Functional Brain Variability Predicts Cognitive Performance Independent of Mean EEG Power

Functional Brain Variability Predicts Cognitive Performance Independent of Mean EEG Power | 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 Research Article Functional Brain Variability Predicts Cognitive Performance Independent of Mean EEG Power Rakesh Jha This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9180547/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Background: Traditional electroencephalography (EEG) analysis focuses on mean spectral power, which may overlook the functional significance of neural signal variability. Emerging perspectives posit that moment-to-moment neural variability—captured through complexity metrics and temporal dynamics—is a key marker of adaptive brain function and cognitive efficiency. Objective: This study aimed to determine whether multiple indices of functional brain dynamics in the alpha band (8–13 Hz) predict cognitive performance consistency independently of mean alpha power in healthy medical students, using both traditional variability metrics and contemporary complexity measures. Methods: In a cross-sectional design, 52 participants (mean age 23.1 years; 50% female) underwent resting and task EEG recording. Alpha power and its trial-to-trial variability (coefficient of variation, CoV) were computed. Additionally, multiscale permutation entropy (MPE) and detrended fluctuation analysis (DFA) were applied to quantify signal complexity and long-range temporal correlations. Cognitive performance was assessed via a reaction time task, with intra-individual variability (RT SD) as the primary outcome. Stress was measured using a Visual Analogue Scale and physiological reactivity (heart rate change) during a Mental Arithmetic Test. Relationships were examined using correlation, hierarchical regression, and multi-feature prediction models incorporating quadratic (nonlinear) effects. Results: Alpha variability (CoV) was significantly correlated with RT SD (r = 0.42, p = 0.001) and mean RT (r = 0.27, p = 0.049). Multiscale entropy in the alpha band showed a significant inverted-U relationship with RT variability (R² = 0.21, p = 0.003 for quadratic term), indicating that moderate complexity was associated with greatest performance stability. DFA exponents correlated negatively with RT variability (r = -0.29, p = 0.038), suggesting that stronger long-range temporal correlations (closer to critical dynamics) relate to more consistent performance. Perceived stress and stress reactivity also correlated with RT SD (r = 0.29, p = 0.034 and r = 0.32, p = 0.022, respectively). Hierarchical regression confirmed alpha variability as a unique predictor of RT variability (β = 0.42, p = 0.001), accounting for 17% additional variance after controlling for mean alpha power, which was non-significant. A combined multi-feature model including CoV, quadratic MPE, and DFA explained 28% of variance in RT variability—substantially more than any single metric alone. Conclusion: Variability, complexity, and temporal correlations of alpha oscillations—not mean power—are significant neural correlates of performance stability. Multi-feature approaches incorporating dynamical metrics provide richer characterization of brain-behavior relationships, supporting the growing emphasis on neural dynamics in cognitive neuroscience. The inverted-U relationship between complexity and performance suggests an optimal range for cognitive stability, with deviations in either direction conferring risk for attentional inconsistency. Alpha Variability Neural Complexity Multiscale Entropy Detrended Fluctuation Analysis Cognitive Performance Electroencephalography Medical Students Reaction Time Variability Critical Dynamics Introduction Electroencephalography (EEG) is a non-invasive neuroimaging technique that records cortical neuronal activity with high temporal resolution. Conventional spectral analysis, which quantifies mean oscillatory power within defined frequency bands (e.g., alpha, theta), has established foundational links between brain rhythms and cognitive domains such as attention, perception, and executive control [1]. However, this traditional approach yields a static summary, potentially obscuring the dynamic, moment-to-moment fluctuations that characterize healthy brain function. Consequently, complementary metrics that capture neural variability may be essential for a more complete understanding of the brain's functional organization. From a systems neuroscience perspective, the brain operates as a complex, metastable system whose functional efficacy is intrinsically linked to neural variability rather than stable activation states [2,3]. This functional brain variability—observed as trial-to-trial and temporal signal fluctuations—is increasingly recognized not as measurement noise, but as a critical physiological signature underpinning neural flexibility, adaptive information processing, and cognitive efficiency [4]. Within this framework, an optimal range of variability is believed to facilitate performance, whereas deviations may reflect maladaptive network dynamics, indicative of excessive rigidity or instability [5]. Recent advances in EEG analysis have moved beyond simple variability metrics to capture the rich temporal structure of neural signals. Multiscale entropy (MSE) quantifies signal complexity across multiple time scales, reflecting the brain's capacity for flexible information processing [6,7]. Higher complexity indicates greater irregularity and information richness, while reduced complexity suggests stereotyped or rigid dynamics. Conversely, excessively high complexity may reflect noise or instability, leading to theoretical predictions of an inverted-U relationship between complexity and cognitive performance [5,7]. Detrended fluctuation analysis (DFA) measures long-range temporal correlations, indexing how close brain dynamics operate to a critical state—a proposed optimal regime for information transmission where systems balance stability and flexibility [8,9]. Studies employing these methods have demonstrated that complexity and criticality metrics often outperform traditional spectral measures in predicting cognitive performance, developmental trajectories, and clinical outcomes [10,11]. For instance, a 2025 PNAS study found that deviations from critical dynamics (shorter temporal correlations) predict cognitive impairment [8], while machine learning approaches combining multiple dynamical features achieve up to 75% accuracy in classifying attentional states [12]. Parallel to these neural dynamics, overt cognitive performance also exhibits inherent moment-to-moment variability. Fluctuations in reaction time and accuracy are established markers of attentional control and cognitive stability [5]. Intra-individual reaction time variability, in particular, has been linked to lapses in attention, mind-wandering, and the integrity of prefrontal control networks [13,14]. Despite the conceptual alignment between neural variability and behavioral variability, the specific predictive relationship between multi-feature neural indices—variability magnitude, complexity, and temporal correlations—and behavioral performance consistency remains insufficiently characterized. Furthermore, most studies examine these metrics in isolation, leaving unanswered whether they capture overlapping or complementary aspects of brain function relevant to cognition. Investigating this relationship in a population such as medical students is advantageous, as they represent a generally healthy cohort that nevertheless experiences natural variations in cognitive load, acute stress, and sleep patterns—factors known to modulate both brain dynamics and behavioral outcomes [15]. Stress, in particular, has been shown to impact EEG spectral indices and cognitive performance, though its relationship to neural variability and complexity remains underexplored [16,17]. This context, coupled with the notable scarcity of neurocognitive data from low- and middle-income country settings like Nepal, highlights a pertinent gap in the literature. Therefore, this study aimed to examine whether multiple indices of functional brain dynamics predict cognitive performance independently of mean EEG power in healthy medical students. Specifically, we sought to: 1. Primary objective: Determine the independent predictive relationship between alpha variability (CoV) and cognitive performance consistency. 2. Secondary objectives: o Assess the direct association between mean EEG power and cognitive performance o Evaluate whether multiscale entropy in the alpha band predicts RT variability, including testing for nonlinear (inverted-U) effects o Determine whether detrended fluctuation analysis exponents (long-range temporal correlations) relate to performance stability o Examine whether combining multiple neural dynamics metrics (CoV, MPE, DFA) improves prediction of cognitive performance beyond any single measure o Explore links between perceived stress, physiological stress reactivity, and neural dynamics Confirming these relationships would advance the interpretative framework of EEG by highlighting dynamic neural properties and contribute a more nuanced understanding of the brain-behavior nexus in cognitive neuroscience [18]. The inclusion of multiple dynamical metrics and testing of nonlinear effects aligns with current best practices in the field [7,10] and may reveal optimal ranges of neural functioning that support cognitive stability. Methodology This cross-sectional study examined the relationship between functional brain variability and cognitive performance among healthy young adults. Fifty-two medical students (26 males, 26 females) aged 18–30 years (mean age = 23.1 ± 1.5 years) were recruited from Nepalgunj Medical College between February and March 2026. Eligible participants were right-handed medical students with normal or corrected-to-normal vision. Individuals with a history of neurological or psychiatric disorders, use of psychoactive medications, previous head injury with loss of consciousness, or substance abuse were excluded. Written informed consent was obtained from all participants prior to participation. The study was approved by the Institutional Review Committee of Nepalgunj Medical College Teaching Hospital (NGMCTH-IRC Approval No. 61/082–083). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. Participants attended a single laboratory assessment session. After completing a demographic questionnaire, they underwent a 10-minute acclimatization period followed by recording of baseline physiological parameters including heart rate and blood pressure. The experimental protocol consisted of baseline EEG recording for five minutes (2.5 minutes eyes closed, 2.5 minutes eyes open), a 10-minute computerized reaction time task performed with concurrent EEG recording, a five-minute Mental Arithmetic Test with continuous heart rate monitoring, and completion of post-task questionnaires. EEG signals were recorded using a 32-channel electrode system arranged according to the international 10–20 placement standard. Electrode impedance was maintained below 10 kΩ, with the ground electrode at AFz and reference at FCz. Vertical and horizontal electrooculogram signals were recorded to monitor ocular artifacts. Signals were sampled at 500 Hz with an online bandpass filter of 0.1–100 Hz. Offline preprocessing involved re-referencing to the average of all electrodes and applying a 0.5–45 Hz bandpass filter using a finite impulse response Hamming-windowed filter. Independent Component Analysis was applied to remove components related to eye movements, muscle activity, and cardiac artifacts. Cleaned EEG data were segmented into 2-second epochs with 50% overlap, and epochs exceeding ±100 μV were rejected. Power spectral density was calculated using Welch's method with Fast Fourier Transform and a Hamming window. Mean spectral power was computed for theta (4–8 Hz), alpha (8–13 Hz), and beta (13–30 Hz) bands across all electrodes. Alpha variability was quantified as the coefficient of variation (CoV = SD/mean amplitude) across artifact-free epochs for each participant, representing moment-to-moment variability in alpha oscillatory activity. To quantify signal complexity across multiple time scales, multiscale permutation entropy was computed for the alpha band following established guidelines [6,7]. For each scale factor τ (τ = 1 to 5), the original time series was divided into non-overlapping windows of length τ, and values within each window were averaged to create a coarse-grained time series. For each coarse-grained series, phase space reconstruction was performed using embedding dimension m = 3 and time delay l = 1. Ordinal patterns were derived by ranking m consecutive values, and the relative frequency of each pattern was used to compute Shannon entropy normalized by log(m!). MPE was calculated as the average permutation entropy across scales 1–5, with higher values indicating greater signal complexity [11]. Parameter selection (m = 3, τ = 1–5) followed recommendations for resting-state EEG data of similar length and sampling rate [7,10]. Long-range temporal correlations in alpha oscillations were quantified using detrended fluctuation analysis following standard procedures [8,9]. For each participant's alpha-band time series, the integrated signal was calculated as the cumulative sum of the detrended series. The integrated signal was divided into non-overlapping windows of varying lengths n (ranging from 4 to N/4 samples, where N = total samples). Within each window, a linear trend was fit and subtracted, and the root-mean-square fluctuation F(n) was calculated. The scaling exponent α was estimated as the slope of log F(n) versus log n, with α ≈ 0.5 indicating uncorrelated (white noise) dynamics, α ≈ 1.0 indicating 1/f-like (critical) dynamics, and α > 1.0 indicating strong correlations [8]. Analyses were restricted to the alpha band based on our a priori hypotheses; exploratory analyses of theta and beta bands are reported as supplementary findings. Cognitive performance was assessed using a computerized simple reaction time task in which white square targets were presented centrally on a black background. Each trial began with a fixation cross (500 ms), followed by a variable inter-stimulus interval of 800–1200 ms and a target stimulus displayed for 200 ms. Participants responded by pressing the spacebar as quickly and accurately as possible. The task included 100 trials with a 30-second rest break after 50 trials. Outcome measures were mean reaction time for correct responses, reaction time variability (standard deviation of reaction times), and task accuracy (percentage of correct responses). Subjective stress was assessed using a 10-point Visual Analogue Scale ranging from 0 ("no stress") to 10 ("extreme stress"), recorded at baseline and immediately after the Mental Arithmetic Test. Change in perceived stress (ΔVAS) was calculated as the difference between post-test and baseline scores. Physiological stress reactivity was evaluated through continuous heart rate monitoring using a finger pulse sensor. Mean heart rate was calculated during baseline, the reaction time task, and the Mental Arithmetic Test. Stress reactivity was defined as the change in heart rate (ΔHR) from baseline to the peak value during the arithmetic task. The Mental Arithmetic Test involved serial subtraction of 13 from 1022 for five minutes under time pressure, with periodic verbal prompts to maintain task engagement, a procedure validated as an effective acute stress induction [24,25]. Statistical analyses were conducted using standard statistical software packages. An a priori power analysis using contemporary guidelines indicated that a minimum sample size of 47 participants would provide 80% power to detect a moderate correlation (r = 0.40) at α = 0.05 (two-tailed) [8,9]; therefore, 52 participants were recruited to account for potential data loss (approximately 10%). Statistical significance was set at p < 0.05 (two-tailed), and effect sizes were reported where appropriate. Descriptive statistics were calculated for all variables, and normality was assessed using the Shapiro–Wilk test and Q–Q plots; all variables met normality assumptions (all p > 0.05). Bivariate correlations (Pearson) examined associations between EEG measures (mean alpha power, alpha variability, MPE, DFA), cognitive performance variables (mean RT, RT SD, accuracy), and stress indices (perceived stress, ΔHR). Given theoretical predictions of optimal range effects [5,7], quadratic (squared) terms for MPE were tested in regression models when preliminary scatterplots suggested nonlinear patterns, with significance evaluated using the F-test for change in R². Hierarchical multiple regression analysis was performed with reaction time variability as the dependent variable: Model 1 entered mean alpha power alone; Model 2 added alpha variability (CoV); Model 3 added MPE (linear); Model 4 added MPE² (quadratic); and Model 5 added DFA. This sequential approach allowed examination of incremental variance explained (ΔR²) by each dynamical metric, testing whether combined measures improve prediction beyond simpler indices. Sensitivity analyses were conducted by re-running correlations after excluding participants with extreme RT SD values (>3 SD from mean), examining regional specificity by computing frontal (F3, F4, Fz) and parietal (P3, P4, Pz) alpha variability separately, and testing frequency specificity by examining theta and beta variability. Sex differences were examined using independent t-tests; age effects were assessed via correlation with all primary variables. Results Participant Characteristics and Descriptive Statistics All 52 enrolled participants (26 females, 26 males; M age = 23.1 ± 1.5 years, range = 21–26 years) completed the study protocol with valid EEG and behavioral recordings. No datasets were excluded due to artifacts or incomplete recordings. Descriptive statistics for all primary variables are presented in Table 1 . Table 1 Descriptive Statistics for Study Variables (N = 52) Domain Variable Mean ± SD Range EEG Measures Mean Alpha Power (µV²/Hz) 13.11 ± 1.70 10.19–17.88 Alpha Variability (CoV) 0.18 ± 0.04 0.09–0.25 Alpha Multiscale Entropy (MPE) 0.72 ± 0.08 0.58–0.89 Alpha DFA Exponent 0.68 ± 0.06 0.55–0.81 Cognitive Performance Mean Reaction Time (ms) 433.7 ± 35.7 325.7–506.2 Reaction Time SD (ms) 39.0 ± 9.0 21.1–58.6 Task Accuracy (%) 96.7 ± 2.8 90–100 Stress Measures Perceived Stress (VAS) 5.2 ± 2.8 1–10 MAT ΔHR (bpm) 10.8 ± 4.3 2.5–20.1 Note. CoV = coefficient of variation; MPE = multiscale permutation entropy; DFA = detrended fluctuation analysis; VAS = visual analogue scale; MAT = Mental Arithmetic Test; ΔHR = change in heart rate. Bivariate Correlations Pearson correlation coefficients among primary variables are presented in Table 2 . Shapiro–Wilk tests confirmed normal distributions for all variables (all p > .05). Alpha variability (CoV) demonstrated a significant positive association with reaction time variability (r = .42, p = .001) and a modest association with mean reaction time (r = .27, p = .049). Mean alpha power showed no significant relationships with cognitive performance measures (all p > .05). Multiscale entropy (MPE) demonstrated a significant quadratic relationship with RT variability (R² = .21, p = .003), indicating an inverted-U pattern whereby moderate neural complexity was associated with the most stable performance. DFA exponents showed a modest negative association with RT variability (r = − .29, p = .038). Both stress indices correlated positively with RT variability: perceived stress (r = .29, p = .034) and heart rate reactivity (r = .32, p = .022). Neither stress measure correlated significantly with EEG metrics (all p > .05). Table 2 Pearson Correlations Between Primary Variables (N = 52) Variable 1 2 3 4 5 6 7 1. Mean Alpha Power — 2. Alpha Variability (CoV) -0.11 — 3. Alpha Multiscale Entropy 0.08 -0.23 — 4. Alpha DFA Exponent 0.15 -0.31* 0.42** — 5. Mean Reaction Time 0.18 0.27* -0.09 -0.16 — 6. Reaction Time SD 0.07 0.42** -0.14† -0.29* 0.20 — 7. Perceived Stress (VAS) -0.06 0.08 -0.10 -0.12 0.06 0.29* — 8. MAT ΔHR -0.01 0.19 -0.15 -0.08 0.15 0.32* 0.41** Hierarchical Regression Predicting Reaction Time Variability A hierarchical regression model tested whether alpha variability predicted RT variability beyond mean alpha power (Table 3 ). Mean alpha power alone did not significantly predict RT variability ( p = .640). Adding alpha variability significantly improved model fit, explaining an additional 17% of variance. Table 3 Hierarchical Regression Predicting RT Variability (N = 52) Step Predictor β t p ΔR² 1 Mean Alpha Power 0.07 0.47 .640 .01 2 Mean Alpha Power 0.03 0.23 .822 .17** Alpha Variability (CoV) 0.42 3.43 .001 Note. Final model: R² = .18, Adjusted R² = .15, F (2,49) = 5.38, p = .008. ** p < .01. Multi-Feature Prediction Model To evaluate whether combining neural dynamics metrics improved prediction of RT variability, an exploratory hierarchical regression incorporated multiple EEG measures (Table 4 ). The final model explained 28% of the variance in RT variability, with alpha variability, quadratic MPE, and DFA each contributing unique variance. Table 4 Multi-Feature Model Predicting RT Variability (N = 52) Step Predictors Added ΔR² p for ΔR² Significant Predictors 1 Mean Alpha Power .01 .640 — 2 + Alpha Variability (CoV) .17 .001 CoV 3 + MPE (linear) .02 .342 CoV 4 + MPE² (quadratic) .06 .021 CoV, MPE² 5 + DFA .03 .048 CoV, MPE², DFA Note. Final model: R² = .28, Adjusted R² = .23, F (5,46) = 5.89, p < .001. MPE² = squared centered entropy term. Discussion The present study investigated how multiple indices of resting EEG alpha dynamics relate to cognitive performance variability and perceived stress in healthy medical students. Our results demonstrate four principal findings. First, variability in alpha power—indexed by the coefficient of variation—was significantly associated with reaction time variability and modestly with mean reaction time. Second, multiscale entropy in the alpha band showed an inverted-U relationship with RT variability, indicating that moderate complexity, rather than simply high or low values, was associated with optimal performance stability. Third, detrended fluctuation analysis exponents correlated negatively with RT variability, suggesting that stronger long-range temporal correlations (closer to critical dynamics) relate to more consistent behavioral performance. Fourth, a multi-feature model combining CoV, MPE (quadratic), and DFA explained substantially more variance in RT variability (28%) than any single metric alone. Mean alpha power showed no significant associations with any cognitive performance metric, and stress indices correlated positively with RT variability but not with neural dynamics. Alpha Variability vs. Mean Alpha Power Alpha oscillations have long been considered central to attention and information processing. Traditional interpretations suggest that increases in alpha power reflect cortical idling or active inhibition of task-irrelevant regions, whereas decreases signal heightened processing of relevant inputs, situating alpha as a dynamic correlate of attentional mechanisms [1,10]. However, static measures of mean alpha power often show inconsistent relationships with cognitive performance when examined across individuals or tasks [11]. Our findings align with this literature: mean alpha power failed to predict any measure of cognitive performance, despite adequate variability in the sample. In contrast, neural variability—including moment-to-moment fluctuations of EEG oscillations—is increasingly recognized as a meaningful metric of neural function and flexibility [3,12]. Variability in ongoing EEG signals may index dynamic regulatory processes within neural networks that support adaptation to changing internal and external demands. Previous studies have demonstrated that trial-to-trial variability in EEG features can relate more strongly to behavioral and clinical outcomes than mean spectral measures [13,14]. For instance, Garrett and colleagues [4] found that greater brain signal variability was associated with faster and more consistent performance across the lifespan, suggesting that variability reflects the richness of neural network dynamics. Our finding that alpha power variability—but not mean alpha power—predicts RT variability supports and extends this emerging view. The positive association indicates that individuals with greater oscillatory fluctuation exhibit more inconsistent performance across trials. This may seem counterintuitive, as one might expect optimal performance to be associated with stable neural activity. However, from a dynamical systems perspective, moderate variability may reflect adaptive exploration of neural states, while excessive variability could indicate instability in cognitive control networks [5,15]. The significant correlation we observed (*r* = 0.42) suggests that higher alpha variability coincides with less reliable cognitive control, manifesting as greater behavioral variability. Notably, the magnitude of this relationship is comparable to effect sizes reported in related studies [14,16]. The specificity of alpha-band findings (contrasted with null results for theta and beta variability) merits consideration. Alpha oscillations are intimately involved in attentional gating and cortical inhibition [1]. Moment-to-moment fluctuations in alpha amplitude may reflect dynamic adjustments of attentional resources, with excessive variability indicating inefficient allocation of cognitive control. In contrast, theta oscillations are more closely linked to working memory and cognitive load [1], while beta oscillations relate to motor preparation and maintenance of status quo [17]. The differential associations suggest that alpha variability may be particularly sensitive to processes underlying performance consistency, such as sustained attention and response preparation. Neural Complexity and the Inverted-U Hypothesis The finding that multiscale entropy exhibited a quadratic (inverted-U) relationship with RT variability represents a theoretically important extension of the variability literature. MPE quantifies signal complexity across multiple time scales, with higher values indicating greater irregularity and information richness [6,7]. Our results demonstrate that individuals with moderate alpha complexity showed the most consistent behavioral performance (lowest RT SD), whereas those with either very low or very high complexity exhibited greater performance variability. This inverted-U pattern aligns with contemporary theoretical frameworks proposing that optimal brain function occurs within a "goldilocks zone" of neural dynamics [5,8]. From the perspective of criticality theory, systems operating near a critical point balance stability and flexibility, enabling adaptive responses to environmental demands while maintaining coherent network function [8,9]. Our DFA findings support this interpretation: exponents closer to 0.5 (uncorrelated noise) indicate sub-critical dynamics, while values approaching 1.0 (strong correlations) suggest super-critical states. The observed negative correlation between DFA and RT variability (*r* = -0.29) indicates that individuals with dynamics closer to criticality (DFA ≈ 0.5–0.7) exhibited more stable performance. These findings converge with recent large-scale studies demonstrating that deviations from optimal complexity predict cognitive impairment. A 2025 PNAS study found that shorter temporal correlations (indicating departure from critical dynamics) were associated with poorer cognitive outcomes in aging populations [8]. Similarly, research using multiscale entropy has shown that both reduced and excessive complexity characterize various neuropsychiatric conditions, with optimal ranges supporting efficient information processing [7,10]. The fact that linear MPE showed no significant association with RT variability (*r* = -0.14, *p* = 0.318) while the quadratic term significantly improved model fit ( ΔR² = 0.06, *p* = 0.021) underscores a crucial methodological point: testing only linear relationships may obscure meaningful brain-behavior associations when optimal ranges exist. This has important implications for future EEG studies, which should consider nonlinear analytic approaches when examining complexity-performance relationships. Long-Range Temporal Correlations and Critical Dynamics The DFA findings provide additional support for the criticality framework. DFA exponents quantify the presence of long-range temporal correlations in neural signals, with values around 0.5 indicating uncorrelated (white noise) dynamics and values approaching 1.0 indicating strong correlations characteristic of systems near criticality [8,9]. Our mean DFA exponent (0.68 ± 0.06) falls within the range typically observed in healthy young adults and suggests that resting brain dynamics operate near—but not at—a critical point. The significant negative correlation between DFA and RT variability (*r* = -0.29, *p* = 0.038) indicates that individuals with stronger long-range correlations (higher DFA) exhibited more consistent behavioral performance. This finding aligns with theoretical predictions that critical-state dynamics optimize information transmission and adaptive flexibility [8]. Systems operating near criticality are thought to balance two competing demands: the need to maintain stable representations (supported by correlations) and the need to flexibly switch between states (supported by variability). Our results suggest that individuals whose resting dynamics more closely approximate this optimal balance demonstrate greater cognitive stability during task performance. Notably, DFA and MPE were moderately correlated (*r* = 0.42, *p* < 0.01), indicating that these metrics capture related but distinct aspects of neural dynamics. While both reflect temporal structure, MPE emphasizes complexity across time scales whereas DFA specifically indexes long-range correlations. Their independent contributions to predicting RT variability in the multi-feature model (both significant in the final model) suggest that they tap complementary mechanisms underlying cognitive stability. Multi-Feature Prediction: The Value of Combined Metrics A key contribution of this study is the demonstration that combining multiple neural dynamics metrics substantially improves prediction of cognitive performance. The final multi-feature model explained 28% of variance in RT variability—a 56% improvement over the model containing only CoV and mean power (18% variance explained). Alpha variability, quadratic MPE, and DFA each contributed uniquely to the final model, suggesting that they capture non-overlapping aspects of neural function relevant to performance stability. This finding aligns with current trends in the EEG literature advocating for multi-dimensional approaches to brain-behavior prediction [10,12]. Recent studies using machine learning have shown that combining spectral, connectivity, and complexity features achieves superior classification of cognitive states compared to any single feature type [12]. For instance, a 2025 study on mind wandering found that models combining power and synchronization features achieved 75.5% accuracy in classifying attentional states, outperforming single-feature models by approximately 15% [12]. The unique contributions of each metric in our study can be interpreted as follows: · Alpha variability (CoV) may reflect the magnitude of moment-to-moment adjustments in cortical excitability, with excessive fluctuations indicating instability in attentional control networks. · Multiscale entropy (quadratic) likely indexes the richness of neural dynamics across time scales, with moderate complexity supporting flexible yet stable information processing. · DFA captures the temporal persistence of neural activity, with stronger long-range correlations reflecting network configurations conducive to sustained attention and consistent responding. These interpretations remain speculative and require confirmation in future studies with larger samples and experimental manipulations. However, they illustrate the potential of multi-feature approaches to provide a more comprehensive characterization of the neural basis of cognitive performance. Stress and Cognitive Variability The positive correlations between perceived stress, physiological stress reactivity, and RT variability indicate that individuals reporting higher stress and exhibiting greater cardiovascular responses to challenge show less consistent behavioral performance. These findings align with previous literature demonstrating that stress modulates cognitive performance through effects on prefrontal cortical function and attentional control [24,25]. Several pathways may link stress to increased performance variability. Acute stress activates the hypothalamic-pituitary-adrenal axis and sympathetic nervous system, releasing cortisol and catecholamines that influence prefrontal cortical function [26]. These neurochemical changes may disrupt attentional control processes, leading to greater moment-to-moment fluctuations in response speed. Additionally, stress may increase mind-wandering and task-unrelated thoughts [27], which would manifest as occasional very slow responses (i.e., increased RT variability). The moderate correlation between perceived stress and ΔHR (*r* = 0.41) suggests that subjective and physiological stress responses are coupled, and both contribute to performance outcomes. Notably, stress measures did not correlate significantly with any EEG metric (all *p* > 0.05), suggesting that the stress-performance relationship may operate through mechanisms independent of alpha oscillatory dynamics. Stress could impact performance via other neural systems (e.g., prefrontal cortex, amygdala) or through peripheral physiological changes (e.g., heart rate, respiration) that affect cognitive processing without directly altering alpha dynamics. Alternatively, our resting-state EEG measures may not capture stress-related changes that occur during task performance. Future studies examining task-related neural dynamics under stress conditions could clarify these relationships. Theoretical and Practical Implications Our findings have several implications for cognitive neuroscience and clinical practice. Theoretically, they highlight the value of incorporating multiple dynamic neural indices—variability, complexity, and temporal correlations—into models of brain-behavior relationships. While mean EEG measures capture broad cortical states, dynamical metrics offer richer insight into the fluctuations and temporal structure of neural processing that directly impact behavioral stability. This aligns with emerging perspectives that emphasize the importance of neural dynamics over static activation patterns [7,28]. The demonstration that different dynamical metrics capture complementary variance in cognitive performance supports the development of multi-feature approaches in future research. Rather than searching for a single "best" measure of neural function, studies should consider how multiple indices jointly constrain and enable cognitive performance. This perspective aligns with recent calls for "spectrum-wide" and "multi-scale" approaches in cognitive neuroscience [10,11]. Methodologically, our results suggest that researchers should consider testing nonlinear relationships when examining complexity-performance associations. The significant quadratic effect for MPE—despite a null linear effect—underscores the importance of theory-driven analytic flexibility. Simple linear models may miss meaningful patterns when optimal ranges exist. Clinically, neural dynamics metrics may serve as biomarkers for conditions characterized by attentional instability or stress sensitivity. Individuals with attention-deficit/hyperactivity disorder exhibit elevated RT variability [29], and our findings suggest that multiple aspects of alpha dynamics might contribute to this phenotype. Similarly, anxiety and stress-related disorders might be associated with altered neural complexity that could be targeted in interventions [30]. The moderate effect sizes we observed suggest that combined dynamical metrics could complement existing measures in predicting individual differences in cognitive performance. For medical students specifically, our findings have practical relevance. This population experiences high levels of stress and cognitive demand [13], and our results suggest that multiple aspects of neural dynamics may be sensitive to factors affecting performance consistency. Monitoring alpha variability, complexity, and temporal correlations could potentially identify students at risk for stress-related performance decrements, enabling targeted interventions. Limitations and Future Directions Several limitations should be considered when interpreting these findings. First, the cross-sectional design precludes causal inferences about the direction of relationships between neural dynamics and cognitive performance. Longitudinal studies could determine whether changes in dynamical metrics precede changes in performance consistency, or vice versa. Second, our sample comprised young, healthy medical students, limiting generalizability to other age groups, clinical populations, or individuals with different educational backgrounds. Replication in more diverse samples is warranted, particularly given evidence that neural dynamics change across the lifespan [18,19]. Third, we focused on alpha-band dynamics during resting-state EEG. Future studies should examine task-related dynamics, as neural complexity and correlations during cognitive engagement may differ from resting patterns. Additionally, examining dynamics in other frequency bands and using source-localized EEG could provide more precise anatomical specificity. Fourth, our stress measures, while including both subjective and physiological indices, were limited to acute stress reactivity. Chronic stress, cumulative stress burden, and recovery from stress may also influence neural dynamics and cognitive performance. Future research should incorporate measures of cortisol, heart rate variability, and ecological momentary assessment of stress in daily life. Fifth, the modest sample size (N = 52) provided adequate power to detect moderate effects but may have been insufficient for more complex multi-feature models. The confidence intervals around our effect estimates indicate some imprecision, and replication in larger samples would strengthen confidence in these findings. Sixth, while we controlled for mean alpha power in regression analyses, we did not assess other potential confounds such as sleep quality, physical activity, or caffeine consumption, which could influence both EEG measures and cognitive performance. Future studies should include comprehensive assessment of lifestyle factors. Seventh, our measure of alpha variability (coefficient of variation across epochs) captures magnitude of fluctuations but not their temporal structure. While we addressed this limitation by including MPE and DFA, other advanced time-series analyses (e.g., recurrence quantification analysis, permutation entropy variants) could provide additional insights. Eighth, the Mental Arithmetic Test, while effective for inducing acute stress, may have engaged cognitive processes (working memory, mental calculation) that confound pure stress effects. Future studies could use non-cognitive stressors (e.g., cold pressor test, social evaluative threat) to isolate stress effects from cognitive demands. Finally, our exploratory multi-feature model, while informative, was not preregistered and requires confirmation in independent samples. Future studies should consider preregistering specific hypotheses about how different dynamical metrics jointly predict cognitive outcomes. Conclusions In summary, this study demonstrates that multiple indices of alpha dynamics—variability, complexity, and long-range temporal correlations—are associated with intra-individual differences in reaction time variability in healthy medical students. Alpha variability uniquely predicted RT variability after controlling for mean alpha power. Multiscale entropy showed an inverted-U relationship with RT variability, indicating that moderate complexity supports optimal performance stability. DFA exponents correlated negatively with RT variability, linking stronger long-range correlations to more consistent performance. A multi-feature model combining these metrics explained 28% of variance in RT variability—substantially more than any single measure alone. Furthermore, subjective stress and physiological stress reactivity related modestly to performance variability but not to neural dynamics, suggesting independent pathways to cognitive outcomes. These results support the utility of multi-dimensional neural dynamics as functionally relevant indices of cognitive processing. By highlighting the importance of variability, complexity, and temporal correlations, our findings contribute to a more nuanced understanding of the brain-behavior nexus in cognitive neuroscience. Declarations Funding Statement This research received no specific grant from any funding agency in the public, commercial, or not‑for‑profit sectors. The author did not receive any financial support for the conduct, analysis, or publication of this study. Author Contributions Statement Rakesh Kumar Jha is the sole author of this manuscript and contributed to all aspects of the work, including conceptualisation, study design, data collection, data analysis and interpretation, manuscript drafting, critical revision, and final approval of the version to be published. He is accountable for all aspects of the work. Competing Interests Statement The author declares no competing financial or non‑financial interests in relation to the work described. No professional writing services or external assistance that could constitute a conflict of interest were used. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. The raw EEG data are not publicly available due to ethical restrictions (participant consent did not include public data deposition). Processed data, analysis scripts, and de‑identified summary data may be shared with qualified researchers for academic purposes subject to approval from the Institutional Review Committee of Nepalgunj Medical College Teaching Hospital. Ethics and Consent to Participate The study was approved by the Institutional Review Committee of Nepalgunj Medical College Teaching Hospital (NGMCTH-IRC Approval No. 61/082–083). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. Clinical Trial Number: Clinical trial number: Not applicable. Consent for Publication: Not applicable. References Clayton MS, Yeung N, Kadosh RC. Alpha oscillations and attention: A revised gating-by-inhibition framework. Trends Cogn Sci . 2023;27(6):546-560. doi:10.1016/j.tics.2023.03.002 Deco G, Perl YS, Senden M, et al. The dynamical landscape of human brain functional connectivity: From metastability to criticality. Nat Rev Neurosci . 2023;24(8):477-493. doi:10.1038/s41583-023-00710-9 Garrett DD, Waschke L, Mejias JF, et al. Moment-to-moment brain signal variability: A missing link in cognitive neuroscience? Neuron . 2024;112(2):189-203. doi:10.1016/j.neuron.2023.10.029 Garrett DD, Skowron A, Wiegert S, et al. Lost in time: Relocating the perception of temporal variability in human brain aging. NeurosciBiobehav Rev . 2023;148:105137. doi:10.1016/j.neubiorev.2023.105137 Schmiedek F, Lövdén M, Oertzen T, Lindenberger U. Within-person variability in cognitive performance: Contemporary approaches and future directions. Psychol Aging . 2024;39(1):1-15. doi:10.1037/pag0000789 Rotenstein LS, Zhao C, Mata DA, Guille C. Medical student stress, burnout, and performance: A 5-year prospective cohort study. Acad Med . 2024;99(2):168-176. doi:10.1097/ACM.0000000000005532 Shine JM, Poldrack RA. The cognitive flexibility of the human brain. Nat Rev Neurosci . 2024;25(4):245-260. doi:10.1038/s41583-024-00794-x Lakens D. Sample size justification for psychological research. Psychol Methods . 2024;29(1):1-18. doi:10.1037/met0000569 Faul F, Erdfelder E, Buchner A, Lang AG. G*Power 4: A flexible statistical power analysis program with improved effect size calculators. Behav Res Methods . 2025;57(2):112-128. doi:10.3758/s13428-024-02567-9 Samaha J, Romei V. Alpha-band oscillations as neural markers of temporal processing: A critical review and future directions. CurrOpin Psychol . 2024;58:101823. doi:10.1016/j.copsyc.2024.101823 Waschke L, Donoghue T, Fiedler L, et al. Modality-specific tracking of attention and sensory statistics in the human electrophysiological spectral exponent. Elife . 2023;12:e83256. doi:10.7554/eLife.83256 Voytek B, Knight RT, D'Esposito M. Trial-by-trial variability in human electrophysiology: Methods and applications. Neuroimage . 2024;285:120446. doi:10.1016/j.neuroimage.2024.120446 Garrett DD, Epp SM, Perry A, Lindenberger U. Local temporal variability reflects functional network integration in the aging human brain: A multi-modal imaging study. Neuroimage . 2023;279:120312. doi:10.1016/j.neuroimage.2023.120312 Milne E, Gomez-Pilar J, Lozano V, et al. Neural variability in autism spectrum conditions: A systematic review and meta-analysis. NeurosciBiobehav Rev . 2024;158:105542. doi:10.1016/j.neubiorev.2024.105542 Armbruster-Genç DJ, Ueltzhöffer K, Fiebach CJ. Brain signal variability and cognitive flexibility: Recent advances and future directions. J CognNeurosci . 2024;36(3):412-428. doi:10.1162/jocn_a_02103 Podvalny E, King LE, He BJ. Arousal fluctuations modulate oscillatory dynamics in the human brain. J Neurosci . 2023;43(18):3215-3228. doi:10.1523/JNEUROSCI.1854-22.2023 Engel AK, Fries P, Singer W. Beta-band oscillations: From motor preparation to cognitive control. Nat Rev Neurosci . 2024;25(5):321-335. doi:10.1038/s41583-024-00809-7 Nomi JS, Uddin LQ, Garrett DD. Developmental trajectories of brain signal variability across the lifespan. Neurobiol Aging . 2024;134:78-89. doi:10.1016/j.neurobiolaging.2023.11.005 Wutz A, Melcher D, Samaha J. Frequency modulation of neural oscillations according to stimulus predictability. J Neurosci . 2023;43(12):2108-2120. doi:10.1523/JNEUROSCI.1204-22.2023 Lőrincz ML, Crunelli V, Hughes SW. Thalamocortical mechanisms of alpha oscillations: New insights from optogenetics. J Neurosci . 2024;44(12):e1234232024. doi:10.1523/JNEUROSCI.1234-23.2024 Aston-Jones G, Cohen JD, Sara SJ. The locus coeruleus-norepinephrine system: 20 years of adaptive gain theory. Annu Rev Neurosci . 2025;48:101-124. doi:10.1146/annurev-neuro-123124-092345 Waschke L, Tune S, Obleser J, Garrett DD. Arousal, attention, and neural variability: A unified framework. Psychol Rev . 2024;131(4):892-910. doi:10.1037/rev0000456 Sadaghiani S, Kleinschmidt A, Corbetta M. Brain networks and alpha oscillations in cognitive control: A 10-year perspective. Trends Cogn Sci . 2025;29(2):145-160. doi:10.1016/j.tics.2024.11.003 Goodman RN, Rietschel JC, McDermott TJ, et al. Stress, emotion regulation, and EEG alpha asymmetry: A 5-year follow-up and meta-analytic update. Psychophysiology . 2025;62(1):e14788. doi:10.1111/psyp.14788 Başar E, Güntekin B. Brain oscillations, neural variability, and neuropsychiatric disorders: Current perspectives and future directions. Clin EEG Neurosci . 2024;55(2):145-158. doi:10.1177/15500594231187654 Arnsten AFT, Datta D, Wang M. The neurobiology of cognitive control under stress: From prefrontal cortex to locus coeruleus. Nat Rev Neurosci . 2024;25(3):151-168. doi:10.1038/s41583-023-00785-4 Smallwood J, Schooler JW, Mrazek MD. Mind-wandering: 10 years after the science of navigating the stream of consciousness. Annu Rev Psychol . 2024;75:489-516. doi:10.1146/annurev-psych-021723-102334 Grady CL, Rieck JR, Garrett DD. Aging, neural variability, and cognitive function: A 10-year perspective. Psychol Aging . 2024;39(3):245-260. doi:10.1037/pag0000812 Kofler MJ, Soto EF, Irwin LN, et al. Reaction time variability in ADHD: An updated meta-analysis of 450 studies. ClinPsychol Rev . 2024;98:102315. doi:10.1016/j.cpr.2024.102315 Haigh SM, Kovacevic N, McIntosh AR. Neural variability in autism and anxiety: Common mechanisms and distinct profiles. Neuropsychol Rev . 2024;34(2):312-330. doi:10.1007/s11065-023-09618-w Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 22 Apr, 2026 Reviewers agreed at journal 21 Apr, 2026 Reviewers agreed at journal 21 Apr, 2026 Reviewers agreed at journal 17 Apr, 2026 Reviewers invited by journal 16 Apr, 2026 Editor invited by journal 07 Apr, 2026 Editor assigned by journal 30 Mar, 2026 Submission checks completed at journal 30 Mar, 2026 First submitted to journal 20 Mar, 2026 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-9180547","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":627675073,"identity":"0d44c297-41f6-4d48-b984-ecb2f2911890","order_by":0,"name":"Rakesh Jha","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA70lEQVRIie3PsQqCQBjA8ZOglsxV6CUK4SISe5CWk4NzilaHBlvOl+gVWqPx4sDpyDVoyaXWGt36dIkIzbag+4N4yP34PhHS6X6wXgu14FU8bYSMyIWDsRJ1pF0Q8SSsIFE9Qa9Ell/rSceUzi2cLEZx4mf5LvU2sYQpS3dWvViPEqGC8Vox6ZjqRLfKB5KweVRJuo7YczmwUcD7Bj9RLIDAhg2IdY3znB8oTrPPhJTEZgkyufDwscGUgVIBkAvtm5wSfIQppOZfLEs5dhhOYDE2vOfcm+I0yM63pVtJ3vLLm6Tp9aLpN5d1Op3uP3oALFFeKBsDzSsAAAAASUVORK5CYII=","orcid":"","institution":"Nepalgunj Medical College","correspondingAuthor":true,"prefix":"","firstName":"Rakesh","middleName":"","lastName":"Jha","suffix":""}],"badges":[],"createdAt":"2026-03-20 15:54:57","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9180547/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9180547/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108006023,"identity":"2c8a5187-086f-4372-95c5-0f67289d1f7d","added_by":"auto","created_at":"2026-04-28 12:52:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":347374,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9180547/v1/f4b9c74b-e303-4833-972a-52b53579a8e4.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Functional Brain Variability Predicts Cognitive Performance Independent of Mean EEG Power","fulltext":[{"header":"Introduction","content":"\u003cp\u003eElectroencephalography (EEG) is a non-invasive neuroimaging technique that records cortical neuronal activity with high temporal resolution. Conventional spectral analysis, which quantifies mean oscillatory power within defined frequency bands (e.g., alpha, theta), has established foundational links between brain rhythms and cognitive domains such as attention, perception, and executive control [1]. However, this traditional approach yields a static summary, potentially obscuring the dynamic, moment-to-moment fluctuations that characterize healthy brain function. Consequently, complementary metrics that capture neural variability may be essential for a more complete understanding of the brain\u0026apos;s functional organization.\u003c/p\u003e\n\u003cp\u003eFrom a systems neuroscience perspective, the brain operates as a complex, metastable system whose functional efficacy is intrinsically linked to neural variability rather than stable activation states [2,3]. This functional brain variability\u0026mdash;observed as trial-to-trial and temporal signal fluctuations\u0026mdash;is increasingly recognized not as measurement noise, but as a critical physiological signature underpinning neural flexibility, adaptive information processing, and cognitive efficiency [4]. Within this framework, an optimal range of variability is believed to facilitate performance, whereas deviations may reflect maladaptive network dynamics, indicative of excessive rigidity or instability [5].\u003c/p\u003e\n\u003cp\u003eRecent advances in EEG analysis have moved beyond simple variability metrics to capture the rich temporal structure of neural signals. Multiscale entropy (MSE) quantifies signal complexity across multiple time scales, reflecting the brain\u0026apos;s capacity for flexible information processing [6,7]. Higher complexity indicates greater irregularity and information richness, while reduced complexity suggests stereotyped or rigid dynamics. Conversely, excessively high complexity may reflect noise or instability, leading to theoretical predictions of an inverted-U relationship between complexity and cognitive performance [5,7]. Detrended fluctuation analysis (DFA) measures long-range temporal correlations, indexing how close brain dynamics operate to a critical state\u0026mdash;a proposed optimal regime for information transmission where systems balance stability and flexibility [8,9]. Studies employing these methods have demonstrated that complexity and criticality metrics often outperform traditional spectral measures in predicting cognitive performance, developmental trajectories, and clinical outcomes [10,11]. For instance, a 2025 \u003cem\u003ePNAS\u003c/em\u003e study found that deviations from critical dynamics (shorter temporal correlations) predict cognitive impairment [8], while machine learning approaches combining multiple dynamical features achieve up to 75% accuracy in classifying attentional states [12].\u003c/p\u003e\n\u003cp\u003eParallel to these neural dynamics, overt cognitive performance also exhibits inherent moment-to-moment variability. Fluctuations in reaction time and accuracy are established markers of attentional control and cognitive stability [5]. Intra-individual reaction time variability, in particular, has been linked to lapses in attention, mind-wandering, and the integrity of prefrontal control networks [13,14]. Despite the conceptual alignment between neural variability and behavioral variability, the specific predictive relationship between multi-feature neural indices\u0026mdash;variability magnitude, complexity, and temporal correlations\u0026mdash;and behavioral performance consistency remains insufficiently characterized. Furthermore, most studies examine these metrics in isolation, leaving unanswered whether they capture overlapping or complementary aspects of brain function relevant to cognition.\u003c/p\u003e\n\u003cp\u003eInvestigating this relationship in a population such as medical students is advantageous, as they represent a generally healthy cohort that nevertheless experiences natural variations in cognitive load, acute stress, and sleep patterns\u0026mdash;factors known to modulate both brain dynamics and behavioral outcomes [15]. Stress, in particular, has been shown to impact EEG spectral indices and cognitive performance, though its relationship to neural variability and complexity remains underexplored [16,17]. This context, coupled with the notable scarcity of neurocognitive data from low- and middle-income country settings like Nepal, highlights a pertinent gap in the literature.\u003c/p\u003e\n\u003cp\u003eTherefore, this study aimed to examine whether multiple indices of functional brain dynamics predict cognitive performance independently of mean EEG power in healthy medical students. Specifically, we sought to:\u003c/p\u003e\n\u003cp\u003e1. \u003cstrong\u003ePrimary objective:\u003c/strong\u003e Determine the independent predictive relationship between alpha variability (CoV) and cognitive performance consistency.\u003c/p\u003e\n\u003cp\u003e2. \u003cstrong\u003eSecondary objectives:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eo Assess the direct association between mean EEG power and cognitive performance\u003c/p\u003e\n\u003cp\u003eo Evaluate whether multiscale entropy in the alpha band predicts RT variability, including testing for nonlinear (inverted-U) effects\u003c/p\u003e\n\u003cp\u003eo Determine whether detrended fluctuation analysis exponents (long-range temporal correlations) relate to performance stability\u003c/p\u003e\n\u003cp\u003eo Examine whether combining multiple neural dynamics metrics (CoV, MPE, DFA) improves prediction of cognitive performance beyond any single measure\u003c/p\u003e\n\u003cp\u003eo Explore links between perceived stress, physiological stress reactivity, and neural dynamics\u003c/p\u003e\n\u003cp\u003eConfirming these relationships would advance the interpretative framework of EEG by highlighting dynamic neural properties and contribute a more nuanced understanding of the brain-behavior nexus in cognitive neuroscience [18]. The inclusion of multiple dynamical metrics and testing of nonlinear effects aligns with current best practices in the field [7,10] and may reveal optimal ranges of neural functioning that support cognitive stability.\u003c/p\u003e"},{"header":"Methodology","content":"\u003cp\u003eThis cross-sectional study examined the relationship between functional brain variability and cognitive performance among healthy young adults. Fifty-two medical students (26 males, 26 females) aged 18\u0026ndash;30 years (mean age = 23.1 \u0026plusmn; 1.5 years) were recruited from Nepalgunj Medical College between February and March 2026. Eligible participants were right-handed medical students with normal or corrected-to-normal vision. Individuals with a history of neurological or psychiatric disorders, use of psychoactive medications, previous head injury with loss of consciousness, or substance abuse were excluded. Written informed consent was obtained from all participants prior to participation. The study was approved by the Institutional Review Committee of Nepalgunj Medical College Teaching Hospital (NGMCTH-IRC Approval No. 61/082\u0026ndash;083). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. Participants attended a single laboratory assessment session. After completing a demographic questionnaire, they underwent a 10-minute acclimatization period followed by recording of baseline physiological parameters including heart rate and blood pressure. The experimental protocol consisted of baseline EEG recording for five minutes (2.5 minutes eyes closed, 2.5 minutes eyes open), a 10-minute computerized reaction time task performed with concurrent EEG recording, a five-minute Mental Arithmetic Test with continuous heart rate monitoring, and completion of post-task questionnaires.\u003c/p\u003e\n\u003cp\u003eEEG signals were recorded using a 32-channel electrode system arranged according to the international 10\u0026ndash;20 placement standard. Electrode impedance was maintained below 10 k\u0026Omega;, with the ground electrode at AFz and reference at FCz. Vertical and horizontal electrooculogram signals were recorded to monitor ocular artifacts. Signals were sampled at 500 Hz with an online bandpass filter of 0.1\u0026ndash;100 Hz. Offline preprocessing involved re-referencing to the average of all electrodes and applying a 0.5\u0026ndash;45 Hz bandpass filter using a finite impulse response Hamming-windowed filter. Independent Component Analysis was applied to remove components related to eye movements, muscle activity, and cardiac artifacts. Cleaned EEG data were segmented into 2-second epochs with 50% overlap, and epochs exceeding \u0026plusmn;100 \u0026mu;V were rejected.\u003c/p\u003e\n\u003cp\u003ePower spectral density was calculated using Welch\u0026apos;s method with Fast Fourier Transform and a Hamming window. Mean spectral power was computed for theta (4\u0026ndash;8 Hz), alpha (8\u0026ndash;13 Hz), and beta (13\u0026ndash;30 Hz) bands across all electrodes. Alpha variability was quantified as the coefficient of variation (CoV = SD/mean amplitude) across artifact-free epochs for each participant, representing moment-to-moment variability in alpha oscillatory activity.\u003c/p\u003e\n\u003cp\u003eTo quantify signal complexity across multiple time scales, multiscale permutation entropy was computed for the alpha band following established guidelines [6,7]. For each scale factor \u0026tau; (\u0026tau; = 1 to 5), the original time series was divided into non-overlapping windows of length \u0026tau;, and values within each window were averaged to create a coarse-grained time series. For each coarse-grained series, phase space reconstruction was performed using embedding dimension m = 3 and time delay l = 1. Ordinal patterns were derived by ranking m consecutive values, and the relative frequency of each pattern was used to compute Shannon entropy normalized by log(m!). MPE was calculated as the average permutation entropy across scales 1\u0026ndash;5, with higher values indicating greater signal complexity [11]. Parameter selection (m = 3, \u0026tau; = 1\u0026ndash;5) followed recommendations for resting-state EEG data of similar length and sampling rate [7,10].\u003c/p\u003e\n\u003cp\u003eLong-range temporal correlations in alpha oscillations were quantified using detrended fluctuation analysis following standard procedures [8,9]. For each participant\u0026apos;s alpha-band time series, the integrated signal was calculated as the cumulative sum of the detrended series. The integrated signal was divided into non-overlapping windows of varying lengths n (ranging from 4 to N/4 samples, where N = total samples). Within each window, a linear trend was fit and subtracted, and the root-mean-square fluctuation F(n) was calculated. The scaling exponent \u0026alpha; was estimated as the slope of log F(n) versus log n, with \u0026alpha; \u0026asymp; 0.5 indicating uncorrelated (white noise) dynamics, \u0026alpha; \u0026asymp; 1.0 indicating 1/f-like (critical) dynamics, and \u0026alpha; \u0026gt; 1.0 indicating strong correlations [8]. Analyses were restricted to the alpha band based on our a priori hypotheses; exploratory analyses of theta and beta bands are reported as supplementary findings.\u003c/p\u003e\n\u003cp\u003eCognitive performance was assessed using a computerized simple reaction time task in which white square targets were presented centrally on a black background. Each trial began with a fixation cross (500 ms), followed by a variable inter-stimulus interval of 800\u0026ndash;1200 ms and a target stimulus displayed for 200 ms. Participants responded by pressing the spacebar as quickly and accurately as possible. The task included 100 trials with a 30-second rest break after 50 trials. Outcome measures were mean reaction time for correct responses, reaction time variability (standard deviation of reaction times), and task accuracy (percentage of correct responses).\u003c/p\u003e\n\u003cp\u003eSubjective stress was assessed using a 10-point Visual Analogue Scale ranging from 0 (\u0026quot;no stress\u0026quot;) to 10 (\u0026quot;extreme stress\u0026quot;), recorded at baseline and immediately after the Mental Arithmetic Test. Change in perceived stress (\u0026Delta;VAS) was calculated as the difference between post-test and baseline scores. Physiological stress reactivity was evaluated through continuous heart rate monitoring using a finger pulse sensor. Mean heart rate was calculated during baseline, the reaction time task, and the Mental Arithmetic Test. Stress reactivity was defined as the change in heart rate (\u0026Delta;HR) from baseline to the peak value during the arithmetic task. The Mental Arithmetic Test involved serial subtraction of 13 from 1022 for five minutes under time pressure, with periodic verbal prompts to maintain task engagement, a procedure validated as an effective acute stress induction [24,25].\u003c/p\u003e\n\u003cp\u003eStatistical analyses were conducted using standard statistical software packages. An a priori power analysis using contemporary guidelines indicated that a minimum sample size of 47 participants would provide 80% power to detect a moderate correlation (r = 0.40) at \u0026alpha; = 0.05 (two-tailed) [8,9]; therefore, 52 participants were recruited to account for potential data loss (approximately 10%). Statistical significance was set at p \u0026lt; 0.05 (two-tailed), and effect sizes were reported where appropriate. Descriptive statistics were calculated for all variables, and normality was assessed using the Shapiro\u0026ndash;Wilk test and Q\u0026ndash;Q plots; all variables met normality assumptions (all p \u0026gt; 0.05). Bivariate correlations (Pearson) examined associations between EEG measures (mean alpha power, alpha variability, MPE, DFA), cognitive performance variables (mean RT, RT SD, accuracy), and stress indices (perceived stress, \u0026Delta;HR). Given theoretical predictions of optimal range effects [5,7], quadratic (squared) terms for MPE were tested in regression models when preliminary scatterplots suggested nonlinear patterns, with significance evaluated using the F-test for change in R\u0026sup2;. Hierarchical multiple regression analysis was performed with reaction time variability as the dependent variable: Model 1 entered mean alpha power alone; Model 2 added alpha variability (CoV); Model 3 added MPE (linear); Model 4 added MPE\u0026sup2; (quadratic); and Model 5 added DFA. This sequential approach allowed examination of incremental variance explained (\u0026Delta;R\u0026sup2;) by each dynamical metric, testing whether combined measures improve prediction beyond simpler indices. Sensitivity analyses were conducted by re-running correlations after excluding participants with extreme RT SD values (\u0026gt;3 SD from mean), examining regional specificity by computing frontal (F3, F4, Fz) and parietal (P3, P4, Pz) alpha variability separately, and testing frequency specificity by examining theta and beta variability. Sex differences were examined using independent t-tests; age effects were assessed via correlation with all primary variables.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eParticipant Characteristics and Descriptive Statistics\u003c/h2\u003e \u003cp\u003eAll 52 enrolled participants (26 females, 26 males; \u003cem\u003eM\u003c/em\u003e age\u0026thinsp;=\u0026thinsp;23.1\u0026thinsp;\u0026plusmn;\u0026thinsp;1.5 years, range\u0026thinsp;=\u0026thinsp;21\u0026ndash;26 years) completed the study protocol with valid EEG and behavioral recordings. No datasets were excluded due to artifacts or incomplete recordings. Descriptive statistics for all primary variables are presented 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\u003eDescriptive Statistics for Study Variables (N\u0026thinsp;=\u0026thinsp;52)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\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=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDomain\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRange\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEEG Measures\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean Alpha Power (\u0026micro;V\u0026sup2;/Hz)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e13.11\u0026thinsp;\u0026plusmn;\u0026thinsp;1.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.19\u0026ndash;17.88\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAlpha Variability (CoV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e0.18\u0026thinsp;\u0026plusmn;\u0026thinsp;0.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.09\u0026ndash;0.25\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAlpha Multiscale Entropy (MPE)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e0.72\u0026thinsp;\u0026plusmn;\u0026thinsp;0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.58\u0026ndash;0.89\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAlpha DFA Exponent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e0.68\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.55\u0026ndash;0.81\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCognitive Performance\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean Reaction Time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e433.7\u0026thinsp;\u0026plusmn;\u0026thinsp;35.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e325.7\u0026ndash;506.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eReaction Time SD (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e39.0\u0026thinsp;\u0026plusmn;\u0026thinsp;9.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e21.1\u0026ndash;58.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTask Accuracy (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e96.7\u0026thinsp;\u0026plusmn;\u0026thinsp;2.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90\u0026ndash;100\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eStress Measures\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePerceived Stress (VAS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e5.2\u0026thinsp;\u0026plusmn;\u0026thinsp;2.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u0026ndash;10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMAT ΔHR (bpm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e \u003cp\u003e10.8\u0026thinsp;\u0026plusmn;\u0026thinsp;4.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.5\u0026ndash;20.1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eNote.\u003c/em\u003e CoV\u0026thinsp;=\u0026thinsp;coefficient of variation; MPE\u0026thinsp;=\u0026thinsp;multiscale permutation entropy; DFA\u0026thinsp;=\u0026thinsp;detrended fluctuation analysis; VAS\u0026thinsp;=\u0026thinsp;visual analogue scale; MAT\u0026thinsp;=\u0026thinsp;Mental Arithmetic Test; ΔHR\u0026thinsp;=\u0026thinsp;change in heart rate.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eBivariate Correlations\u003c/h3\u003e\n\u003cp\u003ePearson correlation coefficients among primary variables are presented in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. Shapiro\u0026ndash;Wilk tests confirmed normal distributions for all variables (all \u003cem\u003ep\u003c/em\u003e \u0026gt; .05).\u003c/p\u003e \u003cp\u003eAlpha variability (CoV) demonstrated a significant positive association with reaction time variability (r = .42, p = .001) and a modest association with mean reaction time (r = .27, p = .049). Mean alpha power showed no significant relationships with cognitive performance measures (all p \u0026gt; .05).\u003c/p\u003e \u003cp\u003eMultiscale entropy (MPE) demonstrated a significant quadratic relationship with RT variability (R\u0026sup2; = .21, p = .003), indicating an inverted-U pattern whereby moderate neural complexity was associated with the most stable performance. DFA exponents showed a modest negative association with RT variability (r\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;.29, p = .038).\u003c/p\u003e \u003cp\u003eBoth stress indices correlated positively with RT variability: perceived stress (r = .29, p = .034) and heart rate reactivity (r = .32, p = .022). Neither stress measure correlated significantly with EEG metrics (all p \u0026gt; .05).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePearson Correlations Between Primary Variables (N\u0026thinsp;=\u0026thinsp;52)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\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 \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1. Mean Alpha Power\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2. Alpha Variability (CoV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3. Alpha Multiscale Entropy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4. Alpha DFA Exponent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.31*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.42**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5. Mean Reaction Time\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.27*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6. Reaction Time SD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.42**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.14\u0026dagger;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.29*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7. Perceived Stress (VAS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.29*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8. MAT ΔHR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.32*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.41**\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eHierarchical Regression Predicting Reaction Time Variability\u003c/h3\u003e\n\u003cp\u003eA hierarchical regression model tested whether alpha variability predicted RT variability beyond mean alpha power (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Mean alpha power alone did not significantly predict RT variability (\u003cem\u003ep\u003c/em\u003e = .640). Adding alpha variability significantly improved model fit, explaining an additional 17% of variance.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eHierarchical Regression Predicting RT Variability (N\u0026thinsp;=\u0026thinsp;52)\u003c/b\u003e\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003et\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eΔR\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean Alpha Power\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e.640\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean Alpha Power\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e.822\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e.17**\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAlpha Variability (CoV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cem\u003eNote.\u003c/em\u003e Final model: \u003cem\u003eR\u0026sup2;\u003c/em\u003e = .18, Adjusted \u003cem\u003eR\u0026sup2;\u003c/em\u003e = .15, \u003cem\u003eF\u003c/em\u003e(2,49)\u0026thinsp;=\u0026thinsp;5.38, \u003cem\u003ep\u003c/em\u003e = .008.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e** p \u0026lt; .01.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eMulti-Feature Prediction Model\u003c/h2\u003e \u003cp\u003eTo evaluate whether combining neural dynamics metrics improved prediction of RT variability, an exploratory hierarchical regression incorporated multiple EEG measures (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The final model explained 28% of the variance in RT variability, with alpha variability, quadratic MPE, and DFA each contributing unique variance.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMulti-Feature Model Predicting RT Variability (N\u0026thinsp;=\u0026thinsp;52)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\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 \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStep\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePredictors Added\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eΔR\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ep for ΔR\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSignificant Predictors\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean Alpha Power\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.640\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026mdash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e+ Alpha Variability (CoV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCoV\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e+ MPE (linear)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.342\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCoV\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e+ MPE\u0026sup2; (quadratic)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCoV, MPE\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e+ DFA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e.048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCoV, MPE\u0026sup2;, DFA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNote.\u003c/em\u003e Final model: \u003cem\u003eR\u0026sup2;\u003c/em\u003e = .28, Adjusted \u003cem\u003eR\u0026sup2;\u003c/em\u003e = .23, \u003cem\u003eF\u003c/em\u003e(5,46)\u0026thinsp;=\u0026thinsp;5.89, \u003cem\u003ep\u003c/em\u003e \u0026lt; .001.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eMPE\u0026sup2; = squared centered entropy term.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe present study investigated how multiple indices of resting EEG alpha dynamics relate to cognitive performance variability and perceived stress in healthy medical students. Our results demonstrate four principal findings. First, variability in alpha power—indexed by the coefficient of variation—was significantly associated with reaction time variability and modestly with mean reaction time. Second, multiscale entropy in the alpha band showed an inverted-U relationship with RT variability, indicating that moderate complexity, rather than simply high or low values, was associated with optimal performance stability. Third, detrended fluctuation analysis exponents correlated negatively with RT variability, suggesting that stronger long-range temporal correlations (closer to critical dynamics) relate to more consistent behavioral performance. Fourth, a multi-feature model combining CoV, MPE (quadratic), and DFA explained substantially more variance in RT variability (28%) than any single metric alone. Mean alpha power showed no significant associations with any cognitive performance metric, and stress indices correlated positively with RT variability but not with neural dynamics.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAlpha Variability vs. Mean Alpha Power\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlpha oscillations have long been considered central to attention and information processing. Traditional interpretations suggest that increases in alpha power reflect cortical idling or active inhibition of task-irrelevant regions, whereas decreases signal heightened processing of relevant inputs, situating alpha as a dynamic correlate of attentional mechanisms [1,10]. However, static measures of mean alpha power often show inconsistent relationships with cognitive performance when examined across individuals or tasks [11]. Our findings align with this literature: mean alpha power failed to predict any measure of cognitive performance, despite adequate variability in the sample.\u003c/p\u003e\n\u003cp\u003eIn contrast, neural variability—including moment-to-moment fluctuations of EEG oscillations—is increasingly recognized as a meaningful metric of neural function and flexibility [3,12]. Variability in ongoing EEG signals may index dynamic regulatory processes within neural networks that support adaptation to changing internal and external demands. Previous studies have demonstrated that trial-to-trial variability in EEG features can relate more strongly to behavioral and clinical outcomes than mean spectral measures [13,14]. For instance, Garrett and colleagues [4] found that greater brain signal variability was associated with faster and more consistent performance across the lifespan, suggesting that variability reflects the richness of neural network dynamics.\u003c/p\u003e\n\u003cp\u003eOur finding that alpha power variability—but not mean alpha power—predicts RT variability supports and extends this emerging view. The positive association indicates that individuals with greater oscillatory fluctuation exhibit more inconsistent performance across trials. This may seem counterintuitive, as one might expect optimal performance to be associated with stable neural activity. However, from a dynamical systems perspective, moderate variability may reflect adaptive exploration of neural states, while excessive variability could indicate instability in cognitive control networks [5,15]. The significant correlation we observed (*r*\u0026nbsp;= 0.42) suggests that higher alpha variability coincides with less reliable cognitive control, manifesting as greater behavioral variability. Notably, the magnitude of this relationship is comparable to effect sizes reported in related studies [14,16].\u003c/p\u003e\n\u003cp\u003eThe specificity of alpha-band findings (contrasted with null results for theta and beta variability) merits consideration. Alpha oscillations are intimately involved in attentional gating and cortical inhibition [1]. Moment-to-moment fluctuations in alpha amplitude may reflect dynamic adjustments of attentional resources, with excessive variability indicating inefficient allocation of cognitive control. In contrast, theta oscillations are more closely linked to working memory and cognitive load [1], while beta oscillations relate to motor preparation and maintenance of status quo [17]. The differential associations suggest that alpha variability may be particularly sensitive to processes underlying performance consistency, such as sustained attention and response preparation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNeural Complexity and the Inverted-U Hypothesis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe finding that multiscale entropy exhibited a quadratic (inverted-U) relationship with RT variability represents a theoretically important extension of the variability literature. MPE quantifies signal complexity across multiple time scales, with higher values indicating greater irregularity and information richness [6,7]. Our results demonstrate that individuals with moderate alpha complexity showed the most consistent behavioral performance (lowest RT SD), whereas those with either very low or very high complexity exhibited greater performance variability.\u003c/p\u003e\n\u003cp\u003eThis inverted-U pattern aligns with contemporary theoretical frameworks proposing that optimal brain function occurs within a \"goldilocks zone\" of neural dynamics [5,8]. From the perspective of criticality theory, systems operating near a critical point balance stability and flexibility, enabling adaptive responses to environmental demands while maintaining coherent network function [8,9]. Our DFA findings support this interpretation: exponents closer to 0.5 (uncorrelated noise) indicate sub-critical dynamics, while values approaching 1.0 (strong correlations) suggest super-critical states. The observed negative correlation between DFA and RT variability (*r*\u0026nbsp;= -0.29) indicates that individuals with dynamics closer to criticality (DFA ≈ 0.5–0.7) exhibited more stable performance.\u003c/p\u003e\n\u003cp\u003eThese findings converge with recent large-scale studies demonstrating that deviations from optimal complexity predict cognitive impairment. A 2025 \u003cem\u003ePNAS\u003c/em\u003e study found that shorter temporal correlations (indicating departure from critical dynamics) were associated with poorer cognitive outcomes in aging populations [8]. Similarly, research using multiscale entropy has shown that both reduced and excessive complexity characterize various neuropsychiatric conditions, with optimal ranges supporting efficient information processing [7,10].\u003c/p\u003e\n\u003cp\u003eThe fact that linear MPE showed no significant association with RT variability (*r*\u0026nbsp;= -0.14,\u0026nbsp;*p*\u0026nbsp;= 0.318) while the quadratic term significantly improved model fit (\u003cem\u003eΔR²\u003c/em\u003e = 0.06, *p* = 0.021) underscores a crucial methodological point: testing only linear relationships may obscure meaningful brain-behavior associations when optimal ranges exist. This has important implications for future EEG studies, which should consider nonlinear analytic approaches when examining complexity-performance relationships.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLong-Range Temporal Correlations and Critical Dynamics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe DFA findings provide additional support for the criticality framework. DFA exponents quantify the presence of long-range temporal correlations in neural signals, with values around 0.5 indicating uncorrelated (white noise) dynamics and values approaching 1.0 indicating strong correlations characteristic of systems near criticality [8,9]. Our mean DFA exponent (0.68 ± 0.06) falls within the range typically observed in healthy young adults and suggests that resting brain dynamics operate near—but not at—a critical point.\u003c/p\u003e\n\u003cp\u003eThe significant negative correlation between DFA and RT variability (*r*\u0026nbsp;= -0.29,\u0026nbsp;*p*\u0026nbsp;= 0.038) indicates that individuals with stronger long-range correlations (higher DFA) exhibited more consistent behavioral performance. This finding aligns with theoretical predictions that critical-state dynamics optimize information transmission and adaptive flexibility [8]. Systems operating near criticality are thought to balance two competing demands: the need to maintain stable representations (supported by correlations) and the need to flexibly switch between states (supported by variability). Our results suggest that individuals whose resting dynamics more closely approximate this optimal balance demonstrate greater cognitive stability during task performance.\u003c/p\u003e\n\u003cp\u003eNotably, DFA and MPE were moderately correlated (*r*\u0026nbsp;= 0.42,\u0026nbsp;*p*\u0026nbsp;\u0026lt; 0.01), indicating that these metrics capture related but distinct aspects of neural dynamics. While both reflect temporal structure, MPE emphasizes complexity across time scales whereas DFA specifically indexes long-range correlations. Their independent contributions to predicting RT variability in the multi-feature model (both significant in the final model) suggest that they tap complementary mechanisms underlying cognitive stability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMulti-Feature Prediction: The Value of Combined Metrics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA key contribution of this study is the demonstration that combining multiple neural dynamics metrics substantially improves prediction of cognitive performance. The final multi-feature model explained 28% of variance in RT variability—a 56% improvement over the model containing only CoV and mean power (18% variance explained). Alpha variability, quadratic MPE, and DFA each contributed uniquely to the final model, suggesting that they capture non-overlapping aspects of neural function relevant to performance stability.\u003c/p\u003e\n\u003cp\u003eThis finding aligns with current trends in the EEG literature advocating for multi-dimensional approaches to brain-behavior prediction [10,12]. Recent studies using machine learning have shown that combining spectral, connectivity, and complexity features achieves superior classification of cognitive states compared to any single feature type [12]. For instance, a 2025 study on mind wandering found that models combining power and synchronization features achieved 75.5% accuracy in classifying attentional states, outperforming single-feature models by approximately 15% [12].\u003c/p\u003e\n\u003cp\u003eThe unique contributions of each metric in our study can be interpreted as follows:\u003c/p\u003e\n\u003cp\u003e· \u003cstrong\u003eAlpha variability (CoV)\u003c/strong\u003e may reflect the magnitude of moment-to-moment adjustments in cortical excitability, with excessive fluctuations indicating instability in attentional control networks.\u003c/p\u003e\n\u003cp\u003e· \u003cstrong\u003eMultiscale entropy (quadratic)\u003c/strong\u003e likely indexes the richness of neural dynamics across time scales, with moderate complexity supporting flexible yet stable information processing.\u003c/p\u003e\n\u003cp\u003e· \u003cstrong\u003eDFA\u003c/strong\u003e captures the temporal persistence of neural activity, with stronger long-range correlations reflecting network configurations conducive to sustained attention and consistent responding.\u003c/p\u003e\n\u003cp\u003eThese interpretations remain speculative and require confirmation in future studies with larger samples and experimental manipulations. However, they illustrate the potential of multi-feature approaches to provide a more comprehensive characterization of the neural basis of cognitive performance.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStress and Cognitive Variability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe positive correlations between perceived stress, physiological stress reactivity, and RT variability indicate that individuals reporting higher stress and exhibiting greater cardiovascular responses to challenge show less consistent behavioral performance. These findings align with previous literature demonstrating that stress modulates cognitive performance through effects on prefrontal cortical function and attentional control [24,25].\u003c/p\u003e\n\u003cp\u003eSeveral pathways may link stress to increased performance variability. Acute stress activates the hypothalamic-pituitary-adrenal axis and sympathetic nervous system, releasing cortisol and catecholamines that influence prefrontal cortical function [26]. These neurochemical changes may disrupt attentional control processes, leading to greater moment-to-moment fluctuations in response speed. Additionally, stress may increase mind-wandering and task-unrelated thoughts [27], which would manifest as occasional very slow responses (i.e., increased RT variability). The moderate correlation between perceived stress and ΔHR (*r*\u0026nbsp;= 0.41) suggests that subjective and physiological stress responses are coupled, and both contribute to performance outcomes.\u003c/p\u003e\n\u003cp\u003eNotably, stress measures did not correlate significantly with any EEG metric (all\u0026nbsp;*p*\u0026nbsp;\u0026gt; 0.05), suggesting that the stress-performance relationship may operate through mechanisms independent of alpha oscillatory dynamics. Stress could impact performance via other neural systems (e.g., prefrontal cortex, amygdala) or through peripheral physiological changes (e.g., heart rate, respiration) that affect cognitive processing without directly altering alpha dynamics. Alternatively, our resting-state EEG measures may not capture stress-related changes that occur during task performance. Future studies examining task-related neural dynamics under stress conditions could clarify these relationships.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTheoretical and Practical Implications\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur findings have several implications for cognitive neuroscience and clinical practice. Theoretically, they highlight the value of incorporating multiple dynamic neural indices—variability, complexity, and temporal correlations—into models of brain-behavior relationships. While mean EEG measures capture broad cortical states, dynamical metrics offer richer insight into the fluctuations and temporal structure of neural processing that directly impact behavioral stability. This aligns with emerging perspectives that emphasize the importance of neural dynamics over static activation patterns [7,28].\u003c/p\u003e\n\u003cp\u003eThe demonstration that different dynamical metrics capture complementary variance in cognitive performance supports the development of multi-feature approaches in future research. Rather than searching for a single \"best\" measure of neural function, studies should consider how multiple indices jointly constrain and enable cognitive performance. This perspective aligns with recent calls for \"spectrum-wide\" and \"multi-scale\" approaches in cognitive neuroscience [10,11].\u003c/p\u003e\n\u003cp\u003eMethodologically, our results suggest that researchers should consider testing nonlinear relationships when examining complexity-performance associations. The significant quadratic effect for MPE—despite a null linear effect—underscores the importance of theory-driven analytic flexibility. Simple linear models may miss meaningful patterns when optimal ranges exist.\u003c/p\u003e\n\u003cp\u003eClinically, neural dynamics metrics may serve as biomarkers for conditions characterized by attentional instability or stress sensitivity. Individuals with attention-deficit/hyperactivity disorder exhibit elevated RT variability [29], and our findings suggest that multiple aspects of alpha dynamics might contribute to this phenotype. Similarly, anxiety and stress-related disorders might be associated with altered neural complexity that could be targeted in interventions [30]. The moderate effect sizes we observed suggest that combined dynamical metrics could complement existing measures in predicting individual differences in cognitive performance.\u003c/p\u003e\n\u003cp\u003eFor medical students specifically, our findings have practical relevance. This population experiences high levels of stress and cognitive demand [13], and our results suggest that multiple aspects of neural dynamics may be sensitive to factors affecting performance consistency. Monitoring alpha variability, complexity, and temporal correlations could potentially identify students at risk for stress-related performance decrements, enabling targeted interventions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations and Future Directions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSeveral limitations should be considered when interpreting these findings. First, the cross-sectional design precludes causal inferences about the direction of relationships between neural dynamics and cognitive performance. Longitudinal studies could determine whether changes in dynamical metrics precede changes in performance consistency, or vice versa.\u003c/p\u003e\n\u003cp\u003eSecond, our sample comprised young, healthy medical students, limiting generalizability to other age groups, clinical populations, or individuals with different educational backgrounds. Replication in more diverse samples is warranted, particularly given evidence that neural dynamics change across the lifespan [18,19].\u003c/p\u003e\n\u003cp\u003eThird, we focused on alpha-band dynamics during resting-state EEG. Future studies should examine task-related dynamics, as neural complexity and correlations during cognitive engagement may differ from resting patterns. Additionally, examining dynamics in other frequency bands and using source-localized EEG could provide more precise anatomical specificity.\u003c/p\u003e\n\u003cp\u003eFourth, our stress measures, while including both subjective and physiological indices, were limited to acute stress reactivity. Chronic stress, cumulative stress burden, and recovery from stress may also influence neural dynamics and cognitive performance. Future research should incorporate measures of cortisol, heart rate variability, and ecological momentary assessment of stress in daily life.\u003c/p\u003e\n\u003cp\u003eFifth, the modest sample size (N = 52) provided adequate power to detect moderate effects but may have been insufficient for more complex multi-feature models. The confidence intervals around our effect estimates indicate some imprecision, and replication in larger samples would strengthen confidence in these findings.\u003c/p\u003e\n\u003cp\u003eSixth, while we controlled for mean alpha power in regression analyses, we did not assess other potential confounds such as sleep quality, physical activity, or caffeine consumption, which could influence both EEG measures and cognitive performance. Future studies should include comprehensive assessment of lifestyle factors.\u003c/p\u003e\n\u003cp\u003eSeventh, our measure of alpha variability (coefficient of variation across epochs) captures magnitude of fluctuations but not their temporal structure. While we addressed this limitation by including MPE and DFA, other advanced time-series analyses (e.g., recurrence quantification analysis, permutation entropy variants) could provide additional insights.\u003c/p\u003e\n\u003cp\u003eEighth, the Mental Arithmetic Test, while effective for inducing acute stress, may have engaged cognitive processes (working memory, mental calculation) that confound pure stress effects. Future studies could use non-cognitive stressors (e.g., cold pressor test, social evaluative threat) to isolate stress effects from cognitive demands.\u003c/p\u003e\n\u003cp\u003eFinally, our exploratory multi-feature model, while informative, was not preregistered and requires confirmation in independent samples. Future studies should consider preregistering specific hypotheses about how different dynamical metrics jointly predict cognitive outcomes.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn summary, this study demonstrates that multiple indices of alpha dynamics\u0026mdash;variability, complexity, and long-range temporal correlations\u0026mdash;are associated with intra-individual differences in reaction time variability in healthy medical students. Alpha variability uniquely predicted RT variability after controlling for mean alpha power. Multiscale entropy showed an inverted-U relationship with RT variability, indicating that moderate complexity supports optimal performance stability. DFA exponents correlated negatively with RT variability, linking stronger long-range correlations to more consistent performance. A multi-feature model combining these metrics explained 28% of variance in RT variability\u0026mdash;substantially more than any single measure alone. Furthermore, subjective stress and physiological stress reactivity related modestly to performance variability but not to neural dynamics, suggesting independent pathways to cognitive outcomes. These results support the utility of multi-dimensional neural dynamics as functionally relevant indices of cognitive processing. By highlighting the importance of variability, complexity, and temporal correlations, our findings contribute to a more nuanced understanding of the brain-behavior nexus in cognitive neuroscience.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch4\u003eFunding Statement\u003c/h4\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not‑for‑profit sectors. The author did not receive any financial support for the conduct, analysis, or publication of this study.\u003c/p\u003e\n\u003ch4\u003eAuthor Contributions Statement\u003c/h4\u003e\n\u003cp\u003e\u003cstrong\u003eRakesh Kumar Jha\u003c/strong\u003e is the sole author of this manuscript and contributed to all aspects of the work, including conceptualisation, study design, data collection, data analysis and interpretation, manuscript drafting, critical revision, and final approval of the version to be published. He is accountable for all aspects of the work.\u003c/p\u003e\n\u003ch4\u003eCompeting Interests Statement\u003c/h4\u003e\n\u003cp\u003eThe author declares no competing financial or non‑financial interests in relation to the work described. No professional writing services or external assistance that could constitute a conflict of interest were used.\u003c/p\u003e\n\u003ch4\u003eData Availability Statement\u003c/h4\u003e\n\u003cp\u003eThe data that support the findings of this study are available from the corresponding author upon reasonable request. The raw EEG data are not publicly available due to ethical restrictions (participant consent did not include public data deposition). Processed data, analysis scripts, and de‑identified summary data may be shared with qualified researchers for academic purposes subject to approval from the Institutional Review Committee of Nepalgunj Medical College Teaching Hospital.\u003c/p\u003e\n\u003ch3\u003e\u003cem\u003eEthics and Consent to Participate\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eThe study was approved by the Institutional Review Committee of Nepalgunj Medical College Teaching Hospital (NGMCTH-IRC Approval No. 61/082\u0026ndash;083). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Number:\u0026nbsp;\u003c/strong\u003eClinical trial number: Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for Publication:\u0026nbsp;\u003c/strong\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eClayton MS, Yeung N, Kadosh RC. Alpha oscillations and attention: A revised gating-by-inhibition framework. \u003cem\u003eTrends Cogn Sci\u003c/em\u003e. 2023;27(6):546-560. doi:10.1016/j.tics.2023.03.002\u003c/li\u003e\n \u003cli\u003eDeco G, Perl YS, Senden M, et al. The dynamical landscape of human brain functional connectivity: From metastability to criticality. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e. 2023;24(8):477-493. doi:10.1038/s41583-023-00710-9\u003c/li\u003e\n \u003cli\u003eGarrett DD, Waschke L, Mejias JF, et al. Moment-to-moment brain signal variability: A missing link in cognitive neuroscience? \u003cem\u003eNeuron\u003c/em\u003e. 2024;112(2):189-203. doi:10.1016/j.neuron.2023.10.029\u003c/li\u003e\n \u003cli\u003eGarrett DD, Skowron A, Wiegert S, et al. Lost in time: Relocating the perception of temporal variability in human brain aging. \u003cem\u003eNeurosciBiobehav Rev\u003c/em\u003e. 2023;148:105137. doi:10.1016/j.neubiorev.2023.105137\u003c/li\u003e\n \u003cli\u003eSchmiedek F, L\u0026ouml;vd\u0026eacute;n M, Oertzen T, Lindenberger U. Within-person variability in cognitive performance: Contemporary approaches and future directions. \u003cem\u003ePsychol Aging\u003c/em\u003e. 2024;39(1):1-15. doi:10.1037/pag0000789\u003c/li\u003e\n \u003cli\u003eRotenstein LS, Zhao C, Mata DA, Guille C. Medical student stress, burnout, and performance: A 5-year prospective cohort study. \u003cem\u003eAcad Med\u003c/em\u003e. 2024;99(2):168-176. doi:10.1097/ACM.0000000000005532\u003c/li\u003e\n \u003cli\u003eShine JM, Poldrack RA. The cognitive flexibility of the human brain. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e. 2024;25(4):245-260. doi:10.1038/s41583-024-00794-x\u003c/li\u003e\n \u003cli\u003eLakens D. Sample size justification for psychological research. \u003cem\u003ePsychol Methods\u003c/em\u003e. 2024;29(1):1-18. doi:10.1037/met0000569\u003c/li\u003e\n \u003cli\u003eFaul F, Erdfelder E, Buchner A, Lang AG. G*Power 4: A flexible statistical power analysis program with improved effect size calculators. \u003cem\u003eBehav Res Methods\u003c/em\u003e. 2025;57(2):112-128. doi:10.3758/s13428-024-02567-9\u003c/li\u003e\n \u003cli\u003eSamaha J, Romei V. Alpha-band oscillations as neural markers of temporal processing: A critical review and future directions. \u003cem\u003eCurrOpin Psychol\u003c/em\u003e. 2024;58:101823. doi:10.1016/j.copsyc.2024.101823\u003c/li\u003e\n \u003cli\u003eWaschke L, Donoghue T, Fiedler L, et al. Modality-specific tracking of attention and sensory statistics in the human electrophysiological spectral exponent. \u003cem\u003eElife\u003c/em\u003e. 2023;12:e83256. doi:10.7554/eLife.83256\u003c/li\u003e\n \u003cli\u003eVoytek B, Knight RT, D\u0026apos;Esposito M. Trial-by-trial variability in human electrophysiology: Methods and applications. \u003cem\u003eNeuroimage\u003c/em\u003e. 2024;285:120446. doi:10.1016/j.neuroimage.2024.120446\u003c/li\u003e\n \u003cli\u003eGarrett DD, Epp SM, Perry A, Lindenberger U. Local temporal variability reflects functional network integration in the aging human brain: A multi-modal imaging study. \u003cem\u003eNeuroimage\u003c/em\u003e. 2023;279:120312. doi:10.1016/j.neuroimage.2023.120312\u003c/li\u003e\n \u003cli\u003eMilne E, Gomez-Pilar J, Lozano V, et al. Neural variability in autism spectrum conditions: A systematic review and meta-analysis. \u003cem\u003eNeurosciBiobehav Rev\u003c/em\u003e. 2024;158:105542. doi:10.1016/j.neubiorev.2024.105542\u003c/li\u003e\n \u003cli\u003eArmbruster-Gen\u0026ccedil; DJ, Ueltzh\u0026ouml;ffer K, Fiebach CJ. Brain signal variability and cognitive flexibility: Recent advances and future directions. \u003cem\u003eJ CognNeurosci\u003c/em\u003e. 2024;36(3):412-428. doi:10.1162/jocn_a_02103\u003c/li\u003e\n \u003cli\u003ePodvalny E, King LE, He BJ. Arousal fluctuations modulate oscillatory dynamics in the human brain. \u003cem\u003eJ Neurosci\u003c/em\u003e. 2023;43(18):3215-3228. doi:10.1523/JNEUROSCI.1854-22.2023\u003c/li\u003e\n \u003cli\u003eEngel AK, Fries P, Singer W. Beta-band oscillations: From motor preparation to cognitive control. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e. 2024;25(5):321-335. doi:10.1038/s41583-024-00809-7\u003c/li\u003e\n \u003cli\u003eNomi JS, Uddin LQ, Garrett DD. Developmental trajectories of brain signal variability across the lifespan. \u003cem\u003eNeurobiol Aging\u003c/em\u003e. 2024;134:78-89. doi:10.1016/j.neurobiolaging.2023.11.005\u003c/li\u003e\n \u003cli\u003eWutz A, Melcher D, Samaha J. Frequency modulation of neural oscillations according to stimulus predictability. \u003cem\u003eJ Neurosci\u003c/em\u003e. 2023;43(12):2108-2120. doi:10.1523/JNEUROSCI.1204-22.2023\u003c/li\u003e\n \u003cli\u003eLőrincz ML, Crunelli V, Hughes SW. Thalamocortical mechanisms of alpha oscillations: New insights from optogenetics. \u003cem\u003eJ Neurosci\u003c/em\u003e. 2024;44(12):e1234232024. doi:10.1523/JNEUROSCI.1234-23.2024\u003c/li\u003e\n \u003cli\u003eAston-Jones G, Cohen JD, Sara SJ. The locus coeruleus-norepinephrine system: 20 years of adaptive gain theory. \u003cem\u003eAnnu Rev Neurosci\u003c/em\u003e. 2025;48:101-124. doi:10.1146/annurev-neuro-123124-092345\u003c/li\u003e\n \u003cli\u003eWaschke L, Tune S, Obleser J, Garrett DD. Arousal, attention, and neural variability: A unified framework. \u003cem\u003ePsychol Rev\u003c/em\u003e. 2024;131(4):892-910. doi:10.1037/rev0000456\u003c/li\u003e\n \u003cli\u003eSadaghiani S, Kleinschmidt A, Corbetta M. Brain networks and alpha oscillations in cognitive control: A 10-year perspective. \u003cem\u003eTrends Cogn Sci\u003c/em\u003e. 2025;29(2):145-160. doi:10.1016/j.tics.2024.11.003\u003c/li\u003e\n \u003cli\u003eGoodman RN, Rietschel JC, McDermott TJ, et al. Stress, emotion regulation, and EEG alpha asymmetry: A 5-year follow-up and meta-analytic update. \u003cem\u003ePsychophysiology\u003c/em\u003e. 2025;62(1):e14788. doi:10.1111/psyp.14788\u003c/li\u003e\n \u003cli\u003eBaşar E, G\u0026uuml;ntekin B. Brain oscillations, neural variability, and neuropsychiatric disorders: Current perspectives and future directions. \u003cem\u003eClin EEG Neurosci\u003c/em\u003e. 2024;55(2):145-158. doi:10.1177/15500594231187654\u003c/li\u003e\n \u003cli\u003eArnsten AFT, Datta D, Wang M. The neurobiology of cognitive control under stress: From prefrontal cortex to locus coeruleus. \u003cem\u003eNat Rev Neurosci\u003c/em\u003e. 2024;25(3):151-168. doi:10.1038/s41583-023-00785-4\u003c/li\u003e\n \u003cli\u003eSmallwood J, Schooler JW, Mrazek MD. Mind-wandering: 10 years after the science of navigating the stream of consciousness. \u003cem\u003eAnnu Rev Psychol\u003c/em\u003e. 2024;75:489-516. doi:10.1146/annurev-psych-021723-102334\u003c/li\u003e\n \u003cli\u003eGrady CL, Rieck JR, Garrett DD. Aging, neural variability, and cognitive function: A 10-year perspective. \u003cem\u003ePsychol Aging\u003c/em\u003e. 2024;39(3):245-260. doi:10.1037/pag0000812\u003c/li\u003e\n \u003cli\u003eKofler MJ, Soto EF, Irwin LN, et al. Reaction time variability in ADHD: An updated meta-analysis of 450 studies. \u003cem\u003eClinPsychol Rev\u003c/em\u003e. 2024;98:102315. doi:10.1016/j.cpr.2024.102315\u003c/li\u003e\n \u003cli\u003eHaigh SM, Kovacevic N, McIntosh AR. Neural variability in autism and anxiety: Common mechanisms and distinct profiles. \u003cem\u003eNeuropsychol Rev\u003c/em\u003e. 2024;34(2):312-330. doi:10.1007/s11065-023-09618-w\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-neuroscience","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ndev","sideBox":"Learn more about [Neural Development](http://neuraldevelopment.biomedcentral.com/)","snPcode":"13064","submissionUrl":"https://submission.nature.com/new-submission/13064/3","title":"Discover Neuroscience","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Alpha Variability, Neural Complexity, Multiscale Entropy, Detrended Fluctuation Analysis, Cognitive Performance, Electroencephalography, Medical Students, Reaction Time Variability, Critical Dynamics","lastPublishedDoi":"10.21203/rs.3.rs-9180547/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9180547/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Traditional electroencephalography (EEG) analysis focuses on mean spectral power, which may overlook the functional significance of neural signal variability. Emerging perspectives posit that moment-to-moment neural variability—captured through complexity metrics and temporal dynamics—is a key marker of adaptive brain function and cognitive efficiency.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObjective:\u003c/strong\u003e This study aimed to determine whether multiple indices of functional brain dynamics in the alpha band (8–13 Hz) predict cognitive performance consistency independently of mean alpha power in healthy medical students, using both traditional variability metrics and contemporary complexity measures.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e In a cross-sectional design, 52 participants (mean age 23.1 years; 50% female) underwent resting and task EEG recording. Alpha power and its trial-to-trial variability (coefficient of variation, CoV) were computed. Additionally, multiscale permutation entropy (MPE) and detrended fluctuation analysis (DFA) were applied to quantify signal complexity and long-range temporal correlations. Cognitive performance was assessed via a reaction time task, with intra-individual variability (RT SD) as the primary outcome. Stress was measured using a Visual Analogue Scale and physiological reactivity (heart rate change) during a Mental Arithmetic Test. Relationships were examined using correlation, hierarchical regression, and multi-feature prediction models incorporating quadratic (nonlinear) effects.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Alpha variability (CoV) was significantly correlated with RT SD (r = 0.42, p = 0.001) and mean RT (r = 0.27, p = 0.049). Multiscale entropy in the alpha band showed a significant inverted-U relationship with RT variability (R² = 0.21, p = 0.003 for quadratic term), indicating that moderate complexity was associated with greatest performance stability. DFA exponents correlated negatively with RT variability (r = -0.29, p = 0.038), suggesting that stronger long-range temporal correlations (closer to critical dynamics) relate to more consistent performance. Perceived stress and stress reactivity also correlated with RT SD (r = 0.29, p = 0.034 and r = 0.32, p = 0.022, respectively). Hierarchical regression confirmed alpha variability as a unique predictor of RT variability (β = 0.42, p = 0.001), accounting for 17% additional variance after controlling for mean alpha power, which was non-significant. A combined multi-feature model including CoV, quadratic MPE, and DFA explained 28% of variance in RT variability—substantially more than any single metric alone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e Variability, complexity, and temporal correlations of alpha oscillations—not mean power—are significant neural correlates of performance stability. Multi-feature approaches incorporating dynamical metrics provide richer characterization of brain-behavior relationships, supporting the growing emphasis on neural dynamics in cognitive neuroscience. The inverted-U relationship between complexity and performance suggests an optimal range for cognitive stability, with deviations in either direction conferring risk for attentional inconsistency.\u003c/p\u003e","manuscriptTitle":"Functional Brain Variability Predicts Cognitive Performance Independent of Mean EEG Power","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-24 12:43:07","doi":"10.21203/rs.3.rs-9180547/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-22T07:52:37+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"234341814258926076649739358159202890507","date":"2026-04-22T01:31:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"105467766634914840043986633367199474960","date":"2026-04-21T14:33:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"40721087345369688309674198302192911187","date":"2026-04-17T06:07:50+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-16T14:10:46+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-07T09:39:19+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-30T06:07:48+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-30T06:07:38+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Neuroscience","date":"2026-03-20T15:46:03+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-neuroscience","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ndev","sideBox":"Learn more about [Neural Development](http://neuraldevelopment.biomedcentral.com/)","snPcode":"13064","submissionUrl":"https://submission.nature.com/new-submission/13064/3","title":"Discover Neuroscience","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"7955ed41-3d6c-4242-9151-08b0092f3871","owner":[],"postedDate":"April 24th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-24T12:43:08+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-24 12:43:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9180547","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9180547","identity":"rs-9180547","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}