Texture analysis is crucial for understanding images by extracting features that define spatial patterns. Recently, bi-dimensional extensions of entropy measures have gained attention due to their simplicity and strong theoretical foundations. However, existing methods primarily operate in the spatial domain and thus overlook frequency-domain and multiscale information. To address this, we introduce bidimensional wavelet increment entropy (wavelet IncrEn(2D)). A one-level discrete wavelet transform (DWT) with the Haar wavelet decomposes each image into approximation (low-frequency) and, for some neuroimaging data, detail (high-frequency) subbands; IncrEn(2D) is then applied both to capture global structural patterns and fine, detailed texture variations. We evaluated wavelet IncrEn(2D) on synthetic and real datasets, demonstrating its effectiveness in distinguishing between different noise types (white Gaussian, salt-and-pepper, and speckle noise). Comparisons between periodic and synthesized images revealed lower wavelet IncrEn(2D) values for periodic textures. Tests on real texture datasets highlight the method's ability to differentiate various patterns. In particular, wavelet IncrEn(2D) achieved 86.69% accuracy in distinguishing MRI images of healthy versus multiple sclerosis-affected brains. Overall, wavelet IncrEn(2D) offers a robust, frequency-aware descriptor that outperforms existing 2D entropy methods.

Multiscale diversity entropy (MDivE) quantifies nonlinear signal complexity by constructing pattern probability distributions across multiple scales using cosine similarity. However, its coarse-graining mainly relies on a mean operator, which may over smooth bearing fault signals and consequently attenuate critical impulsive and transient components. Moreover, MDivE estimates entropy from pattern occurrence statistics at each scale and does not explicitly account for temporal transitions among pattern states, which limits its ability to capture fault induced dynamic evolution. To address these limitations, this paper proposes a high- order multiscale symbolic transfer diversity entropy (HMSTDE). HMSTDE incorporates time-shifted segmentation and a high-order amplitude difference operator to strengthen the dynamic response of fault impulses while preserving multiscale compression. In addition, symbolic transition analysis is implemented by constructing a state transition matrix and computing entropy over the transition distribution, thereby embedding temporal ordering and directional information into the feature representation. Simulation studies and experiments on multiple aero-engine bearing datasets demonstrate that HMSTDE provides improved noise robustness and fault class separability compared with MDivE and other widely used entropy measures. These results indicate that merging enhanced multiscale coarse-graining with transition modeling yields an effective complexity indicator for aero-engine bearing fault feature extraction.

Accurate clinical diagnosis for Alzheimer's disease dementia (AD), amnestic mild cognitive impairment (aMCI), and non-amnestic MCI (naMCI) is essential for timely management. The diagnosis is made using a range of factors including cognitive testing. Explainable artificial intelligence (XAI)-based SHAP (SHapley Additive exPlanations) is a machine learning interpretability tool that can provide insights into specific features that drive classification decisions. We used XAI with support vector machines (SVM) to identify key cognitive features of Toronto Cognitive Assessment (TorCA), a user-friendly cognitive assessment administered by frontline clinicians, for differentiating neurocognitive disorder diagnoses. We used data from the Toronto Dementia Research Alliance (TDRA) database, comprising of participants with AD, aMCI, naMCI, or normal cognition (NC) seen in memory clinics across Toronto. An SVM model with radial basis function (RBF) kernel was configured with 10-fold cross-validation. XAI was integrated using SHAP values to identify the most important critical features contributing to the model predictions. Classification accuracies, defined as the proportion of correct classifications for each pairwise comparison, were calculated using TorCA total scores and specific features from subtests. We included 695 participants (149 AD, 189 aMCI, 304 naMCI, and 53 NC). Classification accuracy for distinguishing AD vs NC was excellent, whether using all TorCA subtests (0.97±0.03), or the top 5 features (0.96±0.04) (Delayed Recall, Immediate Recall Trials 1 and 2, Sentence Comprehension, Benson Figure Recall), but lower (0.93±0.06) with only TorCA total score. Classification accuracy for MCI or naMCI vs NC was also very good (0.86±0.03 to 0.89±0.04) using the top 5 features. TorCA combined with XAI can accurately differentiate common clinical neurocognitive disorder phenotypes in ambulatory settings. These findings point to specific cognitive subtests important for diagnosis and may help improve the efficiency of cognitive testing. Future studies should investigate the differentiation of other neurocognitive disorders using these tools and further validate these findings using formal neuropsychological testing.

EEG microstates (brief, stable scalp topographies) reveal dynamic brain network activity through their transi tions. However, existing analyses typically examine single spatial scales and a limited set of features, potentially missing crucial aspects of network complexity. Here, we introduce a multiscale framework that captures both finegrained local activity and coarse-scale global network organization, and we define three novel metrics: Segregation (network specialization via spatial consistency), Integration (global coordination via transition unpredictability), and Complexity (their balance, reflecting the richness of the dynamic repertoire). We validated this framework on resting-state EEG from two cohorts: 188 healthy adults (young: 20-35 years; old: 60-80 years) and 65 par ticipants with dementia (29 controls, 36 Alzheimer's disease [AD]). Our multiscale analysis revealed distinct aging signatures: healthy aging showed decreased Segregation and Complexity with increased Integration, sug gesting compensatory network reorganization. In contrast, AD exhibited reduced Segregation and Integration but paradoxically elevated Complexity, indicating pathological network disorganization with erratic state transitions. Coarse-scale metrics outperformed fine-scale analyses, achieving robust group discrimination (AUC approximate to 0.76 for age groups; AUC approximate to 0.86 for AD detection). Both Integration and Complexity significantly correlated with cog nitive performance, establishing them as interpretable biomarkers. These findings demonstrate that multiscale microstate analysis captures the balance between network specialization and global coordination, distinguishing compensatory changes in healthy aging from pathological disorganization in AD. As the first study to opera tionalize multiscale segregation and integration in microstate dynamics, this framework establishes a versatile foundation applicable to diverse biomedical signals (e.g., EEG, fMRI) and opens new avenues for developing novel metrics of spatiotemporal network organization.

Alpha-band default mode network (DMN) connectivity declines with aging and Alzheimer's disease (AD), yet most electroencephalography (EEG) connectivity studies used pairwise (two-order) measures, such as mutual information rate (MIR). We leveraged O-information rate (OIR) to quantify three-order interactions and to separate redundant from synergistic information processing across frontal, temporal, and parietal DMN regions. We hypothesized that, extending established findings of reduced pairwise connectivity, (i) OIR (and its components) would be reduced in older versus younger adults and in AD versus healthy controls (HC); (ii) combining MIR with OIR would improve classification compared with MIR alone; and (iii) OIR measures would correlate positively with global cognition (as assessed by the Montreal Cognitive Assessment (MoCA)). Resting-state EEG from two samples—healthy adult lifespan aging (95 younger; 93 older) and AD spectrum (44 HC; 84 amnestic mild cognitive impairment [aMCI]; 41 AD)—was source-localized using eLORETA to DMN regions. Alpha band (8–13 Hz) MIR and OIR were computed through multivariate spectral analysis. Group differences were tested using t-tests or analysis of covariance (ANCOVA) with multiple comparison correction. Classification (OIR, MIR, demographic, and combined feature sets) used cross-validated logistic-regression, linear-SVM, and random-forest models, with bootstrap 95

Recent trends in nonlinear two-dimensional entropy methods are revolutionizing the quantification of image irregularity. Traditional measures—such as two-dimensional sample entropy (SamEn2D), two-dimensional dispersion entropy (DisEn2D), and two-dimensional fuzzy entropy (FuzEn2D)—effectively characterize textures and spatial dependencies but often suffer from high sensitivity to parameter settings and noise. To address this gap, we propose a novel approach that integrates fuzzy membership functions into the dispersion entropy framework, yielding enhanced robustness and stability in the assessment of image irregularity. The proposed Two-Dimensional Fuzzy Dispersion Entropy (FuzDisEn2D) method is based on Shannon entropy and incorporates a fuzzy mapping strategy for smoother quantization. Unlike the traditional quantization in DisEn2D—where a hard quantization operator assigns each pixel to a single discrete class and can increase sensitivity to noise and parameter settings—FuzDisEn2D replaces hard quantization with fuzzy membership-based quantization, enabling pixels near class boundaries to contribute smoothly to adjacent classes. This soft assignment reduces information loss, mitigates boundary ambiguity, and improves robustness.Evaluations on synthetic datasets—including MIX2D images, colored noise, and composite textures—demonstrate that FuzDisEn2D captures a range from purely periodic to fully random structures, detects the presence of frequency-dependent irregularities, and distinguishes heterogeneous and complex patterns, respectively, thereby surpassing existing techniques in terms of stability, noise resilience, and discriminatory power. Similarly, evaluations on diverse real-world datasets—including industrial screw inspection, granite texture classification, lung computed tomography (CT) analysis, and colorectal cancer histology—employing an optimized Support Vector Machine (SVM) framework, highlight the method’s broad practical relevance in industrial quality control and biomedical diagnostic imaging. The results confirm the superiority of the proposed method.

Alzheimer's disease (AD) profoundly affects motor control and cognitive functions, often resulting in impaired speech characteristics such as vocal clarity, emotional expressiveness, and prosodic richness. To detect such abnormalities, we examine the role of three specific acoustic features to differentiate participants with AD from healthy controls (HC) and study the association between these acoustic features and global cognition. Speech data from 237 participants (115 HC, 110 AD) in the ADReSS-M dataset, collected during the “Cookie Theft” picture description task, were analyzed. This dataset has been matched for age and gender by propensity score to prevent bias. The HC group averaged 66.4 years (SD: 6.64), and the AD group averaged 69.4 years (SD: 6.92), with significantly lower Mini-Mental State Examination (MMSE) scores in AD (AD: 17.9; HC: 29.0). Features quantifying vocal clarity, articulatory precision (spectral contrast), vocal tone (pitch mean), and prosodic variability (pitch standard deviation) were extracted. Group differences in these features were assessed using t-tests, and Pearson correlation analyses were conducted to examine associations between acoustic measures and MMSE scores. There were significant differences between AD and HC groups for spectral contrast (t(235) = 4.26, p  <0.0001), pitch mean (t(235) = 3.54, p  = 0.0005), and pitch standard deviation (t(235) = 3.62, p  = 0.0004). Cohen's d values for these features ranged from -0.5 to -0.6, indicating medium effect sizes, with lower values observed in the AD group. We also found a significant correlation ( p  <0.01) between MMSE scores and each of the features (Pearson's r = 0.22 for pitch mean, 0.23 for pitch standard deviation, and 0.25 for spectral contrast). This preliminary study highlights the physiological basis of altered speech patterns in AD and their diagnostic relevance. Future work will focus on refining preprocessing algorithms and incorporating advanced feature extraction methods to enhance the effect sizes and correlation for AD detection and cognitive assessment.

Mild cognitive impairment (MCI) is an early stage of non-age-related cognitive decline with an increased risk of progressing to dementia. Early detection of MCI is essential for implementing preventative strategies that can delay or prevent the onset of dementia, ultimately improving patient outcomes and reducing healthcare costs. electroencephalograms (EEGs) and event-related potentials (ERPs) have shown significant promise in detecting MCI due to their affordability, real-time monitoring capabilities, and non-invasiveness. EEG provides continuous brain activity data, while ERPs offer insights into specific cognitive processes by analyzing brain responses to stimuli. These methods can complement each other in MCI diagnosis by providing a comprehensive view of overall brain function and detailed information on specific cognitive processes. However, EEG and ERP are susceptible to noise and inter-individual variability, which can hinder their reliability. Additionally, applying machine learning models on EEG or ERP for MCI detection presents challenges such as the risk of overfitting and difficulties in interpreting the underlying decision-making process. This review emphasizes recent advancements in signal processing and feature extraction methods applied to EEG and ERP data and explores the use of machine learning and deep learning techniques to enhance diagnostic accuracy and interpretative depth. By integrating these methodologies, the review highlights how EEG and ERP can contribute to a more effective understanding and monitoring of cognitive changes associated with MCI, underscoring the importance of early diagnosis for timely intervention and improved patient care. Finally, the review focuses on future research directions, including the development of advanced analytical techniques and multimodal integration approaches involving EEG and ERP to further improve diagnostic accuracy and clinical application.

Background Multiscale dispersion entropy (MDEnt) is a nonlinear EEG measure that quantifies brain complexity across time scales, reflecting both local and global brain dynamics. Previous research indicates lower complexity at short time scales in Alzheimer's disease (AD) compared to mild cognitive impairment (MCI) and healthy controls (HCs), with MCI also showing lower values than HCs. Major depressive disorder (MDD) has also been preliminarily linked to reduced complexity during acute episodes.Objective To assess whether MDEnt at short time scales can distinguish AD from MCI and HCs, and to examine complexity differences across additional groups, remitted MDD (rMDD) and rMDD + MCI, while exploring associations with cognitive performance.Methods The study included 316 older adults: 44 HCs, 46 with rMDD, 114 with MCI, 71 with rMDD + MCI, and 41 with AD. Resting-state, eyes-closed EEGs were analyzed using MDEnt at 24 ms (short) and 60 ms (long) time scales. Cognitive function was measured with the Montreal Cognitive Assessment and a composite cognitive score.Results Short time scale complexity was lowest in AD, followed by MCI, and highest in HCs; rMDD presence had no impact. Only AD showed reduced complexity at long time scales. Complexity at both time scales was significantly correlated with cognitive performance.Conclusions This study highlights the value of MDEnt to assess complexity at short time scale and differentiate individuals with AD, MCI, or HCs. Reduced complexity in these individuals may underlie their cognitive impairment. In contrast, our study suggests that any MDD impact on complexity is likely related to active depressive symptoms.

Previous literature has identified slowing of resting state electroencephalography (EEG) rhythm and abnormal cortical excitation in Alzheimer’s Dementia (AD). However, the relationship between these two divergent functional abnormalities and cognitive symptoms of AD are not well understood. Resting state EEG signal was recorded in participants with AD and HCs for 5 minutes with eyes closed. Relative resting state EEG power was measured for the five frequency bands. Participants underwent a single pulse transcranial magnetic stimulation (TMS) combined with EEG. Cortical evoked activity (CEA) was assessed using TMS-evoked potential (TEP) rectified area under the curve (AUC) from 25 to 80 ms post-TMS stimulus, and the TEPs peak amplitudes were calculated by taking the maximal peak in the following time windows: 25 – 35 ms (P30), 40 – 50 ms (N45), and 55 – 65 ms (P60). Compared to 32 HC (18 females; mean ± SD age: 69.3 ± 7.9 years), 52 participants with AD (32 females; 74.2 ± 8.5 years) had higher relative theta power than HCs (t(66.6) = 5.34, p<0.001). AD participants also had a significantly lower alpha power (t(76) = -3.03, p=0.003) and beta power (t(76) = -2.51, p=0.014) than HCs. Controlling for sex, age and years of education, AD participants showed a positive association between theta power and CEA (r partial =0.573, p=0.008); and an inverse association between alpha power and CEA (r partial =-0.471, p=0.036). Theta power in AD participants showed a positive association with P60 (r partial =-0.637, p=0.003) and an inverse association with N45 (r partial =-0.512, p=0.025), while alpha power was inversely associated with P60 (r partial =-0.531, p=0.016). This study showed a positive association between resting state EEG slowing and increased cortical excitability in AD, indicating a possible shared mechanistic pathway between these abnormalities.

Multiscale dispersion entropy (MDE) is a nonlinear approach for assessing the complexity of brain activity using electroencephalograms (EEGs). MDE captures EEG dynamics across biologically relevant time scales, with short‐scales reflecting high‐frequency oscillations and local neuronal activity, and long‐scales representing low‐frequency oscillations and large‐scale network processes. Previous studies suggest that patients with Alzheimer's dementia (AD) have decreased complexity at short time scales compared to those with mild cognitive impairment (MCI) or healthy controls (HCs), and individuals with MCI show reduced complexity compared to HCs. There is also preliminary evidence suggesting that adult patients with acute depression –a high‐risk condition for AD– have decreased complexity at a short time scale. Thus, we conducted a study in older participants with AD, MCI, HC, remitted major depressive disorder (rMDD), or rMDD+MCI, hypothesizing reduced short‐scale MDE in AD vs. MCI and MCI vs. HC. We also explored MDE at short and long time scales across all diagnostic groups and their relationships with cognitive performance. The study included 44 HC, 46 rMDD, 114 MCI, 71 rMDD+MCI, and 41 AD participants. MDE was generated using resting‐state EEG with 24ms as the short time scale and 60ms as the long time scale. Cognition was assessed using the Montreal Cognitive Assessment and a cognitive composite score from a comprehensive neuropsychological battery. MDE at 24ms was decreased in AD vs. MCI and in MCI vs. HCs. rMDD had no impact. At 60ms, only the AD group differed from the other groups. Cognitive performance was associated with MDE at 24ms but not 60ms. This study highlights the value of MDE at a short time scale, related to local neuronal activity, to separate individuals with AD vs. MCI vs. HCs. Reduced complexity in these individuals may underlie their cognitive impairment. In contrast, our study suggests that any MDD impact on complexity is likely related to active depressive symptoms.

Good sleep quality is essential for both physiological and mental health. It helps in clearing TAU and beta-amyloid aggregates and consolidating memory, key processes in delaying dementia. Poor sleep is linked to reduced cognitive flexibility in daily life, likely due to decreased brain complexity, reflecting a reduced range of adaptive spatiotemporal brain dynamics. This study introduces a novel approach using non-linear EEG analysis focused on low conventional bands to classify sleep quality in individuals with mild cognitive impairment (MCI), based on brain complexity. Resting-state EEG was collected from 22 participants with MCI aged 60+, grouped by sleep quality (Pittsburgh Sleep Quality Index): 11 MCI with good sleep, and 11 MCI with poor sleep (Table 1). EEG data (128 channels, 5-minute recordings) were normalized and decomposed using the Discrete Wavelet Transform to reach delta (1–4 Hz) and theta (4–8 Hz) bands. Ten non-linear complexity features, namely approximate entropy, correlation dimension, detrended fluctuation analysis, energy, Higuchi fractal dimension, Hurst exponent, Katz fractal dimension, Boltzmann Gibbs entropy, Lyapunov exponent and Shannon entropy, were extracted from 5 second segments. Statistical measures (mean, standard deviation, 95th percentile, variance, median, kurtosis) were computed from these time-distribution features. These statistics were then used for training and testing a set of classic machine learning classifiers, employing leave-one-out cross-validation (Figure 2). Brain complexity successfully classified sleep quality in MCI, achieving an accuracy and area under the curve (AUC) of 1 in channel D13 (delta subband) using Quadratic Discriminant Analysis (QDA), and an accuracy of 0.94 and an AUC of 0.95 in channel B17 (theta subband) using the Extra Trees Classifier (ETC) (Figure 3). Specific machine learning classifiers distinguish excellently sleep quality in MCI using spatiotemporal complexity features from slow EEG subbands. The most relevant channels for group discrimination were primarily located in bilateral temporal regions of the neocortex known to be among the first affected in amnestic MCI, as previously shown in neuroimaging studies. Future longitudinal studies could investigate whether changes in brain complexity within these slow-frequency temporal regions, influenced by sleep quality, are associated with an earlier or faster onset of dementia.

Entropy algorithms are widely applied in signal analysis to quantify the irregularity of data. In the realm of two-dimensional data, their two-dimensional forms play a crucial role in analyzing images. Previous works have demonstrated the effectiveness of one-dimensional increment entropy in detecting abrupt changes in signals. Leveraging these advantages, we introduce a novel concept, two-dimensional increment entropy (IncrEn2D), tailored for analyzing image textures. In our proposed method, increments are translated into two-letter words, encoding both the size (magnitude) and direction (sign) of the increments calculated from an image. We validate the effectiveness of this new entropy measure by applying it to MIX2D(p) processes and synthetic textures. Experimental validation spans diverse datasets, including the Kylberg dataset for real textures and medical images featuring colon cancer characteristics. To further validate our results, we employ a support vector machine model, utilizing multiscale entropy values as feature inputs. A comparative analysis with well-known bidimensional sample entropy (SampEn2D) and bidimensional dispersion entropy (DispEn2D) reveals that IncrEn2D achieves an average classification accuracy surpassing that of other methods. In summary, IncrEn2D emerges as an innovative and potent tool for image analysis and texture characterization, offering superior performance compared to existing bidimensional entropy measures.

Aging and poor sleep quality are associated with altered brain dynamics, yet current electroencephalography (EEG) analyses often overlook regional complexity. This study addresses this gap by introducing a novel integration of intra- and inter-regional complexity analysis using multivariate multiscale dispersion entropy (mvMDE) from awake resting-state EEG for the first time. Moreover, assessing both intra- and inter-regional complexity provides a comprehensive perspective on the dynamic interplay between localized neural activity and its coordination across brain regions, which is essential for understanding the neural substrates of aging and sleep quality. Data from 58 participants—24 young adults (mean age = 24.7 ± 3.4) and 34 older adults (mean age = 72.9 ± 4.2)—were analyzed, with each age group further divided based on Pittsburgh Sleep Quality Index (PSQI) scores. To capture inter-regional complexity, mvMDE was applied to the most informative group of sensors, with one sensor selected from each brain region using four methods: highest average correlation, highest entropy, highest mutual information, and highest principal component loading. This targeted approach reduced computational cost and enhanced the effect sizes (ESs), particularly at large scale factors (e.g., 25) linked to delta-band activity, with the PCA-based method achieving the highest ESs (1.043 for sleep quality in older adults). Overall, we expect that both inter- and intra-regional complexity will play a pivotal role in elucidating neural mechanisms as captured by various physiological data modalities—such as EEG, magnetoencephalography, and magnetic resonance imaging—thereby offering promising insights for a range of biomedical applications.

Background: Aging, frontotemporal dementia (FTD), and Alzheimer's dementia (AD) manifest electroencephalography (EEG) alterations, particularly in the beta-to-theta power ratio derived from linear power spectral density (PSD). Given the brain's nonlinear nature, the EEG nonlinear features could provide valuable physiological indicators of aging and cognitive impairment. Multiscale dispersion entropy (MDE) serves as a sensitive nonlinear metric for assessing the information content in EEGs across biologically relevant time scales. Objective: To compare the MDE-derived beta-to-theta entropy ratio with its PSD-based counterpart to detect differences between healthy young and elderly individuals and between different dementia subtypes. Methods: Scalp EEG recordings were obtained from two datasets: 1) Aging dataset: 133 healthy young and 65 healthy older adult individuals; and 2) Dementia dataset: 29 age-matched healthy controls (HC), 23 FTD, and 36 AD participants. The beta-to-theta ratios based on MDE vs. PSD were analyzed for both datasets. Finally, the relationships between cognitive performance and the beta-to-theta ratios were explored in HC, FTD, and AD. Results: In the Aging dataset, older adults had significantly higher beta-to-theta entropy ratios than young individuals. In the Dementia dataset, this ratio outperformed the beta-to-theta PSD approach in distinguishing between HC, FTD, and AD. The AD participants had a significantly lower beta-to-theta entropy ratio than FTD, especially in the temporal region, unlike its corresponding PSD-based ratio. The beta-to-theta entropy ratio correlated significantly with cognitive performance. Conclusion: Our study introduces the beta-to-theta entropy ratio using nonlinear MDE for EEG analysis, highlighting its potential as a sensitive biomarker for aging and cognitive impairment.

Background and Objective: We present NLDyn, an open-source MATLAB toolbox tailored for in-depth analysis of nonlinear dynamics in biomedical signals. Our objective is to offer a user-friendly yet comprehensive platform for researchers to explore the intricacies of time series data.

Electroencephalography (EEG) is useful for studying brain activity in major depressive disorder (MDD), particularly focusing on theta and alpha frequency bands via power spectral density (PSD). However, PSD-based analysis has often produced inconsistent results due to difficulties in distinguishing between periodic and aperiodic components of EEG signals. We analyzed EEG data from 114 young adults, including 74 healthy controls (HCs) and 40 MDD patients, assessing periodic and aperiodic components alongside conventional PSD at both source and electrode levels. Machine learning algorithms classified MDD versus HC based on these features. Sensor-level analysis showed stronger Hedge’s g effect sizes for parietal theta and frontal alpha activity than source-level analysis. MDD individuals exhibited reduced theta and alpha activity relative to HC. Logistic regression-based classifications showed that periodic components slightly outperformed PSD, with the best results achieved by combining periodic and aperiodic features (AUC = 0.82). Strong negative correlations were found between reduced periodic parietal theta and frontal alpha activities and higher scores on the Beck Depression Inventory, particularly for the anhedonia subscale. This study emphasizes the superiority of sensor-level over source-level analysis for detecting MDD-related changes and highlights the value of incorporating both periodic and aperiodic components for a more refined understanding of depressive disorders.

This paper proposes a new approach to multiscale entropy (MSE) analysis using max-pooling for coarse-graining, inspired by convolutional neural networks, to efficiently capture important features while reducing data dimensionality. The proposed method, maxMSE, is compared with traditional multiscale entropy techniques using synthetic white and $1/f$ noise, as well as resting-state EEG data for healthy young versus elderly individuals. The results demonstrate that maxMSE shows improved stability for short time series, as evidenced by a lower mean and standard deviation compared to the traditional MSE algorithm when tested on white and pink noise. maxMSE also offers competitive computational efficiency. The method's application to EEG data yields statistically significant results in distinguishing age groups, with maxMSE achieving a lower p-value $(p=0.0217)$ and a higher effect size (0.3270) compared to the traditional MSE measure with $p= 0.1113$ and undefined effect size. maxMSE complements conventional multiscale entropy by using max-pooling for coarse-graining, which retains the maximum values of data segments, whereas traditional MSE relies on moving averages.