u003e\u003cspan additionalcitationids=\"CR29\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. Conventional spindle metrics, which prioritise amplitude or count, may therefore overlook early dysfunction. The disruption of SO-coupled temporal structure in iRBD may serve as a physiological marker of subcortical-cortical disintegration, potentially refining early detection strategies in at-risk populations\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eOf note, FM exhibited the most spatially and spectrally circumscribed alterations among the clinical groups\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. While spindle density and phase\u0026ndash;amplitude coupling metrics did not differ significantly from controls, increased TF-peak activity was observed in the theta (4\u0026ndash;6 Hz) and low-alpha (8\u0026ndash;10 Hz) bands during NREM\u0026thinsp;+\u0026thinsp;REM combined stages (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), with frontal and occipital derivations most affected\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. This enhancement of low-frequency activity may reflect increased cortical excitability or sensory gain, possibly consistent with prior reports of alpha intrusions and heightened arousability in FM\u003csup\u003e33,34\u003c/sup\u003e. The absence of fast sigma or phase-coupling disruptions suggests that thalamocortical rhythm generators remain functionally intact, despite localised spectral deviations. These findings align with the interpretation of FM as a condition of altered perceptual filtering rather than global network dysfunction, and they highlight the sensitivity of TF-peak methods to subtle, non-structural alterations in sleep microarchitecture.\u003c/p\u003e\u003cp\u003eStage-specific analyses confirmed that the most robust deviations occurred during NREM sleep. Across all disorders, NREM histograms and components revealed clearer separation from controls than REM sleep-only modes (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e4\u003c/span\u003e, \u003cb\u003eSI Appendix Fig. S5, S10-S11\u003c/b\u003e). REM sleep-stage data, although more variable, proved informative in RBD, where phase decoupling remained detectable. This asymmetry underscores the differential vulnerability of NREM and REM sleep to circuit-level disruption. While NREM sleep remains the dominant substrate for large-scale oscillatory synchrony, REM sleep may still expose more subtle instabilities, particularly in conditions involving degeneration or dysregulation of subcortical modulatory pathways.\u003c/p\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eClinical Implications\u003c/h2\u003e\u003cp\u003eThese findings indicate that transient oscillatory dynamics, referenced to slow oscillation (SO) power and phase, may offer a principled means of probing thalamocortical and subcortical function. The observed alterations, disorder-specific in topology and stage sensitivity, support a view of sleep microstructure as a physiologically meaningful readout of underlying circuit dynamics\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. Importantly, such features are not only mechanistically interpretable but also clinically accessible, requiring minimal instrumentation and offering millisecond-level temporal precision\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eThis approach may, thus, carry translational utility across a range of diagnostic contexts. For instance, the attenuation of SO-coupled phase synchrony in iRBD could serve as an early physiological indicator of brainstem-cortical disintegration, potentially anticipating neurodegenerative trajectories\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. In NT1, reduced spindle-SO alignment may reflect latent thalamocortical dysregulation, with relevance to cognitive and affective symptoms. More broadly, TF-peak\u0026ndash;derived signatures may aid in differential classification where behavioural phenotypes are ambiguous, and in refining diagnostic boundaries across syndromic spectra such as parasomnia and hypersomnia.\u003c/p\u003e\u003cp\u003eBeyond diagnosis, the high dimensionality and temporal specificity of this method may support longitudinal tracking in clinical trials, where changes in oscillatory coordination could precede overt symptomatology. Compared to conventional sleep metrics, such as spindle density or stage proportions, TF-peak histograms may arguably provide a richer, state-dependent fingerprint of network behaviour(14). Future studies incorporating longitudinal designs, cognitive phenotyping, and pharmacological modulation may clarify whether these microstructural features anticipate clinical progression, therapeutic response, or phenotypic conversion.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eLimitations and Future Directions\u003c/h2\u003e\u003cp\u003eSeveral important limitations must be acknowledged. First, this study is exploratory in nature and draws on modestly sized clinical samples, which limits both generalizability and the statistical power for detailed subgroup analyses. While effect sizes were moderate to large, replication in independent, larger cohorts is essential to confirm these findings. In particular, the fibromyalgia (FM) and non-REM parasomnia (NREMP) groups had modest sample sizes (n\u0026thinsp;=\u0026thinsp;11 and n\u0026thinsp;=\u0026thinsp;16, respectively), which, while typical in exploratory EEG studies of sleep pathology, limit statistical power for detecting subtler or regionally restricted effects. As such, the observed deviations in these cohorts, especially those limited to a single derivation or frequency band, should be interpreted cautiously. Larger replication samples are required to robustly confirm the presence, spatial extent, and clinical specificity of these alterations, especially given the known heterogeneity within both FM and NREMP populations.\u003c/p\u003e\u003cp\u003eSecond, the use of standard six-channel EEG constrains spatial resolution and precludes precise topographical or source-level inferences. Future studies employing high-density EEG or source-localized magnetoencephalography (MEG) may yield more anatomically specific insights into the spatial organization of transient oscillatory dynamics. Third, while PCA and ICA provided interpretable decompositions of SO-coupled features, both methods are sensitive to preprocessing decisions and inter-dataset variability. ICA, in particular, lacks component ordering and can be less stable across subsamples; thus, our interpretations emphasized PCA components, which demonstrated a strong reproducibility in split-half analyses. Alternative dimensionality reduction approaches, such as Uniform Manifold Approximation and Projection (UMAP) or supervised embedding, may reveal additional structure not captured here. Fourth, phase metrics in scalp EEG are inherently noisier, particularly during REM sleep, than power-based measures. Consequently, observed phase effects should be interpreted with caution, especially where effect sizes are small or trends did not reach statistical significance.\u003c/p\u003e\u003cp\u003eFifth, although classification performance was strong, particularly for NT1, iRBD, and NREMP, these results were derived from cross-validation within the same dataset used to construct PCA features. Without independent test sets, there is a risk of model overfitting. Moreover, the control and patient groups were diagnostically well-defined and non-overlapping, which may have accentuated group separability. As such, our models should not be interpreted as clinically diagnostic tools at this stage. Broader validation in heterogeneous, real-world populations with overlapping symptoms and comorbidities will be necessary to determine translational applicability. The development of a publicly available benchmarking dataset for TF-peak\u0026ndash;based phenotyping could also accelerate reproducibility and standardization across laboratories. Finally, although we standardized referencing during preprocessing, control and patient datasets were recorded using different reference schemes, which may introduce residual variance. While all analyses were conducted on re-referenced and standardized data, and statistical tests were channel-specific, future work using uniformly acquired EEG across cohorts would help mitigate this potential confound. The observed component patterns and group differences were robust across derivations and not restricted to any single site or channel, suggesting minimal influence from data acquisition variability.\u003c/p\u003e\u003cp\u003eImportantly, this study does not establish causal or longitudinal relationships between the observed oscillatory features and clinical outcomes. Longitudinal designs, particularly in at-risk populations such as idiopathic RBD, will be essential to determine whether the observed deviations in transient dynamics have prognostic utility. Similarly, the behavioural correlates of these electrophysiological features, such as memory consolidation, arousal thresholds, or dream phenomenology, remain important avenues for future investigation.\u003c/p\u003e\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eBy mapping transient sleep oscillations onto continuous measures of SO power and phase, this study reveals reproducible alterations across several clinical populations. These deviations, whether reductions in synchronized spindling, phase dispersion, or increased sub-sigma activity, reflect distinct pathophysiological signatures of network disruption. Together, our findings argue for a redefinition of sleep microstructure not in terms of discrete events, but as a spectrum of dynamic, state-dependent oscillatory activity. This paradigm may inform the development of electrophysiological phenotypes in sleep medicine, with future applications in diagnosis, monitoring, and mechanistic understanding.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003eA retrospective exploratory cross-sectional study was conducted on 99 adults (≥18 years) across five diagnostic categories: idiopathic REM sleep behavior disorder (iRBD; n=17), narcolepsy type 1 (NT1; n=16), non-REM parasomnia (NREMP; n=16), fibromyalgia syndrome (FM; n=11), and healthy controls (n=39; from The Montreal Archive of Sleep Studies) \u003csup\u003e36\u003c/sup\u003e. All clinical diagnoses were made in accordance with the International Classification of Sleep Disorders – Third Edition (ICSD-3) and confirmed by board-certified sleep specialists\u003csup\u003e16,17\u003c/sup\u003e.\u0026nbsp;Participants included were ≥18 years of age and had no major psychiatric or neurological comorbidities, substance dependence, or use of medications known to alter sleep architecture.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEthical approval for the study was granted by the institutional Research Ethics Committee (Project No. 12436) \u003csup\u003e37,38\u003c/sup\u003e. The analysis was conducted on fully anonymized retrospective data, in compliance with the UK Data Protection Act and the General Data Protection Regulation (GDPR) (Regulation (EU) 2016/679). Informed consent was not required due to the retrospective design and the use of non-identifiable data\u003csup\u003e37,38\u003c/sup\u003e. The study was carried out in accordance with the Declaration of Helsinki(WMA, 2013).\u003c/p\u003e\n\u003cp\u003eAll overnight polysomnographic (PSG) recordings included standard six-channel EEG using a 10–20 montage (F3, F4, C3, C4, O1, O2). Referencing schemes were harmonized during preprocessing to ensure consistency across datasets. Sleep staging was performed manually in accordance with American Academy of Sleep Medicine (AASM) criteria\u003csup\u003e39\u003c/sup\u003e. Summary socio-demographic and sleep architecture metrics are provided in \u003cstrong\u003eSI Appendix, Table S1.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEEG Preprocessing and TF-Peak Extraction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEEG preprocessing was conducted using MNE-Python (v1.5.0) \u003csup\u003e40\u003c/sup\u003e and the DYNAM-O toolbox (version 1.0) \u003csup\u003e14\u003c/sup\u003e. Signals were down-sampled to 100 Hz, bandpass filtered (0.1–40 Hz), and re-referenced. Channels with persistent artefacts were excluded per DYNAM-O's artefact rejection pipeline.\u003c/p\u003e\n\u003cp\u003eTF-peaks were extracted using a watershed-based algorithm that identifies local maxima across the 4–25 Hz frequency range \u003csup\u003e14\u003c/sup\u003e. Each TF-peak was annotated by the concurrent SO power (0.3–1.5 Hz) and SO phase at its time of occurrence, yielding two-dimensional histograms per channel for both power- and phase-anchored events.\u003c/p\u003e\n\u003cp\u003eHistograms were generated for three conditions: NREM+REM combined, NREM-only, and REM-only. For REM, only the interquartile range (25–75%) of SO power was included to reduce noise, reflecting the lower SO amplitude in this stage.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical Decomposition and Group Comparisons\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEach histogram (per stage and channel) was standardized \u003csup\u003e41\u003c/sup\u003e and submitted to PCA \u003csup\u003e42\u003c/sup\u003e and ICA \u003csup\u003e43\u003c/sup\u003e using Scikit-learn (v1.3.2) \u003csup\u003e44\u003c/sup\u003e. The first 10 PCA components (capturing ~70% of total variance) were retained. ICA was performed with 5 components for SO-power and 3–4 for SO-phase histograms, based on convergence criteria.\u003c/p\u003e\n\u003cp\u003eGroup comparisons were conducted using Kruskal–Wallis tests \u003csup\u003e45\u003c/sup\u003e across all five groups, with false discovery rate (FDR) control via the Benjamini–Yekutieli method \u003csup\u003e46\u003c/sup\u003e (α = 0.1). When significant, post hoc Mann–Whitney U tests \u003csup\u003e47\u003c/sup\u003e were performed between controls and each patient group, with Bonferroni-adjusted \u003csup\u003e48\u003c/sup\u003e p-values (p ≤ 0.05) reported in \u003cstrong\u003eSI Appendix, Tables S2-S28\u003c/strong\u003e.\u0026nbsp;η2 effect size\u0026nbsp;\u003csup\u003e49\u003c/sup\u003e was calculated for all the cases (\u003cstrong\u003eTables S2-S28\u003c/strong\u003e), and Cliff’s delta metric\u0026nbsp;\u003csup\u003e50\u003c/sup\u003ewas additionally computed for key components.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReliability and Classification Analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSplit-half reliability was assessed by randomly dividing subjects into balanced subsets, performing PCA separately, and correlating eigenvectors between subsets and the full sample. Components with Spearman \u003csup\u003e51\u003c/sup\u003e r \u0026gt; |0.75| (first five PCs) were deemed highly reliable; those with r \u0026gt; |0.5| moderately reliable.\u003c/p\u003e\n\u003cp\u003eTo evaluate the discriminative potential of histogram-derived PCA features, we trained logistic regression models \u003csup\u003e41,52\u003c/sup\u003e using five-fold cross-validation \u003csup\u003e53,54\u003c/sup\u003e. Each subject served once in the validation fold. Performance was measured by AUC-ROC \u003csup\u003e55\u003c/sup\u003e. Statistical significance of classification was tested via 1,000 permutation tests \u003csup\u003e56\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVisualization and Software\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData visualizations were generated using DYNAM-O (v1.0) \u003csup\u003e14\u003c/sup\u003e, Matplotlib (v3.8.0) \u003csup\u003e57\u003c/sup\u003e, and Seaborn (v0.13.0) \u003csup\u003e58\u003c/sup\u003e. Statistical calculations were performed by SciPy (v1.11.4) \u003csup\u003e59\u003c/sup\u003e and Statsmodels (v0.14.1) \u003csup\u003e46\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors gratefully acknowledge current and former members of the Sleep and Brain Plasticity Laboratory, as well as colleagues from the Sleep Disorders Centre at Guy\u0026rsquo;s Hospital, for their contributions, support, and collaborative input throughout the development of this work. We extend our sincere thanks to the patients and patient advocacy groups whose participation and insights helped shape the aims and direction of this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: Ivana Rosenzweig, Robert Leech\u003c/p\u003e\n\u003cp\u003eMethodology and Analysis: Olga Ivanenko, Nazanin Biabani, Zoran Cvetkovic, Robert Leech\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; Original Draft: Nazanin Biabani, Olga Ivanenko, Ivana Rosenzweig\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; Review \u0026amp; Editing: All authors\u003c/p\u003e\n\u003cp\u003eSupervision: Ivana Rosenzweig, Robert Leech\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and Code Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe custom analysis code used in this study will be made publicly available via GitHub upon publication [link]. Due to ethical and legal restrictions, the raw clinical EEG data cannot be shared. These data include sensitive health information and are governed by the data protection policies of Guy\u0026rsquo;s and St Thomas\u0026rsquo; NHS Foundation Trust and King\u0026rsquo;s College London. Requests for access to derived data or anonymized summary metrics may be considered on a case-by-case basis by the corresponding author and are subject to review by the Trust\u0026rsquo;s Research and Development Office and the institutional Data Protection Officer, in accordance with GDPR and NHS research governance frameworks.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interest Statement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was funded in whole, or in part, by the Wellcome Trust (103952/Z/14/Z). For the purpose of open access, the author IR has applied a CC BY public copyright license to any author accepted manuscript version arising from this submission.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis article represents independent research in part funded by the NIHR Maudsley Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King\u0026rsquo;s College London.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAndrillon, T. \u0026amp; Oudiette, D. What is sleep exactly? Global and local modulations of sleep oscillations all around the clock. \u003cem\u003eNeuroscience \u0026amp; Biobehavioral Reviews\u003c/em\u003e \u003cstrong\u003e155\u003c/strong\u003e, 105465 (2023).\u003c/li\u003e\n\u003cli\u003ePeyrache, A.\u003cem\u003e et al.\u003c/em\u003e Spatiotemporal dynamics of neocortical excitation and inhibition during human sleep. \u003cem\u003eProceedings of the national academy of sciences\u003c/em\u003e \u003cstrong\u003e109\u003c/strong\u003e, 1731-1736 (2012).\u003c/li\u003e\n\u003cli\u003eOudiette, D., Antony, J. W., Creery, J. D. \u0026amp; Paller, K. A. The role of memory reactivation during wakefulness and sleep in determining which memories endure. \u003cem\u003eJournal of Neuroscience\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 6672-6678 (2013).\u003c/li\u003e\n\u003cli\u003eSteriade, M. \u0026amp; Deschenes, M. The thalamus as a neuronal oscillator. \u003cem\u003eBrain Research Reviews\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 1-63 (1984).\u003c/li\u003e\n\u003cli\u003eBuzsaki, G. \u0026amp; Draguhn, A. Neuronal oscillations in cortical networks. \u003cem\u003escience\u003c/em\u003e \u003cstrong\u003e304\u003c/strong\u003e, 1926-1929 (2004).\u003c/li\u003e\n\u003cli\u003eHorie, K.\u003cem\u003e et al.\u003c/em\u003e Automated sleep stage scoring employing a reasoning mechanism and evaluation of its explainability. \u003cem\u003eScientific reports\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 12799 (2022).\u003c/li\u003e\n\u003cli\u003eHimanen, S.-L. \u0026amp; Hasan, J. Limitations of rechtschaffen and kales. \u003cem\u003eSleep medicine reviews\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 149-167 (2000).\u003c/li\u003e\n\u003cli\u003eMoser, D.\u003cem\u003e et al.\u003c/em\u003e Sleep classification according to AASM and Rechtschaffen \u0026amp; Kales: effects on sleep scoring parameters. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 139-149 (2009).\u003c/li\u003e\n\u003cli\u003eKlinzing, J. G.\u003cem\u003e et al.\u003c/em\u003e Spindle activity phase-locked to sleep slow oscillations. \u003cem\u003eNeuroimage\u003c/em\u003e \u003cstrong\u003e134\u003c/strong\u003e, 607-616 (2016).\u003c/li\u003e\n\u003cli\u003eCakan, C.\u003cem\u003e et al.\u003c/em\u003e Spatiotemporal patterns of adaptation-induced slow oscillations in a whole-brain model of slow-wave sleep. \u003cem\u003eFrontiers in computational neuroscience\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 800101 (2022).\u003c/li\u003e\n\u003cli\u003eOlbrich, E. \u0026amp; Achermann, P. Analysis of oscillatory patterns in the human sleep EEG using a novel detection algorithm. \u003cem\u003eJournal of sleep research\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 337-346 (2005).\u003c/li\u003e\n\u003cli\u003eDimitrov, T., He, M., Stickgold, R. \u0026amp; Prerau, M. J. Sleep spindles comprise a subset of a broader class of electroencephalogram events. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e, doi:10.1093/sleep/zsab099 (2021).\u003c/li\u003e\n\u003cli\u003eMylonas, D.\u003cem\u003e et al.\u003c/em\u003e A Novel Approach to Estimating the Cortical Sources of Sleep Spindles Using Simultaneous EEG/MEG. \u003cem\u003eFrontiers in Neurology\u003c/em\u003e \u003cstrong\u003eVolume 13 - 2022\u003c/strong\u003e, doi:10.3389/fneur.2022.871166 (2022).\u003c/li\u003e\n\u003cli\u003eStokes, P. A.\u003cem\u003e et al.\u003c/em\u003e Transient oscillation dynamics during sleep provide a robust basis for electroencephalographic phenotyping and biomarker identification. \u003cem\u003eSleep\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, zsac223, doi:10.1093/sleep/zsac223 (2023).\u003c/li\u003e\n\u003cli\u003eChen, S., He, M., Brown, R. E., Eden, U. T. \u0026amp; Prerau, M. J. Individualized temporal patterns drive human sleep spindle timing. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e, e2405276121, doi:10.1073/pnas.2405276121 (2025).\u003c/li\u003e\n\u003cli\u003eTreede, R. D.\u003cem\u003e et al.\u003c/em\u003e Chronic pain as a symptom or a disease: the IASP Classification of Chronic Pain for the International Classification of Diseases (ICD-11). \u003cem\u003ePain\u003c/em\u003e \u003cstrong\u003e160\u003c/strong\u003e, 19-27, doi:10.1097/j.pain.0000000000001384 (2019).\u003c/li\u003e\n\u003cli\u003eSateia, M. J. International classification of sleep disorders-third edition: highlights and modifications. \u003cem\u003eChest\u003c/em\u003e \u003cstrong\u003e146\u003c/strong\u003e, 1387-1394, doi:10.1378/chest.14-0970 (2014).\u003c/li\u003e\n\u003cli\u003eGibbs, S. A.\u003cem\u003e et al.\u003c/em\u003e Sleep-related epileptic behaviors and non-REM-related parasomnias: Insights from stereo-EEG. \u003cem\u003eSleep Medicine Reviews\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 4-20, doi:https://doi.org/10.1016/j.smrv.2015.05.002 (2016).\u003c/li\u003e\n\u003cli\u003eH\u0026ouml;gl, B., Stefani, A. \u0026amp; Videnovic, A. Idiopathic REM sleep behaviour disorder and neurodegeneration \u0026mdash; an update. \u003cem\u003eNature Reviews Neurology\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 40-55, doi:10.1038/nrneurol.2017.157 (2018).\u003c/li\u003e\n\u003cli\u003eThannickal, T. C.\u003cem\u003e et al.\u003c/em\u003e Reduced number of hypocretin neurons in human narcolepsy. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 469-474 (2000).\u003c/li\u003e\n\u003cli\u003eHelfrich, R. F., Mander, B. A., Jagust, W. J., Knight, R. T. \u0026amp; Walker, M. P. Old brains come uncoupled in sleep: slow wave-spindle synchrony, brain atrophy, and forgetting. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e97\u003c/strong\u003e, 221-230. e224 (2018).\u003c/li\u003e\n\u003cli\u003eDenis, D.\u003cem\u003e et al.\u003c/em\u003e Slow oscillation‐spindle coupling is negatively associated with emotional memory formation following stress. \u003cem\u003eEuropean Journal of Neuroscience\u003c/em\u003e \u003cstrong\u003e55\u003c/strong\u003e, 2632-2650 (2022).\u003c/li\u003e\n\u003cli\u003eHarel, B. T.\u003cem\u003e et al.\u003c/em\u003e The nature and magnitude of cognitive impairment in narcolepsy type 1, narcolepsy type 2, and idiopathic hypersomnia: a meta-analysis. \u003cem\u003eSLEEP Advances\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, doi:10.1093/sleepadvances/zpae043 (2024).\u003c/li\u003e\n\u003cli\u003eCastelnovo, A., Lopez, R., Proserpio, P., Nobili, L. \u0026amp; Dauvilliers, Y. NREM sleep parasomnias as disorders of sleep-state dissociation. \u003cem\u003eNature Reviews Neurology\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 470-481 (2018).\u003c/li\u003e\n\u003cli\u003eLecci, S.\u003cem\u003e et al.\u003c/em\u003e Coordinated infraslow neural and cardiac oscillations mark fragility and offline periods in mammalian sleep. \u003cem\u003eScience Advances\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, e1602026, doi:doi:10.1126/sciadv.1602026 (2017).\u003c/li\u003e\n\u003cli\u003eCastelnovo, A.\u003cem\u003e et al.\u003c/em\u003e Conscious experiences during non-rapid eye movement sleep parasomnias. \u003cem\u003eNeuroscience \u0026amp; Biobehavioral Reviews\u003c/em\u003e \u003cstrong\u003e167\u003c/strong\u003e, 105919, doi:https://doi.org/10.1016/j.neubiorev.2024.105919 (2024).\u003c/li\u003e\n\u003cli\u003eMutti, C.\u003cem\u003e et al.\u003c/em\u003e Commonalities and Differences in NREM Parasomnias and Sleep-Related Epilepsy: Is There a Continuum Between the Two Conditions? \u003cem\u003eFront Neurol\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 600026, doi:10.3389/fneur.2020.600026 (2020).\u003c/li\u003e\n\u003cli\u003eBarone, D. A. \u0026amp; Henchcliffe, C. Rapid eye movement sleep behavior disorder and the link to alpha-synucleinopathies. \u003cem\u003eClinical Neurophysiology\u003c/em\u003e \u003cstrong\u003e129\u003c/strong\u003e, 1551-1564 (2018).\u003c/li\u003e\n\u003cli\u003eGrinberg, L. T., Rueb, U. \u0026amp; Heinsen, H. Brainstem: Neglected Locus in Neurodegenerative Diseases. \u003cem\u003eFrontiers in Neurology\u003c/em\u003e \u003cstrong\u003evolume 2 - 2011\u003c/strong\u003e, doi:10.3389/fneur.2011.00042 (2011).\u003c/li\u003e\n\u003cli\u003eGelegen, C.\u003cem\u003e et al.\u003c/em\u003e Relevance of sleep and associated structural changes in GBA1 mouse to human rapid eye movement behavior disorder. \u003cem\u003eScientific Reports\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 7973, doi:10.1038/s41598-022-11516-x (2022).\u003c/li\u003e\n\u003cli\u003eIranzo, A.\u003cem\u003e et al.\u003c/em\u003e Rapid-eye-movement sleep behaviour disorder as an early marker for a neurodegenerative disorder: a descriptive study. \u003cem\u003eThe Lancet Neurology\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, 572-577 (2006).\u003c/li\u003e\n\u003cli\u003eBigatti, S. M., Hernandez, A. M., Cronan, T. A. \u0026amp; Rand, K. L. Sleep disturbances in fibromyalgia syndrome: relationship to pain and depression. \u003cem\u003eArthritis and rheumatism\u003c/em\u003e \u003cstrong\u003e59\u003c/strong\u003e, 961-967, doi:10.1002/art.23828 (2008).\u003c/li\u003e\n\u003cli\u003eVijayan, S., Klerman, E. B., Adler, G. K. \u0026amp; Kopell, N. J. Thalamic mechanisms underlying alpha-delta sleep with implications for fibromyalgia. \u003cem\u003eJournal of neurophysiology\u003c/em\u003e \u003cstrong\u003e114\u003c/strong\u003e, 1923-1930, doi:10.1152/jn.00280.2015 (2015).\u003c/li\u003e\n\u003cli\u003eDrewes, A. M., Nielsen, K. D., Jennum, P. \u0026amp; Andreasen, A. Alpha intrusion in fibromyalgia. \u003cem\u003eJournal of Musculoskeletal Pain\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 223-228 (1993).\u003c/li\u003e\n\u003cli\u003eBoeve, B. F., Silber, M. H., Ferman, T. J., Lucas, J. A. \u0026amp; Parisi, J. E. Association of REM sleep behavior disorder and neurodegenerative disease may reflect an underlying synucleinopathy. \u003cem\u003eMovement disorders: official journal of the Movement Disorder Society\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 622-630 (2001).\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Reilly, C., Gosselin, N., Carrier, J. \u0026amp; Nielsen, T. Montreal Archive of Sleep Studies: an open-access resource for instrument benchmarking and exploratory research. \u003cem\u003eJ Sleep Res\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 628-635, doi:10.1111/jsr.12169 (2014).\u003c/li\u003e\n\u003cli\u003eWorld Medical, A. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. \u003cem\u003eJAMA\u003c/em\u003e \u003cstrong\u003e310\u003c/strong\u003e, 2191-2194, doi:10.1001/jama.2013.281053 (2013).\u003c/li\u003e\n\u003cli\u003eCouncil, E. P. a. o. t. Vol. Regulation (EU) 2016/679 (ed EU) (Official Journal of the European Union, 2016).\u003c/li\u003e\n\u003cli\u003eIber, C. The AASM manual for the scoring of sleep and associated events: rules, terminology, and technical specification. \u003cem\u003e(No Title)\u003c/em\u003e (2007).\u003c/li\u003e\n\u003cli\u003eGramfort, A.\u003cem\u003e et al.\u003c/em\u003e MEG and EEG data analysis with MNE-Python. \u003cem\u003eFront Neurosci\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 267, doi:10.3389/fnins.2013.00267 (2013).\u003c/li\u003e\n\u003cli\u003eBishop, C. M. \u0026amp; Nasrabadi, N. M. \u003cem\u003ePattern recognition and machine learning\u003c/em\u003e. Vol. 4 (Springer, 2006).\u003c/li\u003e\n\u003cli\u003eTipping, M. E. \u0026amp; Bishop, C. M. Probabilistic principal component analysis. \u003cem\u003eJournal of the Royal Statistical Society Series B: Statistical Methodology\u003c/em\u003e \u003cstrong\u003e61\u003c/strong\u003e, 611-622 (1999).\u003c/li\u003e\n\u003cli\u003eHyvarinen, A. \u0026amp; Oja, E. Independent component analysis: algorithms and applications. \u003cem\u003eNeural Netw\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 411-430, doi:10.1016/s0893-6080(00)00026-5 (2000).\u003c/li\u003e\n\u003cli\u003ePedregosa, F.\u003cem\u003e et al.\u003c/em\u003e Scikit-learn: Machine learning in Python. \u003cem\u003ethe Journal of machine Learning research\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 2825-2830 (2011).\u003c/li\u003e\n\u003cli\u003eKruskal, W. H. \u0026amp; Wallis, W. A. Use of ranks in one-criterion variance analysis. \u003cem\u003eJournal of the American statistical Association\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 583-621 (1952).\u003c/li\u003e\n\u003cli\u003eBenjamini, Y. \u0026amp; Yekutieli, D. The control of the false discovery rate in multiple testing under dependency. \u003cem\u003eAnnals of statistics\u003c/em\u003e, 1165-1188 (2001).\u003c/li\u003e\n\u003cli\u003eMann, H. B. \u0026amp; Whitney, D. R. On a test of whether one of two random variables is stochastically larger than the other. \u003cem\u003eThe annals of mathematical statistics\u003c/em\u003e, 50-60 (1947).\u003c/li\u003e\n\u003cli\u003eBonferroni, C. E. \u003cem\u003eTeoria statistica delle classi e calcolo delle probabilit\u0026agrave;\u003c/em\u003e. (Seeber, 1936).\u003c/li\u003e\n\u003cli\u003eCohen, J. \u003cem\u003eStatistical Power Analysis for the Behavioral Sciences\u003c/em\u003e. (Academic Press, 1969).\u003c/li\u003e\n\u003cli\u003eCliff, N. Dominance statistics: Ordinal analyses to answer ordinal questions. \u003cem\u003ePsychological Bulletin\u003c/em\u003e \u003cstrong\u003e114\u003c/strong\u003e, 494-509, doi:10.1037/0033-2909.114.3.494 (1993).\u003c/li\u003e\n\u003cli\u003eSpearman, C. The proof and measurement of association between two things. (1961).\u003c/li\u003e\n\u003cli\u003eFan, R.-E., Chang, K.-W., Hsieh, C.-J., Wang, X.-R. \u0026amp; Lin, C.-J. LIBLINEAR: A library for large linear classification. \u003cem\u003ethe Journal of machine Learning research\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 1871-1874 (2008).\u003c/li\u003e\n\u003cli\u003eStone, M. Cross‐validatory choice and assessment of statistical predictions. \u003cem\u003eJournal of the royal statistical society: Series B (Methodological)\u003c/em\u003e \u003cstrong\u003e36\u003c/strong\u003e, 111-133 (1974).\u003c/li\u003e\n\u003cli\u003eHastie, T., Tibshirani, R., Friedman, J. H. \u0026amp; Friedman, J. H. \u003cem\u003eThe elements of statistical learning: data mining, inference, and prediction\u003c/em\u003e. Vol. 2 (Springer, 2009).\u003c/li\u003e\n\u003cli\u003eHanley, J. A. \u0026amp; McNeil, B. J. The meaning and use of the area under a receiver operating characteristic (ROC) curve. \u003cem\u003eRadiology\u003c/em\u003e \u003cstrong\u003e143\u003c/strong\u003e, 29-36, doi:10.1148/radiology.143.1.7063747 (1982).\u003c/li\u003e\n\u003cli\u003eOjala, M. \u0026amp; Garriga, G. C. Permutation tests for studying classifier performance. \u003cem\u003eJournal of machine learning research\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e (2010).\u003c/li\u003e\n\u003cli\u003eHunter, J. D. Matplotlib: A 2D graphics environment. \u003cem\u003eComputing in science \u0026amp; engineering\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 90-95 (2007).\u003c/li\u003e\n\u003cli\u003eWaskom, M. L. Seaborn: statistical data visualization. \u003cem\u003eJournal of Open Source Software\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 3021 (2021).\u003c/li\u003e\n\u003cli\u003eVirtanen, P.\u003cem\u003e et al.\u003c/em\u003e SciPy 1.0: fundamental algorithms for scientific computing in Python. \u003cem\u003eNature methods\u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e, 261-272 (2020).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"sleep microstructure, slow oscillations, sleep spindles, time-frequency analysis, sleep disorders","lastPublishedDoi":"10.21203/rs.3.rs-7660761/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7660761/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eTransient sleep oscillations reflect the dynamic coordination of cortical and subcortical circuits, modulated by slow oscillatory activity. However, the disorder-specific signatures of these events across neurological, pain, and sleep disorders remain poorly characterized. In this exploratory study, we analyzed transient oscillatory dynamics in 99 individuals, including healthy controls and patients with narcolepsy type 1, non-REM parasomnia, idiopathic REM sleep behavior disorder, and fibromyalgia syndrome. Using slow oscillatory referenced time-frequency peak histograms, we applied principal and independent component analysis to uncover spectral and phase-coupling patterns across non-REM and REM stages. We identified reproducible, trait-like oscillatory structures in controls and disorder-specific deviations in patient groups, particularly during NREM sleep. Specifically, patients with narcolepsy type 1 and non-REM parasomnia exhibited altered fast sigma coupling and phase dispersion, while idiopathic REM sleep behavior disorder patients showed reduced fast sigma density and diminished phase synchrony, despite retention of spindle-like spectral structure. In internal cross-validation, slow oscillatory-power features supported robust group-level discrimination in select EEG derivations; however, broader validation in independent samples is required. These findings highlight distinctive, stage-specific microstructural alterations in sleep and pain pathologies and support the future potential of time-frequency peak analysis as a non-invasive tool for phenotyping thalamocortical and subcortical circuit function.\u003c/p\u003e","manuscriptTitle":"Disorder-specific alterations of transient oscillatory dynamics during sleep across cortical and subcortical networks","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-08 07:08:50","doi":"10.21203/rs.3.rs-7660761/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-27T05:16:36+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-06T06:27:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-02T08:37:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"49826970535841857098160968972314987655","date":"2025-09-29T06:24:43+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"287617689995397806840969467228327834744","date":"2025-09-26T05:41:17+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-24T09:26:12+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-24T04:33:51+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-22T21:53:10+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-20T05:09:19+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-09-19T18:02:37+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"80195bad-2cca-4a94-9174-3d34232fbb0c","owner":[],"postedDate":"October 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":55763884,"name":"Health sciences/Neurology"},{"id":55763885,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2026-01-26T16:12:24+00:00","versionOfRecord":{"articleIdentity":"rs-7660761","link":"https://doi.org/10.1038/s41598-025-33669-1","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2026-01-20 15:58:27","publishedOnDateReadable":"January 20th, 2026"},"versionCreatedAt":"2025-10-08 07:08:50","video":"","vorDoi":"10.1038/s41598-025-33669-1","vorDoiUrl":"https://doi.org/10.1038/s41598-025-33669-1","workflowStages":[]},"version":"v1","identity":"rs-7660761","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7660761","identity":"rs-7660761","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}