Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia

Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia | 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 Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia Gabor Csukly, Frigyes Racz, Kinga Farkas, Melinda Becske, Hajnalka Molnar, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6908048/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract Spectral features of the electroencephalogram (EEG) are essential for providing clinically relevant biomarkers in schizophrenia (SZ). Despite literature indicating altered short-scale neural dynamics in SZ, however, band-limited power (BLP) indices are rarely assessed in a time-resolved manner. To address this, here we evaluated static and dynamic BLP indices in a sample of 30 SZ patients and 31 healthy control (HC) individuals. Guided by recent findings on power spectral dynamics in SZ, we estimated total, and also decomposed fractal and oscillatory BLP in a sliding window manner from resting-state EEG recordings collected in eyes-closed (EC) resting-state. The SZ cohort was characterized by elevated baseline (mean over time) relative power in lower frequency regimes (delta, theta), mainly attributable to aperiodic fractal activity. In the higher regimes (alpha, beta), however, baseline was similar to HC with instead a widespread lessening in temporal fluctuations of both fractal and oscillatory activity. Variability in beta-BLP over the dorsal attention network was found correlated with negative symptoms in SZ. Finally, surrogate data testing indicated a loss of nonlinearity in neural dynamics as a potential mechanism for diminished BLP fluctuations. Health sciences/Diseases/Psychiatric disorders/Schizophrenia Biological sciences/Neuroscience Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Developing and understanding biomarkers of schizophrenia (SZ) is a critical endeavor: such physiology-based features can not only advance understanding the pathomechanisms of this disorder but could also directly affect clinical practice that is yet heavily burdened with costly trial-and-error practices, especially in terms of therapy planning and optimization [ 1 ]. For instance, the inability to select the most ideal pharmaceutical agent in advance can lead to prolongment of effective therapy, increasing patient distress and impairment, reduction of treatment engagement, and an increased risk of danger to self and others [ 2 ]. The relevance of this issue is illustrated by the fact that despite having a lifetime prevalence just under 1\% [ 3 ], SZ draws the most research effort in identifying personalized biomarkers second only to depression [ 4 ]. Among others, functional neuroimaging markers intuitively appear as a suitable choice for this purpose [ 5 ], providing an expanding knowledge on potential diagnostic features, future pharmaceutical targets or early indicators of disease progression and therapy response [ 1 , 4 ]. Currently, the most consistent resting-state electroencephalography (EEG) alterations identified in SZ include an increase in delta- and theta-band activity, while a decrease in alpha power [ 6 , 7 ], though contradictory findings are also reported [ 8 ]. In contrast to resting-state EEG, evoked potentials such as the P50, P300 or mismatch negativity also show potential not only to distinguish between SZ patients and healthy individuals, but even to provide some predictive insight on disease progression [ 8 , 9 ]. Furthermore, extensive dynamic functional connectivity literature indicates that short-scale neural dynamics are altered in SZ, often in a way that is undetectable for conventional, time-invariant analytical approaches [ 10 , 11 ]. Interestingly, however, despite all this evidence, the temporal dynamics of resting-state EEG power spectral features are rarely considered in SZ. Altered cortical excitation/inhibition (E/I) ratio – which can be characterized via the 1/f slope ( \(\:\beta\:\) ) of electrophysiological signals [ 12 ] – is another pathological feature that is considered central in SZ [ 13 ] with explicit potential for clinical utility [ 14 , 15 ]. Since there are indications that the \(\:1/f\) slope might fluctuate over time in SZ [ 16 ], in our recent study [ 17 ] we performed a time-resolved analysis of EEG spectral slopes along these considerations. This analysis revealed that while the overall baseline (average over time) of the E/I ratio is comparable in patients with SZ and healthy controls (HC), the patient group showed significantly diminished temporal variability in \(\:\beta\:\) in the 20–45 Hz regime. These changes were most consistent over cortical locations corresponding to the dorsal attention network, and they were also correlated to clinical symptom scores. In that work, however, we focused exclusively on the spectral slope and yet omitted conventional EEG measures such as spectral power in canonical frequency regimes [ 6 , 8 ]. Therefore, our goal here was to provide a comprehensive picture and understanding of dynamic EEG alterations in SZ by conducting a time-resolved analysis of EEG spectra decomposed into aperiodic \(\:1/f\) and oscillatory components, hypothesizing diminished temporal variability in fractal spectral power in higher-frequency regimes that (iii) correlate with clinical symptom scores as captured via the Positive and Negative Syndrome Scale (PANSS) [ 18 ]. Materials and Methods Participants and Clinical Measures We analyzed EEG recordings of the same study cohorts as in our recent study; therefore, here we only provide a brief description of the study populations and EEG data collection protocols with more details, please see [ 17 ] The SZ group was comprised of 30 SZ patients (11 female, age: 33.07 ± 9.73 years), with an age- and sex-matched healthy control (HC) group of 31 individuals (13 female, age: 33.06 ± 10.31 years). There was no statistically significant difference between the two groups in terms of age, proportion of sexes, or years in education. The Semmelweis Regional and Institutional Committee of Science and Research Ethics reviewed and approved the study (approval number: 197/2015), which was conducted in line with the Declaration of Helsinki. All participants were informed about study details and provided prior written consent for participation. Further study cohort details and inclusion/exclusion criteria are provided in the Supplementary Material . EEG Recording, Quality Control and Pre-processing EEG was collected from 64 standard 10–10 locations at 1000 Hz resolution using a Neuroscan amplifier. Channels were referenced to linked mastoids and electrode impedances were kept under \(\:5k{\Omega\:}\) . Resting state data was collected for two minutes in eyes closed (EC) resting-state condition. Data was visually inspected by two investigators independently, blinded to study group association; segments free of overt artifacts were selected and only data segments deemed admissible by both investigators were included for further analysis. Channels AF7, AF8, AFz, F1, F2 TP7 and TP8 (along with M1/M2) were excluded from further analysis as they were identified as noisy/bad in more than 10 participants according to standard criteria [ 17 ], resulting in a final channel count of 55. The electrode montage (generated using the BrainNet Viewer tool [ 19 ]) is illustrated on Fig. 1 A. Finally, 30-second segments were selected randomly from the admitted data for each participant for further analysis. Data was processed and analyzed in Matlab (MathWorks, Natick, MA) using the EELAB toolbox [ 20 ] along with custom functions and scripts. Data was first band-pass filtered between 0.1 and 128 Hz using the default finite impulse response filter of EEGLAB with additional notch filters at 50 Hz and 100 Hz, and then downsampled to 256 Hz for computational efficiency. EEG signals were then decomposed via independent component analysis (ICA) and artifacts corresponding to non-neural sources (such as potential eye movements, muscle activity, heart activity, head movement) were identified and removed in an automatized manner using the Multiple Artifact Rejection Algorithm (MARA) toolbox [ 21 , 22 ] before reverse-ICA transformation. This resulted in the removal of 27.44 ± 8.22 out of 55 components, respectively. There was no statistical difference ( p > 0.35) in the number of removed components between HC and SZ (see online Supplementary Material ) Finally, after re-referencing to the common average, EEG signals were standardized (z-scored) according to \(\:z\left(t\right)=\left(x\right(t)-\mu\:(x\left)\right)/\sigma\:\left(x\right)\) , where \(\:x\left(t\right)\) is the original time series with \(\:z\left(t\right)\) its z-scored version, and \(\:\mu\:\left(x\right)\) and \(\:\sigma\:\left(x\right)\) are the mean and standard deviation of \(\:x\left(t\right)\) , respectively. As z-scoring renders all time series zero-centered and with a standard deviation of 1, this step was introduced to help avoid potential confounding effects on power estimates by varying impedances across channels and subjects. Data analysis We employed a sliding window approach to obtain time-resolved band-limited power (BLP) estimates after separating \(\:1/f\) aperiodic and oscillatory components of the EEG spectra. In that, we utilized a window size of 8 seconds and a step size of 1 second (87.5% overlap), resulting in 23 estimates from the 30-second epochs. Note that with z-scoring performed on the complete 30-second segments, these window estimates reflect distribution of power along the complete broadband spectrum, while at the same time capturing relevant dynamics over the 30-second period (in contrast to z-scoring in every window individually). Similar to our previous approach [ 17 ], in every window we decomposed the raw (mixed) power spectrum into fractal and oscillatory components using the Irregular Resampling Auto-Spectral Analysis (IRASA) technique [ 23 ]. The IRASA technique is illustrated on Fig. 1 B, while for details, please see the original publication of Wen & Liu [ 23 ]. Frequency resolution was set to 0.0625 Hz and the resampling factor \(\:h\) ranged from 1.1 to 2.6 in increments of 0.1 [ 17 ]. The following conventional frequency bands were defined: delta ( \(\:\delta\:\) ): 1–4 Hz, theta ( \(\:\theta\:\) ): 4–8 Hz, alpha ( \(\:\alpha\:\) ): 8–13 Hz, beta ( \(\:\beta\:\) ): 13–25 Hz and gamma ( \(\:\gamma\:\) ): 25–45 Hz. Frequency regimes over 45 Hz were inaccessible due to the filtering effects introduced by resampling in IRASA [ 24 ]. BLP estimates in all regimes were obtained via integration. For mixed and fractal spectra, BLP estimates were log-transformed, while for oscillatory spectra – where positive power is not strictly ensured by IRASA – we computed the difference between log-transformed mixed- and log-transformed fractal-power (i.e., the proportion of oscillatory power). Finally, each BLP index was characterized by their temporal mean (baseline) and standard deviation (fluctuation), taken over the 23 windows. Throughout this manuscript, we will use the following notation: frequency band is indicated by its preceding Greek-letter symbol (e.g., $ \alpha $ ), physiological state (EC) is indicated in superscript, spectrum type ( \(\:mixd\) , \(\:frac\) and \(\:osci\) for raw/mixed, fractal and oscillatory spectra, respectively) is indicated in subscript, and the mean or variance is indicated by encapsulating \(\:\mu\:(\bullet\:)\) and \(\:\sigma\:(\bullet\:)\) , respectively. For example, temporal fluctuation of eyes-closed BLP obtained from oscillatory spectra in the alpha regime is denoted \(\:\alpha\:{BLP}_{osci}^{EC}\) . Note that as part of the IRASA analysis [ 23 ], \(\:1/f\) spectral slope estimates in low- (1–4 Hz, \(\:{\beta\:}_{lo}\) ) and high-frequency (20–45 Hz, \(\:{\beta\:}_{hi}\) ) regimes were also obtained as in [ 17 ]. For dimensionality reduction and simplified physiological interpretability, channel-level estimates were also collapsed over six, established resting-state brain networks [ 25 ]. We followed our previous approaches in this process [ 17 , 26 , 27 ] – building on the work of Giacometti and colleagues [ 28 ] – and assigned EEG channels to the Visual Network (VN), the Somatomotor Network (SM), the Dorsal Attention Network (DA), the Frontoparietal Network (FP), the combined Ventral Attention and Limbic Networks (VAL) and the Default Mode Network (DMN). In addition, we also performed all analyses on the global average indices taken over all 55 channels. The RSN assignment is illustrated on Fig. 1 A. A summary of analysis aspects is provided in Table 1 . Table 1 Summary of analysis aspects and their short description. EEG: electroencephalography; IRASA: Irregular Resampling Auto-Spectral Analysis Aspect Notation Summary Spectral power BLP Band-limited power (BLP) is obtained via Welch's periodogram method, then integrated in the given frequency range and log-transformed to facilitate normality. Frequency range \(\:\delta\:\) , \(\:\theta\:\) , \(\:\alpha\:\) , \(\:\beta\:\) and \(\:\gamma\:\) Canonical frequency bands Delta ( \(\:\delta\:\) ): 1–4 Hz, Theta ( \(\:\theta\:\) ): 4–8 Hz, Alpha ( \(\:\alpha\:\) ): 8–13 Hz, Beta ( \(\:\beta\:\) ): 13–25 Hz and Gamma ( \(\:\gamma\:\) ): 25–45 Hz. Spectrum type \(\:mixd\) , \(\:frac\) and \(\:osci\) Power spectra separated by the IRASA method [ 23 ]. The raw ( \(\:mixd\) ) power spectrum is assumed to be the sum of \(\:1/{f}^{\beta\:}\) fractal ( \(\:frac\) ) and narrow-band oscillatory ( \(\:osci\) ) components, with the latter obtained by subtracting the fractal from the raw spectrum. Statistical measure \(\:\mu\:(\bullet\:)\:and\:\sigma\:(\bullet\:)\) EEG is analyzed in a time-resolved, sliding window fashion. Note that continuous 30-second EEG segments are standardized to ensure identical distribution of total power among participants, but individual windows (epochs) are not to allow for BLP dynamics to emerge. Then, brain activity is characterized by the baseline i.e., mean ( \(\:\mu\:(\bullet\:)\) ) and extent of temporal fluctuation i.e., standard deviation ( \(\:\sigma\:(\bullet\:)\) ) of BLP estimates. Topology VN, SM, DA, VAL FP, DMN and global EEG channels are collapsed onto brain resting-state networks: the Visual Network (VN), the Somatomotor Network (SM), the Dorsal Attention Network (DA), the Frontoparietal Network (FP), the combined Ventral Attention and Limbic Networks (VAL) and the Default Mode Network (DMN). In addition, global average indices taken over all 55 channels are also assessed. See Fig. 1 A for channel assignment. Statistical Evaluation Spectral features were contrasted between the HC and SZ groups on global-to-global and RSN-to-RSN levels. Channel-to-channel comparisons were performed only for exploratory and visualization purposes according to similar principles. In all cases, data normality was first probed with Lilliefors test, then either two-sample t-test or Mann-Whitney U test was employed accordingly. Outcomes were adjusted for multiple comparisons using False Discovery Rate (FDR) [ 29 ]. Effect sizes were characterized via Cohen's d for normally distributed data, while according to \(\:r=\left|z\right|/\sqrt{N}\) otherwise, where \(\:z\) and \(\:N\) are the standardized test statistic and the total sample size, respectively [ 30 ]. Those global and RSN-level spectral features that revealed significant between-group difference – and therefore could be considered characteristic traits of SZ – were further subjected to correlation analysis with general (GEN-PANSS), negative (NEG-PANSS), positive (POS-PANSS) and total (SUM-PANSS) symptom scores. Possible confounding variables age, sex, years in education, disease duration and CPZ equivalent dose was estimated in both spectral indices and PANSS scores via multiple linear regression. Then, normality of the residuals was assessed via Lilliefors test, and the relationship between spectral indices and clinical symptoms were characterized via Pearson or Spearman correlation coefficient accordingly. Outcomes were not adjusted for multiple comparisons and therefore these outcomes should be treated as exploratory. Effect sizes were interpreted according to Cohen's guidelines, with Pearson or Spearman \(\:\left|r\right|=0.1\) , \(\:0.3\) and \(\:0.5\) indicating small, moderate or large effect [ 31 ]. Results Baseline Distribution of Power Grand average power spectra illustrating baseline power are presented in the upper panels of Fig. 2 . We found more power distributed to lower frequencies in SZ compared to HC, which could be attributed to the fractal spectral component. Specifically, both \(\:\mu\:\left(\delta\:{BLP}_{frac}^{EC}\right)\) ( Supplementary Fig. 1 ) and \(\:\mu\:\left(\theta\:{BLP}_{frac}^{EC}\right)\) ( Supplementary Fig. 2 ) was increased in SZ in all RSNs – also including the global average – except for the SM network. Additionally, increased \(\:\mu\:\left(\alpha\:{BLP}_{frac}^{EC}\right)\) was found in SZ compared to HC most prominently over the VN, while \(\:\mu\:\left(\alpha\:{BLP}_{osci}^{EC}\right)\) exhibited a tendency towards being higher in HC, although this never reached the level of statistical significance after adjusting for multiple comparisons ( Supplementary Fig. 3 ). Effect sizes were ranging between 0.2854 and 0.5680, indicating moderate effect. Spectral Power Dynamics Temporal variance in grand average power spectra is presented in the lower panels of Fig. 2 . SZ patients expressed reduced temporal variability of alpha and beta power. Specifically, there was a ubiquitous reduction of \(\:\sigma\:\left(\alpha\:{BLP}_{mixd}^{EC}\right)\) ( Supplementary Fig. 4 ) and \(\:\sigma\:\left(\beta\:{BLP}_{mixd}^{EC}\right)\) ( Supplementary Fig. 5 ) in SZ compared to HC. Effect sizes ranged between 0.2632 and 0.4313, indicating moderate effect except for 2 cases. While this decrease in the alpha band could be attributed to oscillatory power ( \(\:\sigma\:\left(\alpha\:{BLP}_{osci}^{EC}\right)\) , effect size: \(\:\in\:[0.3519;\:1.0503]\) ), it was contingent with fractal power in the beta band ( \(\:\sigma\:\left(\beta\:{BLP}_{frac}^{EC}\right)\) , effect size: \(\:\in\:[0.2706;\:0.7643]\) ) and more constrained to posterior cortical networks VN, DA and SM networks. Notably, the strongest effect was seen for global \(\:\sigma\:\left(\alpha\:{BLP}_{osci}^{EC}\right)\) ( p = 0.0001, \(\:{t}_{59}\) =4.1540, Cohen's d = 1.0503) followed by that over the FP network ( p = 0.0006, \(\:{t}_{59}\) =3.6512, Cohen's d = 0.9232), while for \(\:\sigma\:\left(\beta\:{BLP}_{frac}^{EC}\right)\) over VN ( p = 0.0.0037, \(\:{t}_{59}\) =3.0231, Cohen's d = 0.7643) and DA ( p = 0.0056, \(\:{t}_{59}\) =2.8770, Cohen's d = 0.7274). Spectral Features and Clinical Symptoms Correlation analysis revealed significant associations between beta-band dynamics over the DA network and clinical symptoms Fig. 3 . Precisely, \(\:\sigma\:\left(\beta\:{BLP}_{mixd}^{EC}\right)\) over DA was inversely correlated to PANSS-NEG (Spearman r =-0.4994, p = 0.0055) and to PANSS-SUM (Spearman r =-0.3967, p = 0.0308), indicating that higher PANSS scores were associated with greater reduction in temporal variability in beta BLP. Note that these outcomes were not adjusted for multiple comparisons. Discussion Our findings demonstrate that a time-resolved approach to BLP analysis in SZ is warranted, with separating aperiodic scale-free and oscillatory EEG spectral components. We found increased baseline spectral power in SZ in the delta and theta bands in line with previous literature [ 5 , 7 – 9 , 32 ] while decreased variability in the alpha and beta bands. Decomposition of the spectrum indicated that the slow-wave baseline shift ( Supplementary Figs. 1&2 ) and decrease in beta fluctuations ( Supplementary Fig. 5 ) was predominantly attributed to broadband fractal activity, while the decrease in alpha fluctuations were more likely of oscillatory origin ( Supplementary Fig. 4 ). Baseline distinguishing EEG features were uncorrelated with disease symptoms, but beta-band dynamics showed associations with clinical presentations (see Fig. 3 ). These results point towards a general slowing of brain activity resulting in the allocation of more power towards smaller frequencies, with at the same time alpha oscillations and beta activity becoming less variable over time. Relationship with E/I ratio In our previous work [ 17 ] we exclusively focused on the fractal component of the power spectra to characterize spectral slopes in 1–4 ( \(\:{\beta\:}_{lo}\) ) and 20–45 Hz ( \(\:{\beta\:}_{hi}\) ) regimes. In a similar pattern, that analysis indicated no between-group difference in the temporal average of slopes, however we have found a reduction in the temporal variance of \(\:{\beta\:}_{hi}\) over the DA network. To address why these two patterns are highly similar, we performed linear mixed-effects (LME) modeling where we regressed the temporal evolution (i.e., over the 23 sliding windows) of global \(\:{\beta\:}_{hi}\) on that of global oscillatory \(\:\alpha\:BLP\) similarly to [ 33 ] with participant as a random effect. This revealed a very strong association between \(\:\sigma\:\left(\alpha\:{BLP}_{osci}^{EC}\right)\) and \(\:{\beta\:}_{hi}\) in both HC ( r = 0.6657, p < \(\:{10}^{-5}\) ) and SZ ( r =0.5059, p < \(\:{10}^{-5}\) ), raising the concern that our previous outcomes were confounded (or driven by) oscillatory alpha instead of broadband activity. We propose two, not necessarily exclusive explanations for this strong dependence. First, our previous analytical approach utilizing IRASA was not delicate enough to compensate for potential bias of oscillatory alpha activity on the 20–45 Hz range when estimating \(\:{\beta\:}_{hi}\) . Even though we defined the specifics of our analytical approach with this concern in mind – i.e., excluding 4–20 Hz regime from slope estimation, setting the rescaling parameter \(\:h\) ranging from 1.1 to 2.6 to avoid leakage and smearing effects [ 23 , 27 ]–, unfortunately we cannot completely exclude this possibility. On the other hand, alpha activity is broadly considered as a neural signature of inhibitory tone [ 34 , 35 ]. While also observed in [ 33 ], recent studies indicated that dynamics in alpha power are related to E/I ratio: namely, alpha bursts with high amplitude might represent short periods with stronger inhibition, alternating with excitation/disinhibition [ 36 , 37 ]. These phenomena are well in line with our findings and could provide an explanation on why we see an alignment between \(\:\alpha\:BLP\) and \(\:{\beta\:}_{hi}\) . Nevertheless, this problem appears very relevant for studies inferring the E/I ratio from EEG analysis and warrants future research. Reduced Power Fluctuations and Loss of Nonlinearity Systems that exhibit power-law ( \(\:1/{f}^{\beta\:}\) ) scaling as a local instead of global property are referred to as multifractals [ 38 , 39 ]. In case of temporal processes this means that the fractal scaling exponent \(\:\beta\:\) changes over time instead of being constant [ 40 – 43 ]. Such phenomena often emerge in biological systems as a result of nonlinear, antagonistic feedback loops such as the net effect of the sympathetic and parasympathetic nervous system [ 44 , 45 ], or the interaction of excitatory and inhibitory neuronal populations [ 46 ]. The temporal variability of \(\:{\beta\:}_{hi}\) therefore implies that the neural processes generating the EEG signal might be multifractal; however, such direct confirmation of multifractality usually demands immense computational capacity [ 47 , 48 ]. Instead, we decided to test if nonlinearity might explain the emergence in the variability of alpha BLP, which we used as a proxy for \(\:{\beta\:}_{hi}\) given their strong correlation. For this purpose, we employed surrogate data testing via phase randomization [ 49 ] to eliminate non-linearity (see Supplementary Material for details). The outcomes of this analysis are presented on Fig. 4 . Locations where nonlinearity was confirmed for higher proportions (i.e., > 15 subjects) are mostly located over fronto-central and parieto-occipital regions. On the group level, \(\:\sigma\:\left(\alpha\:{BLP}_{mixd}^{EC}\right)\) is well above values from surrogate data in the HC group, while the obtained values are mostly within the spread of surrogate values in the SZ group. This analysis therefore strongly indicates that healthy alpha dynamics are governed by nonlinear dynamic processes, which appear mostly absent in the patient cohort. We also made an attempt to reproduce these results regarding \(\:\sigma\:\left(\alpha\:{BLP}_{mixd}^{EC}\right)\) on an independent, open EEG dataset [ 50 ] (see Supplementary Material for details). This analysis did not indicate a significant difference between HC and SZ or suggested a loss of nonlinearity in the SZ group (see Supplementary Fig. 6 ). Specifically, \(\:\sigma\:\left(\alpha\:{BLP}_{mixd}^{EC}\right)\) in SZ was comparable to that in HC (HC: 0.2639 ± 0.1766, SZ: 0.2692 ± 0.0699, p = 0.9185, \(\:{t}_{26}\) =-0.1033, Cohen's d = 0.0379). However, when comparing data from each group in the database of [ 50 ] to their respective match in our dataset, \(\:\sigma\:\left(\alpha\:{BLP}_{mixd}^{EC}\right)\) was significantly lower in our patient sample (median: 0.1350, IQR: [0.1028; 0.2072], p = 0.0002, z = 3.7670, r = 0.5679) compared to those of Olejarczyk & Jernajczyk, while we did not find a statistical difference between HC groups (median: 0.2557, IQR: [0.1472; 0.4927], p = 0.5981, z =-0.5271, r = 0.0786). There are a couple considerations that might explain these outcomes. First, the sample in [ 50 ] contained only paranoid SZ patients in contrast to our cohort, in which only 7 out of 30 patients were diagnosed as paranoid-type SZ. More importantly, patients enrolled in [ 50 ] were subjected to a medication washout period of at least seven days before EEG measurement, while our patients were on medication at the time of assessment. This raises the concern that the observed EEG patterns rather reflect response to medication than neural characteristics of the disease, however a recent review addressing pharmacotherapy effects on EEG in SZ does not indicate any previous works reporting changes in short-term alpha- or beta-band BLP dynamics in response to antipsychotic medication [ 51 ] and the potential effects of antipsychotics on nonlinear characteristics of EEG are also poorly understood, although some reports might indicate a reduction in non-linearity in response to antipsychotic medication [ 52 , 53 ]. Potential Effects of Medication To further explore the potential confounding effects of antipsychotic medication, we correlated global EEG indices with CPZ equivalent dose values (see Supplementary Material ). This analysis revealed a positive correlation between CPZ and \(\:\sigma\:\left(\beta\:{BLP}_{mixd}^{EC}\right)\) $ (Spearman r = 0.4399, p = 0.0150, unadjusted). While this indicates the influence of medication on EEG dynamics, it is also important to highlight that medication was heterogenous in our sample (see Table 1 in [ 17 ] for details) with vastly different mechanisms of action, complicating potential influences on EEG markers [ 54 ]. Pharmaceutical therapy is a multifaceted endeavor in SZ, as often more severe symptoms (higher PANSS scores) warrant more potent pharmaceutical intervention, while a more aggressive treatment plan can result in a more successful repression of symptoms (lower PANSS scores). In our sample, correlation analysis between CPZ doses and PANSS scores ( Supplementary Fig. 7 ) indicated a significant positive correlation between CPZ and PANSS-NEG (Pearson r = 0.3786, p = 0.0391) but no relationship with positive symptoms that can be more efficiently managed by antidopaminergic medication [ 55 ]. However, since the relationship between \(\:\sigma\:\left(\beta\:{BLP}_{mixd}^{EC}\right)\) and PANSS-NEG was anticorrelated after controlling for CPZ in both variables, these outcomes suggest that it is unlikely that EEG features and PANSS scores were both vastly confounded by medication and therefore it is probable that these EEG features could indeed reflect pathophysiology and/or therapy response. Limitations and Future Perspectives The reported EEG patterns in SZ were associated with clinical appearance, and therefore they could potentially be utilized in predicting therapy response and selecting the intervention strategy accordingly, which is still a major challenge for SZ [ 56 ]. This calls for prospective, longitudinal studies with only first-episode, drug-naïve SZ participants, so it can be appropriately assessed if the observed patterns are a response to antipsychotic medication, they reflect ongoing pathology, or both. Longitudinal follow-up will also be impediment to evaluating if dynamic spectral features carry any predictive potential in therapeutic strategy planning. Some methodological concerns also need attention. In terms of separating broadband \(\:1/f\) from oscillatory activity in the EEG spectra, following a personalized approach could render our outcomes more relevant on the level of individual patients and is likely to offer a more appropriate characterization of spectral dynamics [ 57 ] by taking into consideration the inter-individual variability of peak alpha frequency [ 58 ] or scaling range-dependent spectral bimodality [ 59 ]. Besides these issues, we also recognize limitations such as low and heterogenous patient sample, lack of explicit measures on other disease features such as cognitive dysfunction [ 60 ], and our sole focus on EEG signals over e.g., genetic factors [ 61 ]. Nonetheless, we believe that this study is relevant for drawing attention to the relevance of a dynamic EEG analysis approach in SZ and can lay the foundation for future work overcoming these drawbacks. Summary With regards to the relationships between dynamic EEG markers, symptom scores and medication data, we observed three patterns. First, a negative correlation between EEG markers and PANSS scores. This falls in line with the observed between-group differences (lower variability in SZ compared to HC), in that a more extensive loss of temporal BLP fluctuations was associated with more severe clinical symptoms in SZ. Second, there was a positive correlation between the CPZ equivalent dose and dynamic EEG indices, implying that medication in fact increased and not decreased temporal variability in BLP. Therefore, it is unlikely that between-group differences were primarily driven by pharmaceutical effects, even if we failed to replicate this pattern on an independent sample. Third, CPZ was positively correlated with both \(\:\sigma\:\left(\beta\:{BLP}_{mixd}^{EC}\right)\) and PANSS-NEG, and therefore it should not explain the negative correlation between the latter two after controlling for its effect. These outcomes suggest that dynamic EEG indices can indeed be related to disease pathomechanism and/or therapy response in SZ, likely indicating a loss of nonlinearity. Data Availability Statement The eyes-closed resting-state data analyzed this study has been made available in a public Zenodo repository titled “Resting-state EEG, clinical, and demographics data from schizophrenia patients and age-matched healthy controls” at https://doi.org/10.5281/zenodo.14808295 . All data analyses were carried out via Matlab (version 2023b) utilizing the EEGLAB toolbox [ 20 ] (version 2020.0) along with custom analysis codes. Matlab codes developed and used for analyzing the data and visualizing outcomes (along with pre-processed EEG) are provided in a public GitHub repository at https://github.com/samuelracz/schizophrenia_BLP_dynamics . Please note that details of all statistical evaluations, including raw and adjusted p -values, test statistic values and effect size measures will be accessible via the shared code. Declarations Conflict of Interest Statement None of the authors has any financial or other conflicting interests related to the work carried out in this study. Author Contributions F.S.R. designed the analytical pipeline, performed data screening, pre-processing and analysis, results interpretation, and wrote the initial draft of the manuscript. K.F., H.M., and Z.F. contributed to electrophysiological and clinical data collection and study conceptualization. M.B. contributed to data collection, screening and pre-processing. P.M. contributed to developing the analytical pipeline and interpreting results. G.C. conceptualized the study and provided oversight and guidance over data collection, analysis and all other study aspects. All authors contributed to manuscript development and revisions, and all authors have read and approved the final version. Acknowledgments This research was supported by the Hungarian Research Foundation Grant (OTKA FK 138385) and Hungarian Research Foundation Grant (OTKA PD 146424), as well as the Ministry of Innovation and Technology of Hungary from the National Research, Development and Innovation Fund, financed under the TKP2021-EGA-25 funding scheme. References Abi-Dargham, A., et al., Candidate biomarkers in psychiatric disorders: state of the field . World Psychiatry, 2023. 22(2): p. 236–262. Rubio, J.M., et al., Long-term Continuity of Antipsychotic Treatment for Schizophrenia: A Nationwide Study . Schizophr Bull, 2021. 47(6): p. 1611–1620. Kahn, R.S., et al., Schizophrenia . Nat Rev Dis Primers, 2015. 1: p. 15067. Meehan, A.J., et al., Clinical prediction models in psychiatry: a systematic review of two decades of progress and challenges . Mol Psychiatry, 2022. 27(6): p. 2700–2708. Kraguljac, N.V., et al., Neuroimaging Biomarkers in Schizophrenia . Am J Psychiatry, 2021. 178(6): p. 509–521. Boutros, N.N., et al., The status of spectral EEG abnormality as a diagnostic test for schizophrenia . Schizophrenia Research, 2008. 99(1–3): p. 225–237. Newson, J.J. and T.C. Thiagarajan, EEG Frequency Bands in Psychiatric Disorders: A Review of Resting State Studies . Frontiers in Human Neuroscience, 2019. 12. Perrottelli, A., et al., EEG-Based Measures in At-Risk Mental State and Early Stages of Schizophrenia: A Systematic Review . Front Psychiatry, 2021. 12: p. 653642. Lee, H.S. and J.S. Kim, Implication of Electrophysiological Biomarkers in Psychosis: Focusing on Diagnosis and Treatment Response . J Pers Med, 2022. 12(1). Damaraju, E., et al., Dynamic functional connectivity analysis reveals transient states of dysconnectivity in schizophrenia . Neuroimage: Clinical, 2014. 5: p. 298–308. Cattarinussi, G., et al., Dynamic functional connectivity in schizophrenia and bipolar disorder: A review of the evidence and associations with psychopathological features . Prog Neuropsychopharmacol Biol Psychiatry, 2023. 127: p. 110827. Gao, R., E.J. Peterson, and B. Voytek, Inferring synaptic excitation/inhibition balance from field potentials . Neuroimage, 2017. 158: p. 70–78. Kehrer, C., et al., Altered Excitatory-Inhibitory Balance in the NMDA-Hypofunction Model of Schizophrenia . Front Mol Neurosci, 2008. 1: p. 6. Uliana, D.L., et al., The excitatory-inhibitory balance as a target for the development of novel drugs to treat schizophrenia . Biochem Pharmacol, 2024. 228: p. 116298. Lányi, O., et al., Excitation/inhibition imbalance in schizophrenia: a meta-analysis of inhibitory and excitatory TMS-EMG paradigms . Schizophrenia (Heidelb), 2024. 10(1): p. 56. Racz, F.S., et al., Multifractal and Entropy-Based Analysis of Delta Band Neural Activity Reveals Altered Functional Connectivity Dynamics in Schizophrenia . Frontiers in Systems Neuroscience, 2020. 14. Racz, F.S., et al., Reduced temporal variability of cortical excitation/inhibition ratio in schizophrenia . Schizophrenia (Heidelb), 2025. 11(1): p. 20. Kay, S.R., A. Fiszbein, and L.A. Opler, The positive and negative syndrome scale (PANSS) for schizophrenia . Schizophr Bull, 1987. 13(2): p. 261–76. Xia, M.R., J.H. Wang, and Y. He, BrainNet Viewer: A Network Visualization Tool for Human Brain Connectomics . Plos One, 2013. 8(7). Delorme, A. and S. Makeig, EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis . Journal of Neuroscience Methods, 2004. 134(1): p. 9–21. Winkler, I., S. Haufe, and M. Tangermann, Automatic Classification of Artifactual ICA-Components for Artifact Removal in EEG Signals . Behavioral and Brain Functions, 2011. 7. Winkler, I., et al., Robust artifactual independent component classification for BCI practitioners . Journal of Neural Engineering, 2014. 11(3). Wen, H.G. and Z.M. Liu, Separating Fractal and Oscillatory Components in the Power Spectrum of Neurophysiological Signal . Brain Topography, 2016. 29(1): p. 13–26. Racz, F.S., et al., Multiple-Resampling Cross-Spectral Analysis: An Unbiased Tool for Estimating Fractal Connectivity With an Application to Neurophysiological Signals . Frontiers in Physiology, 2022. 13. Yeo, B.T.T., et al., The organization of the human cerebral cortex estimated by intrinsic functional connectivity . Journal of Neurophysiology, 2011. 106(3): p. 1125–1165. Racz, F.S., et al., Multifractal and entropy analysis of resting-state electroencephalography reveals spatial organization in local dynamic functional connectivity . Scientific Reports, 2019. 9. Racz, F.S., et al., Separating scale-free and oscillatory components of neural activity in schizophrenia . Brain Behav, 2021. 11(5): p. e02047. Giacometti, P., K.L. Perdue, and S.G. Diamond, Algorithm to find high density EEG scalp coordinates and analysis of their correspondence to structural and functional regions of the brain . Journal of Neuroscience Methods, 2014. 229: p. 84–96. Benjamini, Y. and Y. Hochberg, Controlling the False Discovery Rate - a Practical and Powerful Approach to Multiple Testing . Journal of the Royal Statistical Society Series B-Statistical Methodology, 1995. 57(1): p. 289–300. Rosenthal, R., H. Cooper, and L. Hedges, Parametric measures of effect size . The handbook of research synthesis, 1994. 621(2): p. 231–244. Brydges, C.R., Effect Size Guidelines, Sample Size Calculations, and Statistical Power in Gerontology . Innov Aging, 2019. 3(4): p. igz036. Owens, E.M., et al., Electrophysiological Endophenotypes for Schizophrenia . Harv Rev Psychiatry, 2016. 24(2): p. 129–47. Becker, R., D. Van de Ville, and A. Kleinschmidt, Alpha Oscillations Reduce Temporal Long-Range Dependence in Spontaneous Human Brain Activity . Journal of Neuroscience, 2018. 38(3): p. 755–764. Klimesch, W., P. Sauseng, and S. Hanslmayr, EEG alpha oscillations: The inhibition-timing hypothesis . Brain Research Reviews, 2007. 53(1): p. 63–88. Jensen, O. and A. Mazaheri, Shaping functional architecture by oscillatory alpha activity: gating by inhibition . Front Hum Neurosci, 2010. 4: p. 186. Lombardi, F., et al., Beyond pulsed inhibition: Alpha oscillations modulate attenuation and amplification of neural activity in the awake resting state . Cell Rep, 2023. 42(10): p. 113162. Sano, M., et al., Analysis of the alpha activity envelope in electroencephalography in relation to the ratio of excitatory to inhibitory neural activity . PLoS One, 2024. 19(6): p. e0305082. Tel, T., Fractals, Multifractals, and Thermodynamics - an Introductory Review. Zeitschrift Fur Naturforschung Section a-a Journal of Physical Sciences, 1988. 43(12): p. 1154–1174. Theiler, J., Estimating Fractal Dimension. Journal of the Optical Society of America a-Optics Image Science and Vision, 1990. 7(6): p. 1055–1073. Ivanov, P.C., et al., Multifractality in human heartbeat dynamics . Nature, 1999. 399(6735): p. 461–5. Kantelhardt, J.W., et al., Multifractal detrended fluctuation analysis of nonstationary time series . Physica A: Statistical Mechanics and Its Applications, 2002. 316(1–4): p. 87–114. Racz, F.S., et al., Multifractal dynamics of resting-state functional connectivity in the prefrontal cortex . Physiological Measurement, 2018. 39(2): p. 024003. Racz, F.S., et al., Multifractal Dynamic Functional Connectivity in the Resting-State Brain . Front Physiol, 2018. 9: p. 1704. Ivanov, P., et al., Stochastic feedback and the regulation of biological rhythms . Europhys Lett, 1998. 43(4): p. 363–8. Ashkenazy, Y., et al., A stochastic model of human gait dynamics . Physica A: Statistical Mechanics and its Applications, 2002. 316(1–4): p. 662–670. Poil, S.S., et al., Critical-State Dynamics of Avalanches and Oscillations Jointly Emerge from Balanced Excitation/Inhibition in Neuronal Networks . Journal of Neuroscience, 2012. 32(29): p. 9817–9823. Chhabra, A.B., et al., Direct determination of the f(alpha) singularity spectrum and its application to fully developed turbulence . Phys Rev A Gen Phys, 1989. 40(9): p. 5284–5294. Kaposzta, Z., et al., Real-Time Algorithm for Detrended Cross-Correlation Analysis of Long-Range Coupled Processes . Frontiers in Physiology, 2022. 13. Theiler, J., et al., Testing for Nonlinearity in Time-Series - the Method of Surrogate Data . Physica D, 1992. 58(1–4): p. 77–94. Olejarczyk, E. and W. Jernajczyk, Graph-based analysis of brain connectivity in schizophrenia . PLoS One, 2017. 12(11): p. e0188629. De Pieri, M., et al., Pharmaco-EEG of antipsychotic treatment response: a systematic review . Schizophrenia (Heidelb), 2023. 9(1): p. 85. Kang, U.G., et al., Non-linear dynamic analysis of clozapine-induced electroencephalographic changes in schizophrenic patients–a preliminary study . Prog Neuropsychopharmacol Biol Psychiatry, 2001. 25(6): p. 1229–39. Takahashi, T., et al., Antipsychotics reverse abnormal EEG complexity in drug-naive schizophrenia: a multiscale entropy analysis . Neuroimage, 2010. 51(1): p. 173–82. Roth, B.L., D.J. Sheffler, and W.K. Kroeze, Magic shotguns versus magic bullets: selectively non-selective drugs for mood disorders and schizophrenia . Nat Rev Drug Discov, 2004. 3(4): p. 353–9. Conley, R.R. and D.L. Kelly, Management of treatment resistance in schizophrenia . Biol Psychiatry, 2001. 50(11): p. 898–911. Carbon, M. and C.U. Correll, Clinical predictors of therapeutic response to antipsychotics in schizophrenia . Dialogues Clin Neurosci, 2014. 16(4): p. 505–24. Donoghue, T., N. Schaworonkow, and B. Voytek, Methodological considerations for studying neural oscillations . Eur J Neurosci, 2022. 55(11–12): p. 3502–3527. Haegens, S., et al., Inter- and intra-individual variability in alpha peak frequency . Neuroimage, 2014. 92(100): p. 46–55. Racz, F.S., et al., Multi-Modal Spectral Parametrization Method (MMSPM) for analyzing EEG activity with distinct scaling regimes. arXiv preprint arXiv:2505.18117, 2025. Keefe, R.S. and P.D. Harvey, Cognitive impairment in schizophrenia . Handb Exp Pharmacol, 2012(213): p. 11–37. Guillozet-Bongaarts, A.L., et al., Altered gene expression in the dorsolateral prefrontal cortex of individuals with schizophrenia . Mol Psychiatry, 2014. 19(4): p. 478–85. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files supplementarymaterialSchizophreniaBLPdynamics.pdf Supplementary Material for Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: revise 22 Oct, 2025 Review # 3 received at journal 11 Oct, 2025 Reviewer # 3 agreed at journal 09 Oct, 2025 Review # 1 received at journal 08 Aug, 2025 Reviewer # 2 agreed at journal 07 Aug, 2025 Reviewer # 1 agreed at journal 21 Jul, 2025 Reviewers invited by journal 27 Jun, 2025 Editor assigned by journal 24 Jun, 2025 Submission checks completed at journal 24 Jun, 2025 First submitted to journal 23 Jun, 2025 Unknown event 17 Jun, 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-6908048","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":477456162,"identity":"a44e4245-26e4-470f-a332-c8a14c958b35","order_by":0,"name":"Gabor Csukly","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABE0lEQVRIie3QsUrEMBjA8S8UbircmlvOJxA+6aAHpfcgLimFmwwIgmOpHOQW5dbCPYbQuRK4W4pzoYLnGxRcTgho0oqDNIKbYP60FD7yI0kBXK6/mQd786H6bftBBoDdYyfsk5C8G5BfEM//ImAn43UVYCwgWm+W29dQhefHqxtB28sUTikbJLS+CJgmSf60TTZcLHhRPYhJjhJm+TBB6gelIaCtxzPJizoWjY8lYG0h46onR4acqXdePL+IRmFqJ9AfLEJDYFTqXYhoAD0rofXiCtkjZSf6LuROJPou8ertFqWP1d7yx+Q9ba/D+bRZSjioiBc7KfGg0inuhnfpIiMaZ9+Hvn29ScH85wUul8v1r/sAik5lLKOLOVYAAAAASUVORK5CYII=","orcid":"","institution":"Semmelweis University","correspondingAuthor":true,"prefix":"","firstName":"Gabor","middleName":"","lastName":"Csukly","suffix":""},{"id":477456163,"identity":"7764f82b-72ef-4d76-bb14-b3a4c9a74863","order_by":1,"name":"Frigyes Racz","email":"","orcid":"https://orcid.org/0000-0001-9077-498X","institution":"University of Texas at Austin","correspondingAuthor":false,"prefix":"","firstName":"Frigyes","middleName":"","lastName":"Racz","suffix":""},{"id":477456164,"identity":"148d9459-2cea-40ac-8531-e4bc7cadee92","order_by":2,"name":"Kinga Farkas","email":"","orcid":"https://orcid.org/0000-0002-1125-3957","institution":"Semmelweis University","correspondingAuthor":false,"prefix":"","firstName":"Kinga","middleName":"","lastName":"Farkas","suffix":""},{"id":477456165,"identity":"89d9b76c-b500-4059-8c1e-69e58889062e","order_by":3,"name":"Melinda Becske","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Melinda","middleName":"","lastName":"Becske","suffix":""},{"id":477456166,"identity":"53d0bd94-93fd-4685-8b53-7bc32872e1c8","order_by":4,"name":"Hajnalka Molnar","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Hajnalka","middleName":"","lastName":"Molnar","suffix":""},{"id":477456167,"identity":"f8cdc525-d515-407d-bd1f-5d94f99b0ea4","order_by":5,"name":"Zsuzsanna Fodor","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Zsuzsanna","middleName":"","lastName":"Fodor","suffix":""},{"id":477456168,"identity":"e3c1000d-701b-4646-a502-ad00874016a7","order_by":6,"name":"Péter Mukli","email":"","orcid":"https://orcid.org/0000-0003-4355-8103","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Péter","middleName":"","lastName":"Mukli","suffix":""}],"badges":[],"createdAt":"2025-06-16 18:40:43","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6908048/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6908048/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85843721,"identity":"f398ade1-20cc-46b7-8e6b-dcfa9e793721","added_by":"auto","created_at":"2025-07-02 09:24:07","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2201391,"visible":true,"origin":"","legend":"\u003cp\u003eElectrode montage, channel groupings and illustration of the IRASA method. Panel A shows the electrode layout, where color indicates the association of each electrode to one of the six resting-state networks. Panel B illustrates the separation of the fractal and oscillatory (inset plots) components from the raw (mixed) power spectrum for healthy (upper) and schizophrenia (lower). In both cases, the group average of global spectra (taken as the average of all channels) obtained from eyes-closed condition is presented for illustration. Shaded areas indicate the standard error of the mean, and dashed lines denote the boundary frequencies between the five investigated frequency bands (delta: 1-4 Hz; theta: 4-8 Hz; alpha: 8-13 Hz; beta: 13-25 Hz; gamma: 25-45 Hz). IRASA: irregular-resampling auto-spectral analysis; VN: visual network; SM: somatomotor network; DA: dorsal attention network; VAL: combined ventral attention and limbic networks; FP: frontoparietal network; DMN: default mode network\u003c/p\u003e","description":"","filename":"fig1layoutmethods.png","url":"https://assets-eu.researchsquare.com/files/rs-6908048/v1/6bf5343e9f2ee489ee6ac0ac.png"},{"id":85843800,"identity":"215e42a0-517f-4fe9-bab3-926528e3bb43","added_by":"auto","created_at":"2025-07-02 09:24:08","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1710679,"visible":true,"origin":"","legend":"\u003cp\u003eGrand average power spectra taken in eyes-closed (EC) resting state for raw/mixed (left), isolated fractal (middle) and oscillatory (right) components. Top and bottom rows present the average and standard deviation of spectral power taken over sliding windows. Asterisk symbols indicate significant between-group difference (\u003cem\u003ep\u003c/em\u003e\u0026lt;0.05) for each 2-Hz frequency bin. BLP: band-limited power; HC: healthy control; SZ: schizophrenia.\u003c/p\u003e","description":"","filename":"fig2GAspectraec.png","url":"https://assets-eu.researchsquare.com/files/rs-6908048/v1/7d083613aea2996b250cab85.png"},{"id":85843818,"identity":"b4db5720-30d2-40be-a384-a466d9cc1f2a","added_by":"auto","created_at":"2025-07-02 09:24:10","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":394892,"visible":true,"origin":"","legend":"\u003cp\u003eCorrelation analysis between \u003cstrong\u003eσ(BLP\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003eEC\u003c/strong\u003e\u003c/sup\u003e\u003csub\u003e\u003cstrong\u003emixd\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e)\u003c/strong\u003e and PANSS scores in the Dorsal Attention (DA) network. The neural index was found significantly correlated (\u003cem\u003ep\u003c/em\u003e\u0026lt;0.05) with negative (left) and total (right) PANSS scores. All variables were adjusted for potential confounding effects of age, sex, years in education and disease duration. PANSS: positive and negative syndrome scale; NEG: negative; SUM: all symptoms combined.\u003c/p\u003e","description":"","filename":"fig3corrpanssrsnall.png","url":"https://assets-eu.researchsquare.com/files/rs-6908048/v1/ff0364dcc66b1341b4215f30.png"},{"id":85843831,"identity":"a2681ec8-b1b0-40c9-9ed9-2f413cdeff2e","added_by":"auto","created_at":"2025-07-02 09:24:12","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":910710,"visible":true,"origin":"","legend":"\u003cp\u003eResults of surrogate data analysis in HC (left) and SZ (right) groups. The top panels illustrate for every cortical location the number of participants in the respective groups where nonlinearity was confirmed. In the HC group nonlinearity could be confirmed for most locations for at least 15 out of 31 participants, while this proportion was mostly 4-6 out of 30 in SZ. Lower panels show the actual group averages for \u003cstrong\u003eσ(BLP\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003eEC\u003c/strong\u003e\u003c/sup\u003e\u003csub\u003e\u003cstrong\u003emixd\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e)\u003c/strong\u003e in red values obtained from surrogate time series indicated in blue. Continuous blue line denotes the mean from surrogates, while dotted line denotes ± standard deviation from the mean.\u003c/p\u003e","description":"","filename":"fig4varalphaBLPsurrogate.png","url":"https://assets-eu.researchsquare.com/files/rs-6908048/v1/3741dd0397c1cb362076eaaf.png"},{"id":85844862,"identity":"2dc70439-2726-4fbb-8129-c3acc5cccf56","added_by":"auto","created_at":"2025-07-02 09:32:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5768043,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6908048/v1/abb3d20f-76e8-4136-989f-641ca896a7db.pdf"},{"id":85843805,"identity":"1756986f-af08-44ed-9769-bdbf4bb15b65","added_by":"auto","created_at":"2025-07-02 09:24:09","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":9343763,"visible":true,"origin":"","legend":"Supplementary Material for Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia","description":"","filename":"supplementarymaterialSchizophreniaBLPdynamics.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6908048/v1/6e0b73ee97cc71b21d992f6b.pdf"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDeveloping and understanding biomarkers of schizophrenia (SZ) is a critical endeavor: such physiology-based features can not only advance understanding the pathomechanisms of this disorder but could also directly affect clinical practice that is yet heavily burdened with costly trial-and-error practices, especially in terms of therapy planning and optimization [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. For instance, the inability to select the most ideal pharmaceutical agent in advance can lead to prolongment of effective therapy, increasing patient distress and impairment, reduction of treatment engagement, and an increased risk of danger to self and others [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. The relevance of this issue is illustrated by the fact that despite having a lifetime prevalence just under 1\\% [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], SZ draws the most research effort in identifying personalized biomarkers second only to depression [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Among others, functional neuroimaging markers intuitively appear as a suitable choice for this purpose [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], providing an expanding knowledge on potential diagnostic features, future pharmaceutical targets or early indicators of disease progression and therapy response [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCurrently, the most consistent resting-state electroencephalography (EEG) alterations identified in SZ include an increase in delta- and theta-band activity, while a decrease in alpha power [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], though contradictory findings are also reported [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. In contrast to resting-state EEG, evoked potentials such as the P50, P300 or mismatch negativity also show potential not only to distinguish between SZ patients and healthy individuals, but even to provide some predictive insight on disease progression [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Furthermore, extensive dynamic functional connectivity literature indicates that short-scale neural dynamics are altered in SZ, often in a way that is undetectable for conventional, time-invariant analytical approaches [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Interestingly, however, despite all this evidence, the temporal dynamics of resting-state EEG power spectral features are rarely considered in SZ.\u003c/p\u003e \u003cp\u003eAltered cortical excitation/inhibition (E/I) ratio \u0026ndash; which can be characterized via the 1/f slope (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e) of electrophysiological signals [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] \u0026ndash; is another pathological feature that is considered central in SZ [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] with explicit potential for clinical utility [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Since there are indications that the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1/f\\)\u003c/span\u003e\u003c/span\u003e slope might fluctuate over time in SZ [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e], in our recent study [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] we performed a time-resolved analysis of EEG spectral slopes along these considerations. This analysis revealed that while the overall baseline (average over time) of the E/I ratio is comparable in patients with SZ and healthy controls (HC), the patient group showed significantly diminished temporal variability in \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e in the 20\u0026ndash;45 Hz regime. These changes were most consistent over cortical locations corresponding to the dorsal attention network, and they were also correlated to clinical symptom scores. In that work, however, we focused exclusively on the spectral slope and yet omitted conventional EEG measures such as spectral power in canonical frequency regimes [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Therefore, our goal here was to provide a comprehensive picture and understanding of dynamic EEG alterations in SZ by conducting a time-resolved analysis of EEG spectra decomposed into aperiodic \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1/f\\)\u003c/span\u003e\u003c/span\u003e and oscillatory components, hypothesizing diminished temporal variability in fractal spectral power in higher-frequency regimes that (iii) correlate with clinical symptom scores as captured via the Positive and Negative Syndrome Scale (PANSS) [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants and Clinical Measures\u003c/h2\u003e \u003cp\u003eWe analyzed EEG recordings of the same study cohorts as in our recent study; therefore, here we only provide a brief description of the study populations and EEG data collection protocols with more details, please see [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThe SZ group was comprised of 30 SZ patients (11 female, age: 33.07\u0026thinsp;\u0026plusmn;\u0026thinsp;9.73 years), with an age- and sex-matched healthy control (HC) group of 31 individuals (13 female, age: 33.06\u0026thinsp;\u0026plusmn;\u0026thinsp;10.31 years). There was no statistically significant difference between the two groups in terms of age, proportion of sexes, or years in education. The Semmelweis Regional and Institutional Committee of Science and Research Ethics reviewed and approved the study (approval number: 197/2015), which was conducted in line with the Declaration of Helsinki. All participants were informed about study details and provided prior written consent for participation. Further study cohort details and inclusion/exclusion criteria are provided in the \u003cb\u003eSupplementary Material\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEEG Recording, Quality Control and Pre-processing\u003c/h3\u003e\n\u003cp\u003eEEG was collected from 64 standard 10\u0026ndash;10 locations at 1000 Hz resolution using a Neuroscan amplifier. Channels were referenced to linked mastoids and electrode impedances were kept under \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:5k{\\Omega\\:}\\)\u003c/span\u003e\u003c/span\u003e. Resting state data was collected for two minutes in eyes closed (EC) resting-state condition. Data was visually inspected by two investigators independently, blinded to study group association; segments free of overt artifacts were selected and only data segments deemed admissible by both investigators were included for further analysis. Channels AF7, AF8, AFz, F1, F2 TP7 and TP8 (along with M1/M2) were excluded from further analysis as they were identified as noisy/bad in more than 10 participants according to standard criteria [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], resulting in a final channel count of 55. The electrode montage (generated using the BrainNet Viewer tool [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]) is illustrated on Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA. Finally, 30-second segments were selected randomly from the admitted data for each participant for further analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eData was processed and analyzed in Matlab (MathWorks, Natick, MA) using the EELAB toolbox [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] along with custom functions and scripts. Data was first band-pass filtered between 0.1 and 128 Hz using the default finite impulse response filter of EEGLAB with additional notch filters at 50 Hz and 100 Hz, and then downsampled to 256 Hz for computational efficiency. EEG signals were then decomposed via independent component analysis (ICA) and artifacts corresponding to non-neural sources (such as potential eye movements, muscle activity, heart activity, head movement) were identified and removed in an automatized manner using the Multiple Artifact Rejection Algorithm (MARA) toolbox [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] before reverse-ICA transformation. This resulted in the removal of 27.44\u0026thinsp;\u0026plusmn;\u0026thinsp;8.22 out of 55 components, respectively. There was no statistical difference (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.35) in the number of removed components between HC and SZ (see \u003cb\u003eonline Supplementary Material\u003c/b\u003e) Finally, after re-referencing to the common average, EEG signals were standardized (z-scored) according to \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:z\\left(t\\right)=\\left(x\\right(t)-\\mu\\:(x\\left)\\right)/\\sigma\\:\\left(x\\right)\\)\u003c/span\u003e\u003c/span\u003e, where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:x\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e is the original time series with \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:z\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e its z-scored version, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:\\left(x\\right)\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(x\\right)\\)\u003c/span\u003e\u003c/span\u003e are the mean and standard deviation of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:x\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e, respectively. As z-scoring renders all time series zero-centered and with a standard deviation of 1, this step was introduced to help avoid potential confounding effects on power estimates by varying impedances across channels and subjects.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cp\u003e We employed a sliding window approach to obtain time-resolved band-limited power (BLP) estimates after separating \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:1/f\\) \u003c/span\u003e \u003c/span\u003e aperiodic and oscillatory components of the EEG spectra. In that, we utilized a window size of 8 seconds and a step size of 1 second (87.5% overlap), resulting in 23 estimates from the 30-second epochs. Note that with z-scoring performed on the complete 30-second segments, these window estimates reflect distribution of power along the complete broadband spectrum, while at the same time capturing relevant dynamics over the 30-second period (in contrast to z-scoring in every window individually). Similar to our previous approach [ \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e ], in every window we decomposed the raw (mixed) power spectrum into fractal and oscillatory components using the Irregular Resampling Auto-Spectral Analysis (IRASA) technique [ \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e ]. The IRASA technique is illustrated on Fig.\u0026nbsp; \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e B, while for details, please see the original publication of Wen \u0026amp; Liu [ \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e ]. Frequency resolution was set to 0.0625 Hz and the resampling factor \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:h\\) \u003c/span\u003e \u003c/span\u003e ranged from 1.1 to 2.6 in increments of 0.1 [ \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e ]. The following conventional frequency bands were defined: delta ( \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\delta\\:\\) \u003c/span\u003e \u003c/span\u003e ): 1\u0026ndash;4 Hz, theta ( \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\theta\\:\\) \u003c/span\u003e \u003c/span\u003e ): 4\u0026ndash;8 Hz, alpha ( \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\alpha\\:\\) \u003c/span\u003e \u003c/span\u003e ): 8\u0026ndash;13 Hz, beta ( \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\beta\\:\\) \u003c/span\u003e \u003c/span\u003e ): 13\u0026ndash;25 Hz and gamma ( \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\gamma\\:\\) \u003c/span\u003e \u003c/span\u003e ): 25\u0026ndash;45 Hz. Frequency regimes over 45 Hz were inaccessible due to the filtering effects introduced by resampling in IRASA [ \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e ]. BLP estimates in all regimes were obtained via integration. For mixed and fractal spectra, BLP estimates were log-transformed, while for oscillatory spectra \u0026ndash; where positive power is not strictly ensured by IRASA \u0026ndash; we computed the difference between log-transformed mixed- and log-transformed fractal-power (i.e., the proportion of oscillatory power). Finally, each BLP index was characterized by their temporal mean (baseline) and standard deviation (fluctuation), taken over the 23 windows. Throughout this manuscript, we will use the following notation: frequency band is indicated by its preceding Greek-letter symbol (e.g., \u003cspan\u003e$\u003c/span\u003e\\alpha\u003cspan\u003e$\u003c/span\u003e), physiological state (EC) is indicated in superscript, spectrum type ( \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:mixd\\) \u003c/span\u003e \u003c/span\u003e , \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:frac\\) \u003c/span\u003e \u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:osci\\) \u003c/span\u003e \u003c/span\u003e for raw/mixed, fractal and oscillatory spectra, respectively) is indicated in subscript, and the mean or variance is indicated by encapsulating \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\mu\\:(\\bullet\\:)\\) \u003c/span\u003e \u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\sigma\\:(\\bullet\\:)\\) \u003c/span\u003e \u003c/span\u003e , respectively. For example, temporal fluctuation of eyes-closed BLP obtained from oscillatory spectra in the alpha regime is denoted \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\alpha\\:{BLP}_{osci}^{EC}\\) \u003c/span\u003e \u003c/span\u003e . Note that as part of the IRASA analysis [ \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e ], \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:1/f\\) \u003c/span\u003e \u003c/span\u003e spectral slope estimates in low- (1\u0026ndash;4 Hz, \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:{\\beta\\:}_{lo}\\) \u003c/span\u003e \u003c/span\u003e ) and high-frequency (20\u0026ndash;45 Hz, \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:{\\beta\\:}_{hi}\\) \u003c/span\u003e \u003c/span\u003e ) regimes were also obtained as in [ \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e ]. \u003c/p\u003e \u003cp\u003e\u0026lt;\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u0026gt;\u003c/p\u003e \u003cp\u003eFor dimensionality reduction and simplified physiological interpretability, channel-level estimates were also collapsed over six, established resting-state brain networks [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. We followed our previous approaches in this process [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] \u0026ndash; building on the work of Giacometti and colleagues [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] \u0026ndash; and assigned EEG channels to the Visual Network (VN), the Somatomotor Network (SM), the Dorsal Attention Network (DA), the Frontoparietal Network (FP), the combined Ventral Attention and Limbic Networks (VAL) and the Default Mode Network (DMN). In addition, we also performed all analyses on the global average indices taken over all 55 channels. The RSN assignment is illustrated on Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA.\u003c/p\u003e \u003cp\u003eA summary of analysis aspects is provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSummary of analysis aspects and their short description. EEG: electroencephalography; IRASA: Irregular Resampling Auto-Spectral Analysis\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAspect\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNotation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSummary\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpectral power\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBLP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBand-limited power (BLP) is obtained via Welch's periodogram method, then integrated in the given frequency range and log-transformed to facilitate normality.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFrequency range\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\delta\\:\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\theta\\:\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\alpha\\:\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\gamma\\:\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCanonical frequency bands Delta (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\delta\\:\\)\u003c/span\u003e\u003c/span\u003e): 1\u0026ndash;4 Hz, Theta (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\theta\\:\\)\u003c/span\u003e\u003c/span\u003e): 4\u0026ndash;8 Hz, Alpha (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\alpha\\:\\)\u003c/span\u003e\u003c/span\u003e): 8\u0026ndash;13 Hz, Beta (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e): 13\u0026ndash;25 Hz and Gamma (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\gamma\\:\\)\u003c/span\u003e\u003c/span\u003e): 25\u0026ndash;45 Hz.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpectrum type\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:mixd\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:frac\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:osci\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePower spectra separated by the IRASA method [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. The raw (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:mixd\\)\u003c/span\u003e\u003c/span\u003e) power spectrum is assumed to be the sum of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1/{f}^{\\beta\\:}\\)\u003c/span\u003e\u003c/span\u003e fractal (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:frac\\)\u003c/span\u003e\u003c/span\u003e) and narrow-band oscillatory (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:osci\\)\u003c/span\u003e\u003c/span\u003e) components, with the latter obtained by subtracting the fractal from the raw spectrum.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStatistical measure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:(\\bullet\\:)\\:and\\:\\sigma\\:(\\bullet\\:)\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEEG is analyzed in a time-resolved, sliding window fashion. Note that continuous 30-second EEG segments are standardized to ensure identical distribution of total power among participants, but individual windows (epochs) are not to allow for BLP dynamics to emerge. Then, brain activity is characterized by the baseline i.e., mean (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:(\\bullet\\:)\\)\u003c/span\u003e\u003c/span\u003e) and extent of temporal fluctuation i.e., standard deviation (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:(\\bullet\\:)\\)\u003c/span\u003e\u003c/span\u003e) of BLP estimates.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTopology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVN, SM, DA, VAL FP, DMN and global\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEEG channels are collapsed onto brain resting-state networks: the Visual Network (VN), the Somatomotor Network (SM), the Dorsal Attention Network (DA), the Frontoparietal Network (FP), the combined Ventral Attention and Limbic Networks (VAL) and the Default Mode Network (DMN). In addition, global average indices taken over all 55 channels are also assessed. See Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA for channel assignment.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eStatistical Evaluation\u003c/h3\u003e\n\u003cp\u003eSpectral features were contrasted between the HC and SZ groups on global-to-global and RSN-to-RSN levels. Channel-to-channel comparisons were performed only for exploratory and visualization purposes according to similar principles. In all cases, data normality was first probed with Lilliefors test, then either two-sample t-test or Mann-Whitney U test was employed accordingly. Outcomes were adjusted for multiple comparisons using False Discovery Rate (FDR) [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Effect sizes were characterized via Cohen's \u003cem\u003ed\u003c/em\u003e for normally distributed data, while according to \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:r=\\left|z\\right|/\\sqrt{N}\\)\u003c/span\u003e\u003c/span\u003e otherwise, where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:z\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:N\\)\u003c/span\u003e\u003c/span\u003e are the standardized test statistic and the total sample size, respectively [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThose global and RSN-level spectral features that revealed significant between-group difference \u0026ndash; and therefore could be considered characteristic traits of SZ \u0026ndash; were further subjected to correlation analysis with general (GEN-PANSS), negative (NEG-PANSS), positive (POS-PANSS) and total (SUM-PANSS) symptom scores. Possible confounding variables age, sex, years in education, disease duration and CPZ equivalent dose was estimated in both spectral indices and PANSS scores via multiple linear regression. Then, normality of the residuals was assessed via Lilliefors test, and the relationship between spectral indices and clinical symptoms were characterized via Pearson or Spearman correlation coefficient accordingly. Outcomes were not adjusted for multiple comparisons and therefore these outcomes should be treated as exploratory. Effect sizes were interpreted according to Cohen's guidelines, with Pearson or Spearman \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left|r\\right|=0.1\\)\u003c/span\u003e\u003c/span\u003e, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:0.3\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:0.5\\)\u003c/span\u003e\u003c/span\u003e indicating small, moderate or large effect [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e].\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eBaseline Distribution of Power\u003c/h2\u003e \u003cp\u003eGrand average power spectra illustrating baseline power are presented in the upper panels of Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. We found more power distributed to lower frequencies in SZ compared to HC, which could be attributed to the fractal spectral component. Specifically, both \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:\\left(\\delta\\:{BLP}_{frac}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e (\u003cb\u003eSupplementary Fig.\u0026nbsp;1\u003c/b\u003e) and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:\\left(\\theta\\:{BLP}_{frac}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e (\u003cb\u003eSupplementary Fig.\u0026nbsp;2\u003c/b\u003e) was increased in SZ in all RSNs \u0026ndash; also including the global average \u0026ndash; except for the SM network. Additionally, increased \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:\\left(\\alpha\\:{BLP}_{frac}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e was found in SZ compared to HC most prominently over the VN, while \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\mu\\:\\left(\\alpha\\:{BLP}_{osci}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e exhibited a tendency towards being higher in HC, although this never reached the level of statistical significance after adjusting for multiple comparisons (\u003cb\u003eSupplementary Fig.\u0026nbsp;3\u003c/b\u003e). Effect sizes were ranging between 0.2854 and 0.5680, indicating moderate effect.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSpectral Power Dynamics\u003c/h3\u003e\n\u003cp\u003eTemporal variance in grand average power spectra is presented in the lower panels of Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. SZ patients expressed reduced temporal variability of alpha and beta power. Specifically, there was a ubiquitous reduction of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e (\u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e) and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e (\u003cb\u003eSupplementary Fig.\u0026nbsp;5\u003c/b\u003e) in SZ compared to HC. Effect sizes ranged between 0.2632 and 0.4313, indicating moderate effect except for 2 cases. While this decrease in the alpha band could be attributed to oscillatory power (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{osci}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e, effect size: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\in\\:[0.3519;\\:1.0503]\\)\u003c/span\u003e\u003c/span\u003e), it was contingent with fractal power in the beta band (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{frac}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e, effect size: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\in\\:[0.2706;\\:0.7643]\\)\u003c/span\u003e\u003c/span\u003e) and more constrained to posterior cortical networks VN, DA and SM networks. Notably, the strongest effect was seen for global \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{osci}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0001, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{59}\\)\u003c/span\u003e\u003c/span\u003e=4.1540, Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.0503) followed by that over the FP network (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0006, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{59}\\)\u003c/span\u003e\u003c/span\u003e=3.6512, Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.9232), while for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{frac}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e over VN (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0.0037, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{59}\\)\u003c/span\u003e\u003c/span\u003e=3.0231, Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.7643) and DA (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0056, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{59}\\)\u003c/span\u003e\u003c/span\u003e=2.8770, Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.7274).\u003c/p\u003e \u003cp\u003e\u0026lt;\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u0026gt;\u003c/p\u003e\n\u003ch3\u003eSpectral Features and Clinical Symptoms\u003c/h3\u003e\n\u003cp\u003eCorrelation analysis revealed significant associations between beta-band dynamics over the DA network and clinical symptoms Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Precisely, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e over DA was inversely correlated to PANSS-NEG (Spearman \u003cem\u003er\u003c/em\u003e=-0.4994, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0055) and to PANSS-SUM (Spearman \u003cem\u003er\u003c/em\u003e=-0.3967, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0308), indicating that higher PANSS scores were associated with greater reduction in temporal variability in beta BLP. Note that these outcomes were not adjusted for multiple comparisons.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e\u0026lt;\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u0026gt;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur findings demonstrate that a time-resolved approach to BLP analysis in SZ is warranted, with separating aperiodic scale-free and oscillatory EEG spectral components. We found increased baseline spectral power in SZ in the delta and theta bands in line with previous literature [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e] while decreased variability in the alpha and beta bands. Decomposition of the spectrum indicated that the slow-wave baseline shift (\u003cb\u003eSupplementary Figs.\u0026nbsp;1\u0026amp;2\u003c/b\u003e) and decrease in beta fluctuations (\u003cb\u003eSupplementary Fig.\u0026nbsp;5\u003c/b\u003e) was predominantly attributed to broadband fractal activity, while the decrease in alpha fluctuations were more likely of oscillatory origin (\u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e). Baseline distinguishing EEG features were uncorrelated with disease symptoms, but beta-band dynamics showed associations with clinical presentations (see Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). These results point towards a general slowing of brain activity resulting in the allocation of more power towards smaller frequencies, with at the same time alpha oscillations and beta activity becoming less variable over time.\u003c/p\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eRelationship with E/I ratio\u003c/h2\u003e \u003cp\u003eIn our previous work [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] we exclusively focused on the fractal component of the power spectra to characterize spectral slopes in 1\u0026ndash;4 (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{lo}\\)\u003c/span\u003e\u003c/span\u003e) and 20\u0026ndash;45 Hz (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e) regimes. In a similar pattern, that analysis indicated no between-group difference in the temporal average of slopes, however we have found a reduction in the temporal variance of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e over the DA network. To address why these two patterns are highly similar, we performed linear mixed-effects (LME) modeling where we regressed the temporal evolution (i.e., over the 23 sliding windows) of global \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e on that of global oscillatory \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\alpha\\:BLP\\)\u003c/span\u003e\u003c/span\u003e similarly to [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e] with participant as a random effect. This revealed a very strong association between \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{osci}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e in both HC (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.6657, \u003cem\u003ep\u003c/em\u003e\u0026lt;\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{10}^{-5}\\)\u003c/span\u003e\u003c/span\u003e) and SZ (\u003cem\u003er\u003c/em\u003e=0.5059, \u003cem\u003ep\u003c/em\u003e\u0026lt;\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{10}^{-5}\\)\u003c/span\u003e\u003c/span\u003e), raising the concern that our previous outcomes were confounded (or driven by) oscillatory alpha instead of broadband activity. We propose two, not necessarily exclusive explanations for this strong dependence.\u003c/p\u003e \u003cp\u003eFirst, our previous analytical approach utilizing IRASA was not delicate enough to compensate for potential bias of oscillatory alpha activity on the 20\u0026ndash;45 Hz range when estimating \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e. Even though we defined the specifics of our analytical approach with this concern in mind \u0026ndash; i.e., excluding 4\u0026ndash;20 Hz regime from slope estimation, setting the rescaling parameter \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:h\\)\u003c/span\u003e\u003c/span\u003e ranging from 1.1 to 2.6 to avoid leakage and smearing effects [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u0026ndash;, unfortunately we cannot completely exclude this possibility. On the other hand, alpha activity is broadly considered as a neural signature of inhibitory tone [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. While also observed in [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e], recent studies indicated that dynamics in alpha power are related to E/I ratio: namely, alpha bursts with high amplitude might represent short periods with stronger inhibition, alternating with excitation/disinhibition [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. These phenomena are well in line with our findings and could provide an explanation on why we see an alignment between \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\alpha\\:BLP\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e. Nevertheless, this problem appears very relevant for studies inferring the E/I ratio from EEG analysis and warrants future research.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eReduced Power Fluctuations and Loss of Nonlinearity\u003c/h2\u003e \u003cp\u003eSystems that exhibit power-law (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1/{f}^{\\beta\\:}\\)\u003c/span\u003e\u003c/span\u003e) scaling as a local instead of global property are referred to as multifractals [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. In case of temporal processes this means that the fractal scaling exponent \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e changes over time instead of being constant [\u003cspan additionalcitationids=\"CR41 CR42\" citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. Such phenomena often emerge in biological systems as a result of nonlinear, antagonistic feedback loops such as the net effect of the sympathetic and parasympathetic nervous system [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e], or the interaction of excitatory and inhibitory neuronal populations [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. The temporal variability of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e therefore implies that the neural processes generating the EEG signal might be multifractal; however, such direct confirmation of multifractality usually demands immense computational capacity [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. Instead, we decided to test if nonlinearity might explain the emergence in the variability of alpha BLP, which we used as a proxy for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{hi}\\)\u003c/span\u003e\u003c/span\u003e given their strong correlation. For this purpose, we employed surrogate data testing via phase randomization [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e] to eliminate non-linearity (see \u003cb\u003eSupplementary Material\u003c/b\u003e for details). The outcomes of this analysis are presented on Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Locations where nonlinearity was confirmed for higher proportions (i.e., \u0026gt;\u0026thinsp;15 subjects) are mostly located over fronto-central and parieto-occipital regions. On the group level, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e is well above values from surrogate data in the HC group, while the obtained values are mostly within the spread of surrogate values in the SZ group. This analysis therefore strongly indicates that healthy alpha dynamics are governed by nonlinear dynamic processes, which appear mostly absent in the patient cohort.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e\u0026lt;\u003cem\u003eInsert\u003c/em\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e \u003cem\u003ehere\u003c/em\u003e\u0026gt;\u003c/p\u003e \u003cp\u003eWe also made an attempt to reproduce these results regarding \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e on an independent, open EEG dataset [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e] (see \u003cb\u003eSupplementary Material\u003c/b\u003e for details). This analysis did not indicate a significant difference between HC and SZ or suggested a loss of nonlinearity in the SZ group (see \u003cb\u003eSupplementary Fig.\u0026nbsp;6\u003c/b\u003e). Specifically, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e in SZ was comparable to that in HC (HC: 0.2639\u0026thinsp;\u0026plusmn;\u0026thinsp;0.1766, SZ: 0.2692\u0026thinsp;\u0026plusmn;\u0026thinsp;0.0699, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.9185, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}_{26}\\)\u003c/span\u003e\u003c/span\u003e=-0.1033, Cohen's \u003cem\u003ed\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0379). However, when comparing data from each group in the database of [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e] to their respective match in our dataset, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\alpha\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e was significantly lower in our patient sample (median: 0.1350, IQR: [0.1028; 0.2072], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0002, \u003cem\u003ez\u003c/em\u003e\u0026thinsp;=\u0026thinsp;3.7670, \u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.5679) compared to those of Olejarczyk \u0026amp; Jernajczyk, while we did not find a statistical difference between HC groups (median: 0.2557, IQR: [0.1472; 0.4927], \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.5981, \u003cem\u003ez\u003c/em\u003e=-0.5271, \u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.0786). There are a couple considerations that might explain these outcomes. First, the sample in [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e] contained only paranoid SZ patients in contrast to our cohort, in which only 7 out of 30 patients were diagnosed as paranoid-type SZ. More importantly, patients enrolled in [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e] were subjected to a medication washout period of at least seven days before EEG measurement, while our patients were on medication at the time of assessment. This raises the concern that the observed EEG patterns rather reflect response to medication than neural characteristics of the disease, however a recent review addressing pharmacotherapy effects on EEG in SZ does not indicate any previous works reporting changes in short-term alpha- or beta-band BLP dynamics in response to antipsychotic medication [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e] and the potential effects of antipsychotics on nonlinear characteristics of EEG are also poorly understood, although some reports might indicate a reduction in non-linearity in response to antipsychotic medication [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003ePotential Effects of Medication\u003c/h2\u003e \u003cp\u003e To further explore the potential confounding effects of antipsychotic medication, we correlated global EEG indices with CPZ equivalent dose values (see \u003cb\u003eSupplementary Material\u003c/b\u003e ). This analysis revealed a positive correlation between CPZ and \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{mixd}^{EC}\\right)\\) \u003c/span\u003e \u003c/span\u003e \u003cspan\u003e$\u003c/span\u003e (Spearman \u003cem\u003er\u003c/em\u003e \u0026thinsp;=\u0026thinsp;0.4399, \u003cem\u003ep\u003c/em\u003e \u0026thinsp;=\u0026thinsp;0.0150, unadjusted). While this indicates the influence of medication on EEG dynamics, it is also important to highlight that medication was heterogenous in our sample (see Table\u0026nbsp; \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e in [ \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e ] for details) with vastly different mechanisms of action, complicating potential influences on EEG markers [ \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e ]. Pharmaceutical therapy is a multifaceted endeavor in SZ, as often more severe symptoms (higher PANSS scores) warrant more potent pharmaceutical intervention, while a more aggressive treatment plan can result in a more successful repression of symptoms (lower PANSS scores). In our sample, correlation analysis between CPZ doses and PANSS scores ( \u003cb\u003eSupplementary Fig.\u0026nbsp;7\u003c/b\u003e ) indicated a significant positive correlation between CPZ and PANSS-NEG (Pearson \u003cem\u003er\u003c/em\u003e \u0026thinsp;=\u0026thinsp;0.3786, \u003cem\u003ep\u003c/em\u003e \u0026thinsp;=\u0026thinsp;0.0391) but no relationship with positive symptoms that can be more efficiently managed by antidopaminergic medication [ \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e ]. However, since the relationship between \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{mixd}^{EC}\\right)\\) \u003c/span\u003e \u003c/span\u003e and PANSS-NEG was anticorrelated after controlling for CPZ in both variables, these outcomes suggest that it is unlikely that EEG features and PANSS scores were both vastly confounded by medication and therefore it is probable that these EEG features could indeed reflect pathophysiology and/or therapy response. \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eLimitations and Future Perspectives\u003c/h2\u003e \u003cp\u003eThe reported EEG patterns in SZ were associated with clinical appearance, and therefore they could potentially be utilized in predicting therapy response and selecting the intervention strategy accordingly, which is still a major challenge for SZ [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. This calls for prospective, longitudinal studies with only first-episode, drug-na\u0026iuml;ve SZ participants, so it can be appropriately assessed if the observed patterns are a response to antipsychotic medication, they reflect ongoing pathology, or both. Longitudinal follow-up will also be impediment to evaluating if dynamic spectral features carry any predictive potential in therapeutic strategy planning. Some methodological concerns also need attention. In terms of separating broadband \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:1/f\\)\u003c/span\u003e\u003c/span\u003e from oscillatory activity in the EEG spectra, following a personalized approach could render our outcomes more relevant on the level of individual patients and is likely to offer a more appropriate characterization of spectral dynamics [\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e] by taking into consideration the inter-individual variability of peak alpha frequency [\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e] or scaling range-dependent spectral bimodality [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. Besides these issues, we also recognize limitations such as low and heterogenous patient sample, lack of explicit measures on other disease features such as cognitive dysfunction [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e], and our sole focus on EEG signals over e.g., genetic factors [\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. Nonetheless, we believe that this study is relevant for drawing attention to the relevance of a dynamic EEG analysis approach in SZ and can lay the foundation for future work overcoming these drawbacks.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eSummary\u003c/h2\u003e \u003cp\u003eWith regards to the relationships between dynamic EEG markers, symptom scores and medication data, we observed three patterns. First, a negative correlation between EEG markers and PANSS scores. This falls in line with the observed between-group differences (lower variability in SZ compared to HC), in that a more extensive loss of temporal BLP fluctuations was associated with more severe clinical symptoms in SZ. Second, there was a positive correlation between the CPZ equivalent dose and dynamic EEG indices, implying that medication in fact increased and not decreased temporal variability in BLP. Therefore, it is unlikely that between-group differences were primarily driven by pharmaceutical effects, even if we failed to replicate this pattern on an independent sample. Third, CPZ was positively correlated with both \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sigma\\:\\left(\\beta\\:{BLP}_{mixd}^{EC}\\right)\\)\u003c/span\u003e\u003c/span\u003e and PANSS-NEG, and therefore it should not explain the negative correlation between the latter two after controlling for its effect. These outcomes suggest that dynamic EEG indices can indeed be related to disease pathomechanism and/or therapy response in SZ, likely indicating a loss of nonlinearity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eData Availability Statement\u003c/h2\u003e \u003cp\u003eThe eyes-closed resting-state data analyzed this study has been made available in a public Zenodo repository titled \u0026ldquo;Resting-state EEG, clinical, and demographics data from schizophrenia patients and age-matched healthy controls\u0026rdquo; at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5281/zenodo.14808295\u003c/span\u003e\u003cspan address=\"10.5281/zenodo.14808295\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. All data analyses were carried out via Matlab (version 2023b) utilizing the EEGLAB toolbox [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] (version 2020.0) along with custom analysis codes. Matlab codes developed and used for analyzing the data and visualizing outcomes (along with pre-processed EEG) are provided in a public GitHub repository at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/samuelracz/schizophrenia_BLP_dynamics\u003c/span\u003e\u003cspan address=\"https://github.com/samuelracz/schizophrenia_BLP_dynamics\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Please note that details of all statistical evaluations, including raw and adjusted \u003cem\u003ep\u003c/em\u003e-values, test statistic values and effect size measures will be accessible via the shared code.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":" \u003ch2\u003eConflict of Interest Statement\u003c/h2\u003e \u003cp\u003eNone of the authors has any financial or other conflicting interests related to the work carried out in this study.\u003c/p\u003e \u003ch2\u003eAuthor Contributions\u003c/h2\u003e \u003cp\u003eF.S.R. designed the analytical pipeline, performed data screening, pre-processing and analysis, results interpretation, and wrote the initial draft of the manuscript. K.F., H.M., and Z.F. contributed to electrophysiological and clinical data collection and study conceptualization. M.B. contributed to data collection, screening and pre-processing. P.M. contributed to developing the analytical pipeline and interpreting results. G.C. conceptualized the study and provided oversight and guidance over data collection, analysis and all other study aspects. All authors contributed to manuscript development and revisions, and all authors have read and approved the final version.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eThis research was supported by the Hungarian Research Foundation Grant (OTKA FK 138385) and Hungarian Research Foundation Grant (OTKA PD 146424), as well as the Ministry of Innovation and Technology of Hungary from the National Research, Development and Innovation Fund, financed under the TKP2021-EGA-25 funding scheme.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbi-Dargham, A., et al., \u003cem\u003eCandidate biomarkers in psychiatric disorders: state of the field\u003c/em\u003e. World Psychiatry, 2023. 22(2): p. 236\u0026ndash;262.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRubio, J.M., et al., \u003cem\u003eLong-term Continuity of Antipsychotic Treatment for Schizophrenia: A Nationwide Study\u003c/em\u003e. Schizophr Bull, 2021. 47(6): p. 1611\u0026ndash;1620.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKahn, R.S., et al., \u003cem\u003eSchizophrenia\u003c/em\u003e. Nat Rev Dis Primers, 2015. 1: p. 15067.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeehan, A.J., et al., \u003cem\u003eClinical prediction models in psychiatry: a systematic review of two decades of progress and challenges\u003c/em\u003e. Mol Psychiatry, 2022. 27(6): p. 2700\u0026ndash;2708.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKraguljac, N.V., et al., \u003cem\u003eNeuroimaging Biomarkers in Schizophrenia\u003c/em\u003e. Am J Psychiatry, 2021. 178(6): p. 509\u0026ndash;521.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoutros, N.N., et al., \u003cem\u003eThe status of spectral EEG abnormality as a diagnostic test for schizophrenia\u003c/em\u003e. Schizophrenia Research, 2008. 99(1\u0026ndash;3): p. 225\u0026ndash;237.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNewson, J.J. and T.C. Thiagarajan, \u003cem\u003eEEG Frequency Bands in Psychiatric Disorders: A Review of Resting State Studies\u003c/em\u003e. Frontiers in Human Neuroscience, 2019. 12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerrottelli, A., et al., \u003cem\u003eEEG-Based Measures in At-Risk Mental State and Early Stages of Schizophrenia: A Systematic Review\u003c/em\u003e. Front Psychiatry, 2021. 12: p. 653642.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, H.S. and J.S. Kim, \u003cem\u003eImplication of Electrophysiological Biomarkers in Psychosis: Focusing on Diagnosis and Treatment Response\u003c/em\u003e. J Pers Med, 2022. 12(1).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDamaraju, E., et al., \u003cem\u003eDynamic functional connectivity analysis reveals transient states of dysconnectivity in schizophrenia\u003c/em\u003e. Neuroimage: Clinical, 2014. 5: p. 298\u0026ndash;308.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCattarinussi, G., et al., \u003cem\u003eDynamic functional connectivity in schizophrenia and bipolar disorder: A review of the evidence and associations with psychopathological features\u003c/em\u003e. Prog Neuropsychopharmacol Biol Psychiatry, 2023. 127: p. 110827.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao, R., E.J. Peterson, and B. Voytek, \u003cem\u003eInferring synaptic excitation/inhibition balance from field potentials\u003c/em\u003e. Neuroimage, 2017. 158: p. 70\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKehrer, C., et al., \u003cem\u003eAltered Excitatory-Inhibitory Balance in the NMDA-Hypofunction Model of Schizophrenia\u003c/em\u003e. Front Mol Neurosci, 2008. 1: p. 6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUliana, D.L., et al., \u003cem\u003eThe excitatory-inhibitory balance as a target for the development of novel drugs to treat schizophrenia\u003c/em\u003e. Biochem Pharmacol, 2024. 228: p. 116298.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL\u0026aacute;nyi, O., et al., \u003cem\u003eExcitation/inhibition imbalance in schizophrenia: a meta-analysis of inhibitory and excitatory TMS-EMG paradigms\u003c/em\u003e. Schizophrenia (Heidelb), 2024. 10(1): p. 56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eMultifractal and Entropy-Based Analysis of Delta Band Neural Activity Reveals Altered Functional Connectivity Dynamics in Schizophrenia\u003c/em\u003e. Frontiers in Systems Neuroscience, 2020. 14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eReduced temporal variability of cortical excitation/inhibition ratio in schizophrenia\u003c/em\u003e. Schizophrenia (Heidelb), 2025. 11(1): p. 20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKay, S.R., A. Fiszbein, and L.A. Opler, \u003cem\u003eThe positive and negative syndrome scale (PANSS) for schizophrenia\u003c/em\u003e. Schizophr Bull, 1987. 13(2): p. 261\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXia, M.R., J.H. Wang, and Y. He, \u003cem\u003eBrainNet Viewer: A Network Visualization Tool for Human Brain Connectomics\u003c/em\u003e. Plos One, 2013. 8(7).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDelorme, A. and S. Makeig, \u003cem\u003eEEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis\u003c/em\u003e. Journal of Neuroscience Methods, 2004. 134(1): p. 9\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWinkler, I., S. Haufe, and M. Tangermann, \u003cem\u003eAutomatic Classification of Artifactual ICA-Components for Artifact Removal in EEG Signals\u003c/em\u003e. Behavioral and Brain Functions, 2011. 7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWinkler, I., et al., \u003cem\u003eRobust artifactual independent component classification for BCI practitioners\u003c/em\u003e. Journal of Neural Engineering, 2014. 11(3).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWen, H.G. and Z.M. Liu, \u003cem\u003eSeparating Fractal and Oscillatory Components in the Power Spectrum of Neurophysiological Signal\u003c/em\u003e. Brain Topography, 2016. 29(1): p. 13\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eMultiple-Resampling Cross-Spectral Analysis: An Unbiased Tool for Estimating Fractal Connectivity With an Application to Neurophysiological Signals\u003c/em\u003e. Frontiers in Physiology, 2022. 13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYeo, B.T.T., et al., \u003cem\u003eThe organization of the human cerebral cortex estimated by intrinsic functional connectivity\u003c/em\u003e. Journal of Neurophysiology, 2011. 106(3): p. 1125\u0026ndash;1165.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eMultifractal and entropy analysis of resting-state electroencephalography reveals spatial organization in local dynamic functional connectivity\u003c/em\u003e. Scientific Reports, 2019. 9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eSeparating scale-free and oscillatory components of neural activity in schizophrenia\u003c/em\u003e. Brain Behav, 2021. 11(5): p. e02047.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiacometti, P., K.L. Perdue, and S.G. Diamond, \u003cem\u003eAlgorithm to find high density EEG scalp coordinates and analysis of their correspondence to structural and functional regions of the brain\u003c/em\u003e. Journal of Neuroscience Methods, 2014. 229: p. 84\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenjamini, Y. and Y. Hochberg, \u003cem\u003eControlling the False Discovery Rate - a Practical and Powerful Approach to Multiple Testing\u003c/em\u003e. Journal of the Royal Statistical Society Series B-Statistical Methodology, 1995. 57(1): p. 289\u0026ndash;300.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosenthal, R., H. Cooper, and L. Hedges, \u003cem\u003eParametric measures of effect size\u003c/em\u003e. The handbook of research synthesis, 1994. 621(2): p. 231\u0026ndash;244.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrydges, C.R., \u003cem\u003eEffect Size Guidelines, Sample Size Calculations, and Statistical Power in Gerontology\u003c/em\u003e. Innov Aging, 2019. 3(4): p. igz036.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOwens, E.M., et al., \u003cem\u003eElectrophysiological Endophenotypes for Schizophrenia\u003c/em\u003e. Harv Rev Psychiatry, 2016. 24(2): p. 129\u0026ndash;47.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecker, R., D. Van de Ville, and A. Kleinschmidt, \u003cem\u003eAlpha Oscillations Reduce Temporal Long-Range Dependence in Spontaneous Human Brain Activity\u003c/em\u003e. Journal of Neuroscience, 2018. 38(3): p. 755\u0026ndash;764.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlimesch, W., P. Sauseng, and S. Hanslmayr, \u003cem\u003eEEG alpha oscillations: The inhibition-timing hypothesis\u003c/em\u003e. Brain Research Reviews, 2007. 53(1): p. 63\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJensen, O. and A. Mazaheri, \u003cem\u003eShaping functional architecture by oscillatory alpha activity: gating by inhibition\u003c/em\u003e. Front Hum Neurosci, 2010. 4: p. 186.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLombardi, F., et al., \u003cem\u003eBeyond pulsed inhibition: Alpha oscillations modulate attenuation and amplification of neural activity in the awake resting state\u003c/em\u003e. Cell Rep, 2023. 42(10): p. 113162.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSano, M., et al., \u003cem\u003eAnalysis of the alpha activity envelope in electroencephalography in relation to the ratio of excitatory to inhibitory neural activity\u003c/em\u003e. PLoS One, 2024. 19(6): p. e0305082.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTel, T., \u003cem\u003eFractals, Multifractals, and Thermodynamics - an Introductory Review.\u003c/em\u003e Zeitschrift Fur Naturforschung Section a-a Journal of Physical Sciences, 1988. 43(12): p. 1154\u0026ndash;1174.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTheiler, J., \u003cem\u003eEstimating Fractal Dimension.\u003c/em\u003e Journal of the Optical Society of America a-Optics Image Science and Vision, 1990. 7(6): p. 1055\u0026ndash;1073.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIvanov, P.C., et al., \u003cem\u003eMultifractality in human heartbeat dynamics\u003c/em\u003e. Nature, 1999. 399(6735): p. 461\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKantelhardt, J.W., et al., \u003cem\u003eMultifractal detrended fluctuation analysis of nonstationary time series\u003c/em\u003e. Physica A: Statistical Mechanics and Its Applications, 2002. 316(1\u0026ndash;4): p. 87\u0026ndash;114.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eMultifractal dynamics of resting-state functional connectivity in the prefrontal cortex\u003c/em\u003e. Physiological Measurement, 2018. 39(2): p. 024003.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eMultifractal Dynamic Functional Connectivity in the Resting-State Brain\u003c/em\u003e. Front Physiol, 2018. 9: p. 1704.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIvanov, P., et al., \u003cem\u003eStochastic feedback and the regulation of biological rhythms\u003c/em\u003e. Europhys Lett, 1998. 43(4): p. 363\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshkenazy, Y., et al., \u003cem\u003eA stochastic model of human gait dynamics\u003c/em\u003e. Physica A: Statistical Mechanics and its Applications, 2002. 316(1\u0026ndash;4): p. 662\u0026ndash;670.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoil, S.S., et al., \u003cem\u003eCritical-State Dynamics of Avalanches and Oscillations Jointly Emerge from Balanced Excitation/Inhibition in Neuronal Networks\u003c/em\u003e. Journal of Neuroscience, 2012. 32(29): p. 9817\u0026ndash;9823.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChhabra, A.B., et al., \u003cem\u003eDirect determination of the f(alpha) singularity spectrum and its application to fully developed turbulence\u003c/em\u003e. Phys Rev A Gen Phys, 1989. 40(9): p. 5284\u0026ndash;5294.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKaposzta, Z., et al., \u003cem\u003eReal-Time Algorithm for Detrended Cross-Correlation Analysis of Long-Range Coupled Processes\u003c/em\u003e. Frontiers in Physiology, 2022. 13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTheiler, J., et al., \u003cem\u003eTesting for Nonlinearity in Time-Series - the Method of Surrogate Data\u003c/em\u003e. Physica D, 1992. 58(1\u0026ndash;4): p. 77\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOlejarczyk, E. and W. Jernajczyk, \u003cem\u003eGraph-based analysis of brain connectivity in schizophrenia\u003c/em\u003e. PLoS One, 2017. 12(11): p. e0188629.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDe Pieri, M., et al., \u003cem\u003ePharmaco-EEG of antipsychotic treatment response: a systematic review\u003c/em\u003e. Schizophrenia (Heidelb), 2023. 9(1): p. 85.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKang, U.G., et al., \u003cem\u003eNon-linear dynamic analysis of clozapine-induced electroencephalographic changes in schizophrenic patients\u0026ndash;a preliminary study\u003c/em\u003e. Prog Neuropsychopharmacol Biol Psychiatry, 2001. 25(6): p. 1229\u0026ndash;39.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTakahashi, T., et al., \u003cem\u003eAntipsychotics reverse abnormal EEG complexity in drug-naive schizophrenia: a multiscale entropy analysis\u003c/em\u003e. Neuroimage, 2010. 51(1): p. 173\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoth, B.L., D.J. Sheffler, and W.K. Kroeze, \u003cem\u003eMagic shotguns versus magic bullets: selectively non-selective drugs for mood disorders and schizophrenia\u003c/em\u003e. Nat Rev Drug Discov, 2004. 3(4): p. 353\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConley, R.R. and D.L. Kelly, \u003cem\u003eManagement of treatment resistance in schizophrenia\u003c/em\u003e. Biol Psychiatry, 2001. 50(11): p. 898\u0026ndash;911.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarbon, M. and C.U. Correll, \u003cem\u003eClinical predictors of therapeutic response to antipsychotics in schizophrenia\u003c/em\u003e. Dialogues Clin Neurosci, 2014. 16(4): p. 505\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonoghue, T., N. Schaworonkow, and B. Voytek, \u003cem\u003eMethodological considerations for studying neural oscillations\u003c/em\u003e. Eur J Neurosci, 2022. 55(11\u0026ndash;12): p. 3502\u0026ndash;3527.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHaegens, S., et al., \u003cem\u003eInter- and intra-individual variability in alpha peak frequency\u003c/em\u003e. Neuroimage, 2014. 92(100): p. 46\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRacz, F.S., et al., \u003cem\u003eMulti-Modal Spectral Parametrization Method (MMSPM) for analyzing EEG activity with distinct scaling regimes.\u003c/em\u003e arXiv preprint arXiv:2505.18117, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeefe, R.S. and P.D. Harvey, \u003cem\u003eCognitive impairment in schizophrenia\u003c/em\u003e. Handb Exp Pharmacol, 2012(213): p. 11\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuillozet-Bongaarts, A.L., et al., \u003cem\u003eAltered gene expression in the dorsolateral prefrontal cortex of individuals with schizophrenia\u003c/em\u003e. Mol Psychiatry, 2014. 19(4): p. 478\u0026ndash;85.\u003c/span\u003e\u003c/li\u003e\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":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6908048/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6908048/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSpectral features of the electroencephalogram (EEG) are essential for providing clinically relevant biomarkers in schizophrenia (SZ). Despite literature indicating altered short-scale neural dynamics in SZ, however, band-limited power (BLP) indices are rarely assessed in a time-resolved manner. To address this, here we evaluated static and dynamic BLP indices in a sample of 30 SZ patients and 31 healthy control (HC) individuals. Guided by recent findings on power spectral dynamics in SZ, we estimated total, and also decomposed fractal and oscillatory BLP in a sliding window manner from resting-state EEG recordings collected in eyes-closed (EC) resting-state. The SZ cohort was characterized by elevated baseline (mean over time) relative power in lower frequency regimes (delta, theta), mainly attributable to aperiodic fractal activity. In the higher regimes (alpha, beta), however, baseline was similar to HC with instead a widespread lessening in temporal fluctuations of both fractal and oscillatory activity. Variability in beta-BLP over the dorsal attention network was found correlated with negative symptoms in SZ. Finally, surrogate data testing indicated a loss of nonlinearity in neural dynamics as a potential mechanism for diminished BLP fluctuations.\u003c/p\u003e","manuscriptTitle":"Diminished Variability of Alpha and Beta Band-limited Power as a Neural Signature in Schizophrenia","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-02 09:23:34","doi":"10.21203/rs.3.rs-6908048/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2025-10-22T09:19:51+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-10-11T07:09:02+00:00","index":3,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-10-09T16:29:40+00:00","index":3,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-08-08T07:08:40+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-08-07T12:15:45+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-07-21T07:30:38+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2025-06-27T14:48:29+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-06-24T09:05:35+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-24T08:58:27+00:00","index":"","fulltext":""},{"type":"submitted","content":"Translational Psychiatry","date":"2025-06-23T19:30:42+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2025-06-17T10:17:58+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"28745874-c450-499b-8720-9050349c8d59","owner":[],"postedDate":"July 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":50702849,"name":"Health sciences/Diseases/Psychiatric disorders/Schizophrenia"},{"id":50702850,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2026-04-17T13:07:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-02 09:23:34","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6908048","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6908048","identity":"rs-6908048","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}