The research team recruited 49 university students aged 18 to 25, 48 of whom were included in the final analysis after one was excluded for incomplete questionnaire data. Each participant filled out standard mental health instruments—the Patient Health Questionnaire-9 (PHQ-9), the seven-item Generalized Anxiety Disorder scale (GAD-7), the Perceived Stress Scale, and the Multidimensional Fatigue Scale. A licensed clinical psychologist, blinded to the EEG data, then classified each participant as experiencing mental fatigue or not, based on a validated fatigue questionnaire and a predefined cutoff score rather than subjective judgment alone. The team treated “mental fatigue” as an umbrella construct encompassing self-reported symptoms of anxiety, depression, and stress.

Each participant then underwent two five-minute EEG recording sessions with a 16-channel dry-electrode cap sampled at 1,200 Hz: one in silence and one while listening to an ambient track, with identical relaxation instructions before each session. The raw signals were band-pass filtered between 1 and 45 Hz, power-line noise was removed with a 50 Hz notch filter, and artifacts—including eye blinks and muscle contamination—were stripped out using independent component analysis. The cleaned recordings were chopped into 10-second epochs, and from each epoch the researchers extracted five complementary families of features: power spectral density across the classic delta, theta, alpha, beta, and gamma bands; integrated band powers; the alpha-theta ratio; the theta-beta ratio; and coherence, a measure of phase synchronization capturing functional connectivity between all pairs of the 16 electrodes. Combined, these produced feature vectors of roughly 19,500 dimensions per epoch.

To classify these patterns, the team built a deep residual multi-layer perceptron—a feedforward neural network with residual skip connections that ease gradient flow through three residual blocks of 512 neurons followed by downsampling layers of 256 and 128 units, with dropout regularization of 0.3. The same architecture was adapted for a multi-output task, adding three parallel sigmoid output heads to simultaneously predict anxiety, depression, and stress, each trained with a weighted binary cross-entropy loss. Crucially, the researchers evaluated the model with Leave-Subject-Out cross-validation, holding each of the 48 participants out of training exactly once, and they aggregated epoch predictions into a single subject-level probability before thresholding—a design choice meant to ensure the reported metrics reflect real-world performance on whole individuals rather than flattering epoch-level statistics.

The results revealed a striking gap between the two recording conditions. When participants listened to music, the model detected mental fatigue with 72.92% accuracy, a recall of 96.67%, an F1 score of 81.69%, and an area under the curve of 69.63%. Without music, accuracy fell to 58.33% with an AUC of 53.15%—barely better than a coin flip. In other words, the very features that failed to separate fatigued from non-fatigued brains in silence became substantially more informative under music stimulation. The high recall in the music condition means the model caught nearly everyone who was fatigued, though at the cost of false positives, a trade-off the authors note may be acceptable in screening contexts where missing true cases is more costly than ordering further evaluation.

The spectral analysis offers clues about why music helps. Without music, participants without mental fatigue showed higher and more stable spectral power—particularly in the alpha band—while fatigued participants displayed elevated frontal theta activity, higher frontal beta power, and greater signal variability. During music listening, theta power decreased in the fatigued group, alpha activity became more prominent in posterior regions for both groups, and the beta-band differences between groups shrank, hinting at a possible normalizing effect of music on cortical arousal. Statistical tests at frontal channels such as Fp1, Fp2, F3, Fz, and F4 showed the strongest group differences, and after Bonferroni correction three channels—T7, P3, and P4—lost significance only in the music condition, consistent with the idea that music normalizes neural activity in temporal and parietal regions.

Interpretability analysis using SHAP (SHapley Additive exPlanations) values added further texture. In the music condition, alpha-theta ratio features from frontal electrodes were among the most important drivers of the model’s decisions, alongside theta-beta ratios from the central electrode Cz, suggesting that the balance between relaxation-associated alpha rhythms and drowsiness-associated theta rhythms becomes especially informative when music is playing. In silence, the model leaned more heavily on raw band-power features, and its overall discrimination was weaker—implying that music organizes brain activity in a way that amplifies fatigue-related neural differences. The deep network also outperformed classical baselines, including Random Forest, linear support vector machines, and logistic regression, in the music condition; notably, the linear models collapsed to near-zero specificity, suggesting the decision boundary between fatigue states is highly non-linear.

The multi-output extension told a more sobering story. Stress appeared to classify almost perfectly at 93.75% accuracy with 100% recall—but the authors are explicit that this figure is an artifact of class imbalance, since 85.7% of participants fell into the stressed category and the balanced accuracy of exactly 50% equals random chance. The AUC of 35.56% confirmed the absence of any genuine stress-specific neural signature. Anxiety (54.17% accuracy) and depression (62.50% accuracy) proved genuinely difficult, as their class distributions were more balanced. These results underscore a recurring hazard in EEG-based mental health machine learning: headline accuracy figures can mask models that simply echo the prevalence of the majority class.

The authors are candid about their limitations. The sample was small and drawn from a single university; the condition order was fixed—silence first, then music—so order, habituation, or time-on-task effects cannot be ruled out as alternative explanations for the performance gap; only one music track was used; recordings were just five minutes per condition; dry electrodes are vulnerable to electromyographic contamination in the gamma band; and labels derived from questionnaires with a cutoff do not constitute formal psychiatric diagnosis. The exploratory epoch-pooled significance tests, the authors caution, are anti-conservative and require confirmation with properly powered subject-level designs, ideally randomized or counterbalanced.

Even with these caveats, the study represents a meaningful step toward objective, AI-driven mental health screening. The central message—that recording context matters, and that music-evoked EEG may expose neural biomarkers invisible in resting-state recordings—could influence how future brain-based assessments are designed, from protocol choices to the selection of relational spectral features. The authors envision follow-up work with larger and more diverse cohorts, personalized music selection, complementary physiological signals such as heart rate variability, and source-level EEG analysis to localize the neural networks most affected by fatigue. Until then, the results remain preliminary and are not intended for clinical use, but they quietly raise an appealing possibility: that something as ordinary as listening to music might help unlock the brain’s most telling signals.

Subject of Research: Music-evoked EEG signals and deep learning classification of mental fatigue, anxiety, depression, and stress for mental health screening

Article Title: Music-evoked EEG signals for mental health assessment using two- and three-class classification

Article References: Belkacem, A. N., Hireche, A., Faris Ali, N., Atef, M., Hamid, A. A. R. M., Elfadel, I. M., & Jelinek, H. F. (2026). Music-evoked EEG signals for mental health assessment using two- and three-class classification. Frontiers in Systems Neuroscience, 20, Article 1886805. https://doi.org/10.3389/fnsys.2026.1886805

Image Credits: AI Generated

DOI: 10.3389/fnsys.2026.1886805

Keywords: EEG, mental fatigue, music therapy, deep learning, residual MLP, depression, anxiety, stress, machine learning, mental health screening, neuroscience, power spectral density

The research, carried out by Andrés Asensio Ramos of the Instituto de Astrofísica de Canarias and the Universidad de La Laguna in Tenerife, addresses a long-standing weakness in the traditional approach. In conventional MOMFBD, the algorithm performs what statisticians call a joint maximum likelihood estimation: it simultaneously estimates both the unknown object being imaged and the unknown atmospheric aberrations that distorted each frame. This joint strategy has powered pipelines at major facilities such as the Swedish 1-meter Solar Telescope and the GREGOR solar telescope for years, but it carries inherent instabilities that can silently degrade scientific results.

The heart of the problem lies in how the algorithm allocates power between signal and noise. When object and aberrations are estimated together, the optimizer can mistakenly attribute noise fluctuations in the data to high-order wavefront aberrations, fabricating atmospheric complexity that was never there. It is a bit like trying to transcribe a whispered conversation in a loud café while simultaneously modeling the background chatter: with no principled separation, the transcript can absorb café noise as words. For solar physics, where quantitative measurements of contrast and intensity feed directly into inferences about magnetic fields and plasma dynamics, such errors matter enormously.

Asensio Ramos’s solution reframes the estimation problem through a Bayesian lens. Rather than treating the observed object as an unknown to be pinned down exactly, the new approach marginalizes over it, integrating out the object’s uncertainty instead of estimating it jointly. This marginal estimator, built on ideas previously explored in adaptive optics for retinal imaging and astronomical adaptive-optics systems, produces what Bayesian practitioners describe as a form of automatic regularization. By acknowledging that the true object is not perfectly known, the method naturally prevents the reconstruction from funneling noise power into spurious high-order aberrations.

Technically, the procedure models the object with a Gaussian prior characterized by a power spectral density, or PSD, which encodes expectations about how image power is distributed across spatial frequencies. In the joint approach, the hyperparameters governing this prior act as delicate tuning knobs whose values strongly influence the final image contrast; set them slightly wrong and the restored Sun appears either unnaturally smooth or artificially grainy. The marginal formulation changes this sensitivity dramatically. Because the object uncertainty is folded into the likelihood analytically, the resulting estimator is far more forgiving, and those PSD hyperparameters can be optimized directly from the data rather than hand-tuned by an experienced pipeline operator.

This robustness yields what the author calls a plug-and-play deployment. Observatories currently depend on specialists who understand the subtle interplay between prior settings and image quality, and transferring pipelines between instruments or observing conditions often demands painstaking recalibration. With the marginal estimator, the need for that expert manual tuning largely disappears, lowering the barrier for smaller facilities and for scientists who want reproducible reconstructions without encyclopedic knowledge of the algorithm’s interior. In an era when next-generation instruments at the Daniel K. Inouye Solar Telescope are producing torrents of data, anything that reduces human intervention per dataset carries real practical weight.

Perhaps the most striking feature of the contribution is its mathematical minimalism. Implementing the marginal estimator requires adding only a log-determinant term to the traditional merit function that the optimizer minimizes. That single addition encapsulates the entire Bayesian marginalization, penalizing solutions in ways that account for the uncertainty of the object being reconstructed. Existing blind deconvolution pipelines can adopt the improvement with minimal modification, which dramatically shortens the path from theoretical insight to operational practice at working observatories.

To make adoption even easier, the method has been fully integrated into torchmfbd, an open-source package built on the PyTorch deep learning framework that Asensio Ramos and colleagues introduced in 2025 in Astronomy & Astrophysics. The package provides flexible multi-object, multi-frame blind deconvolution with automatic differentiation, allowing modern gradient-based optimizers such as Adam to drive the reconstruction. Scripts for reproducing the paper’s results are publicly available on GitHub, putting the technique within reach of the entire solar physics community, from large institutional pipelines to individual researchers analyzing archival data.

The significance extends beyond a single algorithmic tweak. High-resolution ground-based solar imaging underpins much of modern solar physics: measurements of penumbral fine structure, chromospheric magnetic concentrations, and the small-scale dynamo all depend on image restoration pipelines that faithfully preserve photometric contrast. Synthetic observations generated by advanced magnetohydrodynamic codes, such as MANCHA3D, are increasingly compared against restored observations to test our understanding of the solar atmosphere, and any systematic bias introduced during deconvolution propagates directly into those physical inferences. A marginal estimator that controls contrast more reliably therefore strengthens the entire chain from photons on a detector to statements about magnetic fields on the Sun.

The work also illustrates a broader trend in astronomy: the migration of rigorous Bayesian machinery into operational pipelines once dominated by ad hoc regularization. Marginal blind deconvolution has already proven valuable in adaptive-optics imaging of the retina and of night-sky targets, and this study extends the framework to the multi-object, multi-frame setting that solar observations uniquely demand, where wide fields of view must be patched together from many simultaneously recorded channels and frames. By combining accessibility, statistical soundness, and immediate software availability, the marginal approach may well become the default for the next generation of solar image restoration, ensuring that as telescopes grow larger and cameras faster, our pictures of the Sun stay as honest as they are spectacular.

Subject of Research: Bayesian marginal blind deconvolution for high-resolution ground-based solar imaging

Article Title: Marginal Multi-Object Multi-Frame Blind Deconvolution

Article References: Asensio Ramos, A. (2026). Marginal Multi-Object Multi-Frame Blind Deconvolution. Solar Physics, 301(10), Article 155. https://doi.org/10.1007/s11207-026-02752-9

Image Credits: AI Generated

DOI: 10.1007/s11207-026-02752-9

Keywords: solar physics, blind deconvolution, MOMFBD, Bayesian inference, image restoration, atmospheric turbulence, adaptive optics, PyTorch, torchmfbd, power spectral density, regularization, solar telescopes

The study, by Cédric Chavanne of the Institut des sciences de la mer at Université du Québec à Rimouski, revisits a problem that has haunted spectral analysis since the 1960s: how to estimate the power spectral density, or PSD, of a signal sampled at uniform intervals when some of those samples are missing. The PSD describes how the total power of a signal is distributed across frequencies, and its slope in log-log space is the standard diagnostic for turbulence regimes in meteorology, oceanography, and engineering. If the estimated slope is biased, the diagnosis of the physics is wrong, no matter how beautiful the plot looks.

The mathematical backbone of the paper is the Wiener-Khinchin theorem, a cornerstone of signal processing that connects two seemingly different views of a random process. The theorem states that the spectral density of a stationary random process is the Fourier transform of its autocorrelation function, a measure of how strongly the signal at one moment resembles the signal a short time later. For finite, discrete data, Chavanne derives a subtle but consequential refinement: the familiar periodogram estimator is exactly the finite Fourier transform of what he calls the circular unbiased estimator of the autocorrelation, not the standard biased or standard unbiased estimators found in most textbooks. This circular estimator wraps the time series around on itself, treating it as periodic, and it has a crucial property: it is positive semidefinite, which guarantees that its Fourier transform yields only positive spectral values when no data are missing.

Everything changes when samples vanish. Skipping missing samples is mathematically equivalent to replacing them with zeros, which multiplies the underlying signal by a sampling series of ones and zeros. In the frequency domain, that multiplication becomes a convolution, smearing the true spectrum with the spectrum of the sampling pattern. The circular autocorrelation estimator, computed naively, becomes biased because fewer data pairs contribute at each lag. The classical remedy, dating back to work by R. H. Jones in 1962, is to normalize each lag by the actual number of available data pairs, producing an unbiased autocorrelation estimate. But there is a catch that has bothered researchers for decades: this unbiased estimator is no longer guaranteed to be positive semidefinite, so its Fourier transform can dip below zero at some frequencies, producing spectral power estimates that are negative, which is physically nonsensical for any single realization.

That awkward negativity is precisely what a 2021 study by Gao and colleagues tried to eliminate by taking the absolute value of the Fourier transform of the unbiased autocorrelation estimator, forcing every spectral estimate to be positive. The trick was adopted in several subsequent publications, including analyses of satellite wind and wave data, dust storm observations, and even a widely reported claim about hidden turbulence in van Gogh’s The Starry Night. Chavanne’s central message is blunt: the absolute-value fix makes things worse, not better. By folding negative estimates into positive ones, the procedure injects a systematic bias that depends on the number and distribution of the missing samples, and the resulting spectra are more distorted than those from the classical biased estimators the fix was meant to improve. Spectral slopes computed this way, he argues, cannot be trusted for quantitative diagnosis of turbulence regimes.

The counterintuitive resolution is that negative spectral estimates are not a bug to be patched but a necessary feature of unbiasedness. For any single realization of a random process, an unbiased estimator must be allowed to wander below zero occasionally, because only then does the average over many independent realizations converge to the true, always-positive spectrum. Chavanne demonstrates this with synthetic data generated from fractional Brownian motion, a mathematical process whose spectral slope is known exactly. Using a Hurst parameter of one third, which corresponds to the famous minus five-thirds slope of isotropic turbulence in the inertial range, and a Hurst parameter of one half, corresponding to a slope of minus two, he generated ten thousand realizations of 1,024 samples each and then deleted a third of the samples using two different schemes: a Bernoulli mechanism, where samples vanish independently at random, and a batch-Bernoulli mechanism, where samples disappear in consecutive runs of random length, mimicking realistic data outages.

The results were striking. When averaged over ten thousand realizations, the unbiased estimators recovered the theoretical spectral slopes across nearly the entire frequency range for both missing-data schemes, while the standard biased correlogram and the widely used Lomb-Scargle periodogram showed clear distortions, including a leveling of the spectrum at high frequencies and a biased slope for the batch-missing case. The absolute-value estimator of Gao and colleagues was strongly biased under both schemes. Histograms of the estimates at a single frequency told the deeper story: the biased estimators always produced positive values, but their averages sat noticeably away from the true value, whereas the unbiased estimators produced distributions straddling zero, with negative values appearing regularly, yet their averages landed almost exactly on the truth. The price of unbiasedness is variance, and the standard unbiased estimator carried the largest variance of all.

To show the method works on real data, not just synthetic mathematics, Chavanne turned to laboratory measurements of decaying turbulence in an active-grid-generated air flow, a benchmark dataset collected in a return-type Corrsin wind tunnel and originally reported in 2003. The streamwise velocity was sampled at 40 kilohertz, and the fifteen-minute record was divided into 2,195 overlapping segments of 32,768 samples each. He then artificially deleted half the samples in each segment using the same Bernoulli and batch-Bernoulli schemes, a severity beyond which time series are usually considered unusable for conventional spectral analysis. Because the unbiased estimators are noisy with only a few thousand realizations, he applied band-averaging over 25 frequency bins per decade. Even with half the data gone, the unbiased estimators recovered the minus five-thirds spectral slope over the same two-decade frequency range as the complete record, while the absolute-value estimator remained strongly biased and the standard biased estimator failed under the Bernoulli scheme.

Beyond the conceptual correction, the paper offers practical advantages. Chavanne proposes using the circular unbiased autocorrelation estimator rather than the standard unbiased one, because the number of available data pairs at each lag is always greater or equal, which reduces the variance of the resulting spectra. Computing it via the fast Fourier transform also requires only half the memory, since the standard unbiased estimator needs the series zero-padded to separate overlapping contributions. An efficient algorithm, generalized to estimate the cross-spectral density of two gappy signals, is provided in a public Matlab implementation, lowering the barrier for other researchers to adopt the method.

The implications reach well beyond the wind tunnel. Spectral slopes diagnose turbulence regimes in the atmosphere and oceans, calibrate satellite retrievals of wind and waves, and underpin interpretations of scaling in systems as varied as geophysical flows and, controversially, works of art. Wherever the underlying data have gaps, and they almost always do, the temptation to force every spectral estimate positive introduces a hidden bias that scales with the missing fraction. Chavanne’s study, supported by the Natural Sciences and Engineering Research Council of Canada, argues that the community should embrace the ugly negative numbers, average them honestly, and let the truth emerge from the mean. Sometimes, in statistics as in life, the path to a positive answer runs directly through the negatives.

Subject of Research: Unbiased nonparametric estimation of the power spectral density from uniformly spaced data with missing samples

Article Title: Asymptotically-unbiased nonparametric estimation of the power spectral density from uniformly-spaced data with missing samples

Article References: Chavanne, C. (2026). Asymptotically-unbiased nonparametric estimation of the power spectral density from uniformly-spaced data with missing samples. Advances in Statistical Climatology, Meteorology and Oceanography, 12(1), 59-72. https://doi.org/10.5194/ascmo-12-59-2026

Image Credits: AI Generated

DOI: 10.5194/ascmo-12-59-2026

Keywords: power spectral density, missing samples, spectral analysis, turbulence, Wiener-Khinchin theorem, autocorrelation, periodogram, Lomb-Scargle periodogram, fractional Brownian motion, bias, time series, signal processing

The new framework, called EEG-UDAF, was developed by Yufei Chen, Gang Zhou, and colleagues at the State Key Laboratory of Mathematical Engineering and Advanced Computing in Zhengzhou and at Zhengzhou University. Writing in the Journal of Big Data, the team describes an unsupervised domain adaptation system, meaning it can transfer what it has learned from one set of EEG recordings to another without requiring any manually annotated labels in the target data. That distinction matters enormously in practice. In supervised learning, every training example must be tagged with the correct answer, which for EEG means someone must determine what emotion a subject was experiencing at each moment of a recording session. Unsupervised domain adaptation sidesteps that burden by letting a model trained on labeled data from one group of subjects adapt itself to unlabeled data from another.

The first pillar of EEG-UDAF is a dual-view frequency domain feature fusion module. EEG emotion recognition has long relied on frequency features because different emotional states are associated with characteristic patterns of oscillatory brain activity across frequency bands such as theta, alpha, beta, and gamma. Two of the most widely used frequency descriptors are differential entropy, which measures the variance of a signal within a frequency band and captures the intensity of oscillatory activity, and power spectral density, which describes how the signal’s power is distributed across the frequency spectrum. Each view carries complementary information, but most existing systems rely on only one. The researchers extract both and then fuse them using a mutual cross-attention mechanism, a technique borrowed from modern deep learning in which two feature streams interrogate each other to decide which elements deserve emphasis. The result is a richer, more informative frequency representation than either view could provide alone.

Frequency features, however, tell only part of the story. Emotions are also encoded in the spatial arrangement of electrodes across the scalp and in the way signals at different locations evolve over time. The second pillar of the framework is a spatio-temporal feature extraction module built from two complementary architectures. A graph convolutional network treats the EEG electrodes as nodes in a graph, with edges reflecting the brain’s spatial topology, allowing the model to learn how activity at one scalp location relates to activity at its neighbors. A bidirectional long short-term memory network, or BiLSTM, then sweeps through the temporal dimension in both directions, capturing the dynamic characteristics of how EEG channels change over time. Together, these components extract complex spatio-temporal patterns that neither a purely spatial nor a purely temporal model could recover, giving the system a comprehensive view of how emotional states unfold across the brain’s geometry and its history.

The third and arguably most consequential pillar is the unsupervised domain adaptation strategy itself, equipped with a semantic correction mechanism. Domain adaptation addresses a fundamental asymmetry in machine learning: a model trained on data from one distribution, called the source domain, often performs poorly on data from another distribution, the target domain. In EEG research this shows up as the gap between subjects, since each person’s brain produces signals with idiosyncratic amplitude, timing, and spatial patterns, and within subjects over time, since non-stationarity means the same brain does not produce identical signals even in identical conditions. EEG-UDAF aligns the source and target feature distributions so that knowledge acquired from labeled recordings can be applied to unlabeled ones, while the semantic correction mechanism works to preserve the meaning of the features during that alignment, guarding against the well-known failure mode in which distributions are matched but class boundaries are scrambled.

The team evaluated EEG-UDAF on two public EEG emotion recognition datasets, the standard proving grounds for this field. They tested the framework under both intra-subject and inter-subject paradigms. Intra-subject testing asks whether a model can handle shifts within the same person, for example when recordings are taken in different sessions or under different conditions. Inter-subject testing, the harder and more clinically relevant challenge, asks whether a model trained on one group of people can generalize to people it has never seen. The experiments showed that EEG-UDAF can effectively alleviate the non-stationarity and individual differences that have plagued earlier approaches, performing well in both paradigms where conventional models typically degrade sharply.

Why does this matter beyond the benchmark leaderboards? The annotation bottleneck is one of the biggest practical obstacles to deploying EEG technology outside the lab. Collecting EEG data is relatively straightforward with modern consumer-grade headsets, but attaching reliable emotion labels to that data requires either self-reports that are subjective and coarse, or elicitation protocols that assume the stimuli reliably provoke the intended feelings. If a system can be trained on labeled data from a modest group of volunteers and then adapt to new users without any labeling at all, the economics of EEG emotion recognition change completely. Affective brain-computer interfaces could be calibrated for each new user in minutes rather than requiring lengthy labeled recording sessions, opening the door to applications in driver monitoring, adaptive learning systems, neurofeedback therapy, and passive mental state tracking.

The work also reflects a broader convergence in neural signal processing. Graph neural networks, originally developed for molecules and social networks, have proven natural fits for EEG because electrode layouts are literally graphs embedded on the scalp. Attention mechanisms, the engine behind large language models, excel at deciding which features matter, and here they are repurposed to arbitrate between two mathematical views of the same frequency content. Bidirectional recurrent networks, veterans of speech and text processing, bring their temporal modeling strengths to brain dynamics. EEG-UDAF is a reminder that progress in brain-computer interfaces often comes not from exotic new sensors but from assembling proven machine learning components in ways that respect the peculiar structure of neural data: its spatial topology, its oscillatory organization, and its stubborn variability.

Caveats remain, as they always do in this field. The framework was validated on public datasets recorded under controlled conditions with elicited emotions, and real-world emotional experiences are messier than laboratory stimuli. Unsupervised domain adaptation reduces but does not eliminate the distribution gap between subjects, and the semantic correction mechanism must walk a fine line between aligning distributions and distorting class structure. The authors note that their experiments demonstrate effective alleviation of non-stationarity and individual differences, a carefully worded claim that stops short of declaring the problems solved. Still, the direction of travel is clear. As frameworks like EEG-UDAF mature, the vision of technology that understands not just what we do but how we feel moves closer to reality, powered by algorithms that can learn from one brain and speak to another without a single label in between.

Subject of Research: Unsupervised domain adaptation for EEG-based emotion recognition

Article Title: EEG-UDAF: A novel unsupervised domain adaptation framework leveraging dual-view frequency and spatio-temporal features for EEG-based emotion recognition

Article References: Chen, Y., Zhou, G., Wang, P., Nan, Y., Xi, Y., Cao, R., Lu, J., & He, Z. (2026). EEG-UDAF: A novel unsupervised domain adaptation framework leveraging dual-view frequency and spatio-temporal features for EEG-based emotion recognition. Journal of Big Data. https://doi.org/10.1186/s40537-026-01584-5

Image Credits: AI Generated

DOI: 10.1186/s40537-026-01584-5

Keywords: EEG, emotion recognition, unsupervised domain adaptation, graph convolutional network, BiLSTM, differential entropy, power spectral density, cross-attention, brain-computer interface, deep learning, non-stationarity, individual differences

The research, published in PLOS Complex Systems by Axel Hutt of the University of Strasbourg and CNRS, Matteus McCulloch and Jérémie Lefebvre of the University of Ottawa, Anthony Hudetz of the University of Michigan, and Aref Pariz of The Royal’s Institute of Mental Health Research, tackles a long-standing puzzle in electrophysiology. The aperiodic component of the power spectral density, or PSD, is not a fixed signature of the brain. Its slope, quantified by an exponent that describes how quickly power decays at higher frequencies, changes systematically with behavioral state. Steep spectra dominated by large-amplitude slow fluctuations appear during sleep and anesthesia, while flatter spectra accompany arousal and wakefulness. The same exponent also tracks cognitive load, task demands, and even aging, making it one of the most widely used spectral biomarkers in modern neuroscience.

Beyond the slope, the aperiodic spectrum is structured by a spectral knee, a bend that separates distinct scaling regimes and is thought to reflect an underlying neuronal timescale. When the knee moves, the frequency range over which the power law can be meaningfully measured moves with it, dragging estimates of the exponent along. Clinically and scientifically, this matters: if the knee shifts, an apparent change in the exponent may say nothing about a change in the actual asymptotic scaling of the signal, and everything about a change in the timescale of the neural population producing it. Yet the biological origin of both the knee and the exponent has remained hotly contested, with proposed explanations ranging from self-organized criticality to the passive filtering of signals by brain tissue and random ion diffusion.

To dissect the problem, the team built a large-scale recurrent neural network with sparse, balanced excitatory and inhibitory connectivity of the kind long used to model interacting neural populations. Each node in the network represents a local population whose mesoscopic synaptic potential evolves under a linear relaxation process, a constant input, and a nonlinear sigmoidal activation function. Crucially, the network was driven by state-dependent fluctuations, modeled as additive Gaussian noise whose variance stands in for the intensity of endogenous broadband activity. Low noise variance corresponded to low-activation states reminiscent of deep anesthesia or slow-wave sleep, while high variance emulated the elevated, irregular activity of awake and aroused cortex, where endogenous fluctuations are particularly prominent thanks to sensory input, brainstem modulation, and inter-area communication.

The simulations revealed a striking pattern. When the noise variance was small, the network settled into epochs of high-amplitude, low-frequency, synchronous activity, and its power spectrum showed strong low-frequency power with a steep high-frequency decay. As the variance increased, low-frequency power was suppressed, higher frequencies gained ground, and the spectral knee visibly shifted toward higher frequencies. At frequencies beyond the knee, however, the asymptotic scaling exponent approached a value that remained largely invariant across noise levels. In other words, the fluctuations were not changing the fundamental scaling law of the network; they were moving the boundary between scaling regimes, reorganizing the spectrum by tuning the effective timescale over which neural populations evolve.

Capturing this behavior mathematically required the researchers to go beyond standard tools. The conventional approach to computing a PSD in a high-dimensional nonlinear network is to linearize the dynamics around a deterministic equilibrium, but this method is computationally awkward for large networks and, more importantly, structurally blind to the phenomenon at hand. A linearized treatment predicts a Lorentzian spectral profile whose knee and slope are fixed regardless of noise intensity; changing the noise merely rescales the overall amplitude of the spectrum. Since both the simulations and the experimental literature show marked state-dependent shifts in knee position and slope, the team concluded that nonlinear effects are essential, and set out to build an analytical framework that could include them.

By integrating recent advances in random matrix theory, and in particular the circular law that constrains the eigenvalues of random connectivity matrices within a disk in the complex plane, the researchers derived a variational equation inspired by Floquet theory and the master stability function. Rather than linearizing about a deterministic fixed point, they analyzed stability about the stochastic system’s evolution, assuming a Gaussian stationary probability density for the network activity, an approximation validated by extensive simulations. This yielded an effective spectral radius that depends explicitly on the noise variance: increasing the noise contracts the spectral disk, steering the network away from the May-Wigner instability threshold and reinforcing linear, stable behavior. Noise, in this framework, is not merely a nuisance but a stabilizing control parameter.

Substituting this effective spectral radius back into the expression for the mean power spectrum produced an analytical prediction with an explicit, nonlinear dependence on noise variance, and it matched the simulations with remarkable fidelity. The theory predicted, and the numerics confirmed, that the characteristic knee frequency increases with noise variance, indicating a reduction in the effective neuronal timescale. The resulting displacement of the knee flattened the spectrum within fixed frequency bands, decreasing the apparent exponent exactly as observed empirically. These results held across multiple orders of magnitude of noise variance, across different levels of connectivity heterogeneity, in networks conforming to Dale’s law, and in the presence of temporally correlated colored noise, though the magnitude of the effect declined as the noise autocorrelation time grew.

To connect theory with biology, the team revisited local field potential recordings from rats under the anesthetic desflurane, comparing light and deep anesthesia. The spectra were steeper under deep anesthesia, and a paired bootstrap analysis across seven trials found a statistically significant increase in the exponent with anesthetic depth. This aligns with the hypothesis that deeper anesthesia suppresses endogenous fluctuations, potentially through dampening of the ascending reticular arousal system, and thus corresponds to a lower intrinsic noise level, a slower knee frequency, and a steeper spectrum. The experimental trend qualitatively mirrors the model’s central prediction.

The broader implication is provocative: shifts in widely used spectral biomarkers may reflect noise-driven, nonlinear transitions in recurrent neural systems rather than the action of any single biophysical mechanism. Variations in the statistics of endogenous neural activity alone, without any change in connectivity, could account for spectral changes previously attributed to rewiring or altered microarchitecture. The authors are careful to note that their model is deliberately simplified, using random Erdős–Rényi connectivity rather than the hierarchical, small-world architectures of real brains, and that tissue filtering, synaptic filtering, and ion diffusion likely act as complementary mechanisms shaping spectra across different frequency ranges and recording modalities. Still, the message is clear: the brain’s broadband background is dynamic, its characteristic features are generic properties of nonlinear recurrent networks, and the noise that neuroscientists so often average away may be one of the very levers by which the brain reorganizes itself across sleep, wakefulness, and every state in between.

Subject of Research: Mechanisms by which noise-like neural fluctuations drive state-dependent shifts in neuronal timescales and 1/f power spectra

Article Title: Noise-like fluctuations drive shifts in neuronal timescales and 1/ f power spectra across brain states

Article References: Hutt, A., McCulloch, M., Hudetz, A. G., Pariz, A., & Lefebvre, J. (2026). Noise-like fluctuations drive shifts in neuronal timescales and 1/f power spectra across brain states. PLOS Complex Systems, 3(9), e0000133. https://doi.org/10.1371/journal.pcsy.0000133

Image Credits: AI Generated

DOI: 10.1371/journal.pcsy.0000133

Keywords: neuroscience, 1/f power spectrum, spectral knee, neural noise, brain states, recurrent neural networks, random matrix theory, EEG, anesthesia, neuronal timescales, power spectral density, criticality

The challenge stems from the very technique that makes optical wireless so spectrally efficient: high-order quadrature amplitude modulation, or QAM. Schemes such as 256-QAM and 512-QAM pack enormous amounts of data into each symbol by using extremely dense constellations of amplitude and phase combinations. But as the constellation density grows, the decision boundaries between neighboring symbols shrink to razor-thin margins, making the signal exquisitely sensitive to noise, phase fluctuations, and the nonlinear distortions introduced by optical components such as light-emitting diodes and laser diodes. In intensity-modulated direct-detection systems, where light intensity cannot go negative, engineers must also contend with a high peak-to-average power ratio, which further degrades signal quality. Traditional receivers that rely on analytical models and linear approximations simply cannot capture these nonlinearities, and their performance collapses precisely where 6G needs it most.

The research team built their evaluation around a two-user optical NOMA system, in which both users share the same time and frequency resources through power-domain superposition coding. The user with the weaker channel is allocated 80 percent of the transmit power while the stronger user receives 20 percent, and at the receiver a successive interference cancellation procedure separates the overlapping signals. The channel model incorporated a realistic battery of impairments: distance-dependent path loss with an exponent of 2.2, Rayleigh fading, additive white Gaussian noise, optical nonlinear distortion from the light source, and hardware impairments including synchronization errors. This deliberately harsh environment was designed to emulate the conditions a real 6G optical link would face, with independent channel realizations generated for each user to reflect heterogeneous propagation conditions.

Against this backdrop, the authors evaluated six detectors within a single unified framework. The minimum mean square error (MMSE) detector offers a favorable complexity-performance trade-off but remains fundamentally linear. The QR-decomposition with M-algorithm maximum likelihood detector (QRM-MLD) approaches optimal performance through tree search but becomes computationally prohibitive at high modulation orders. Convolutional neural network (CNN) detectors learn spatial signal patterns efficiently but cannot model temporal dependencies, while recurrent neural network (RNN) detectors capture channel memory and time correlations at the cost of slower convergence. Two hybrids, CNN-MMSE and the newly proposed RNN-MMSE, combine statistical estimation with deep learning refinement. Prior studies had examined these techniques only in isolation; the unified comparison is itself a significant contribution.

The proposed RNN-MMSE detector operates in three stages. First, an RNN learns the channel characteristics directly from pilot symbols, replacing conventional pilot-based estimation and capturing temporal variations in the process. Second, the MMSE stage applies a statistically optimal linear estimate that suppresses noise and linear interference, producing a structured initial symbol estimate. Third, a second RNN refines that estimate by learning only the residual nonlinear distortions and temporal dependencies that the linear stage could not remove. The authors provide a theoretical justification rooted in the orthogonality principle of linear estimation: because the MMSE output extracts all linearly available information, the residual presented to the recurrent network has a substantially smaller variance than the raw signal, which smooths the optimization landscape, lowers gradient variance, and accelerates training convergence.

The simulation results are striking. At 64-QAM, conventional optical NOMA detection required approximately 13.5 dB of signal-to-noise ratio to achieve a bit error rate of 10 to the minus 3, while the proposed RNN-MMSE detector reached the same target at just 6.5 to 7 dB. As the modulation order climbed, the gap widened. At 128-QAM the hybrid detector delivered a 7 to 8 dB improvement, at 256-QAM an 8 to 9 dB gain, and at 512-QAM, the most demanding scheme tested, it achieved a 9 to 10 dB advantage, reaching the target error rate at roughly 12.5 to 13 dB where the conventional baseline needed 21 to 22 dB. Standalone CNN and RNN detectors and the CNN-MMSE hybrid improved on the classical methods but consistently fell short of the proposed architecture.

Spectral behavior told a similar story. Using Welch power spectral density estimation with Hamming windowing, the team measured out-of-band leakage across all detectors. Conventional optical NOMA exhibited in-band power spectral density around minus 40 dB/Hz with poor suppression near minus 60 dB/Hz, while the RNN-MMSE detector pushed spectral leakage below minus 130 dB/Hz at 512-QAM, indicating excellent sidelobe suppression and minimal adjacent-channel interference. Training convergence favored the hybrid as well: over 20 epochs the RNN-MMSE model began at roughly 85 percent accuracy and converged to about 90 percent, outperforming standalone models by 18 to 32 percentage points. Statistical validation across 20 independent Monte Carlo runs confirmed the results were reproducible, with a standard deviation of only 0.22 dB on the key SNR metric and a 95 percent confidence interval of plus or minus 0.10 dB.

Practicality was a central concern. The entire framework was implemented in MATLAB R2024a on a workstation with an Intel Core i7 processor, 32 GB of RAM, and an NVIDIA RTX 3060 GPU. Training on 100,000 generated signal samples took only 14 to 16 minutes, and inference required approximately 3.8 milliseconds per OFDM frame, fast enough for near-real-time deployment. Memory consumption stayed below 800 MB during inference. The authors acknowledge that the hybrid architecture carries more computational overhead than a standalone linear detector, with complexity of the order of N cubed for the MMSE stage plus N squared terms for the recurrent stage, but they argue the performance gains justify the cost, particularly since the heavy training can be performed offline once while deployment requires only lightweight forward inference that GPUs, FPGAs, or edge accelerators can handle.

The study is candid about its limits. All results derive from Monte Carlo simulation rather than a physical testbed, and the authors identify experimental validation on an LED- or laser-based intensity-modulation link with software-defined-radio baseband processing as an immediate priority. Their comparison with Transformer-based detectors draws on literature benchmarks obtained under different channel models rather than a controlled head-to-head experiment, and their robustness analysis of imperfect channel estimation, imperfect interference cancellation, and synchronization errors is mechanistic rather than simulated. The two-user configuration, while modular and scalable in principle, will need extension to denser multi-user networks with adaptive grouping and dynamic power allocation. Even so, the findings make a compelling case that hybrid intelligent detection is not an incremental refinement but a necessary evolution for 6G optical networks, and the roadmap toward testbed validation, attention-based comparisons, and large-scale multi-user trials is already clearly drawn.

Subject of Research: Neural network-based signal detection for 6G optical non-orthogonal multiple access systems with high-order QAM modulation

Article Title: Comprehensive analysis of 6G optical NOMA waveforms using neural network-based detection for high-order QAM modulation

Article References: Kumar, A., Revathi, V., Gaur, N., & Nanthaamornphong, A. (2026). Comprehensive analysis of 6G optical NOMA waveforms using neural network-based detection for high-order QAM modulation. Results in Optics, 25, Article 101159. https://doi.org/10.1016/j.rio.2026.101159

Image Credits: AI Generated

DOI: 10.1016/j.rio.2026.101159

Keywords: 6G, optical wireless communication, NOMA, QAM, recurrent neural network, MMSE detection, deep learning, bit error rate, power spectral density, IM/DD, OFDM, signal detection

The framework, called FAGNet, short for Frequency-Guided Adaptive Graph Network, begins with a deceptively simple observation. Financial time series are not single, uniform phenomena. They are superpositions of processes operating at different speeds: fast, high-frequency fluctuations driven by intraday trading and noise, medium-frequency swings tied to weekly and monthly cycles, and slow, low-frequency drifts that reflect macroeconomic fundamentals. When a graph neural network ingests these mixed signals directly, it must somehow learn to disentangle all of these scales simultaneously, a task the authors argue is fundamentally ill-suited to time-domain processing. Their solution is to apply the Fourier transform, a mathematical technique with a century of pedigree in physics and engineering, to project the coupled stock price signals into the frequency domain, where each oscillation speed becomes a distinct, measurable component.

But simply moving to the frequency domain is not enough, and this is where the technical heart of the paper lies. A naive frequency decomposition, such as splitting the spectrum into bands of equal width, would not guarantee that each band carries comparable information. Two bands of identical width can contain wildly different amounts of energy, and therefore wildly different amounts of signal. FAGNet addresses this with an adaptive energy-based spectral partitioning mechanism guided by the power spectral density, a function that describes how the power of a signal is distributed across frequencies. Rather than carving the spectrum into equal slices, the model divides it into sub-bands that each contain roughly equal energy. The result, the authors contend, is genuine scale decoupling at the level of information content: each sub-band represents a comparable share of the market’s total dynamical activity, so no scale is systematically over- or under-represented in the downstream analysis.

Once the spectrum has been partitioned, the framework constructs a dedicated graph neural network for each decoupled sub-band. This is a crucial architectural decision. Risk contagion, the authors argue, is scale-dependent: the network of relationships through which a fast-moving shock propagates may look very different from the network along which slow-moving fundamentals diffuse. A rapid panic might spread through algorithmic trading links and sentiment spillovers, while a slow repricing might follow industrial supply chains and shared analyst coverage, a phenomenon documented in the finance literature. By modelling each frequency band with its own graph network, FAGNet can learn the contagion topology that is specific to that scale, rather than forcing a single graph structure to explain contagion at every speed simultaneously. The separated scale signals thus become the raw material for a family of specialised graph learners, each attuned to one slice of the market’s rhythm.

The final stage of the pipeline tackles a problem that arises whenever multiple models are combined: how should their outputs be fused? A static weighted average would be blind to changing market conditions, yet the relative importance of different scales is precisely what shifts during a crisis. FAGNet therefore employs a cross-scale interaction and context-aware dynamic fusion module. This module does two things. First, it captures synergistic dependencies between risk contagions at different scales, recognising that a slow-moving fundamental weakness can amplify a fast-moving panic, and vice versa. Second, it adaptively adjusts the prediction weight assigned to each scale according to the instantaneous state of the market. In turbulent conditions, the model can lean more heavily on the frequency bands that carry the most relevant information; in calmer periods, it can rebalance toward the slower components. The fusion is not a fixed formula but a learned, state-sensitive mechanism.

The motivation for this frequency-domain perspective draws on a substantial body of prior research. Economists have long known that systematic risk behaves differently at different time scales, with wavelet-based studies of emerging stock markets showing that the relationship between risk and horizon is far from uniform. More recently, the machine learning community has embraced frequency methods for time series: frequency-enhanced decomposed transformers have improved long-term forecasting, frequency-domain multilayer perceptrons have proven to be effective learners, and Fourier-based graph networks have been applied to multivariate prediction problems. FAGNet extends this line of work by combining frequency decomposition with graph-based contagion modelling in a principled way, using the energy content of the spectrum, rather than arbitrary frequency cutoffs, to define the scales themselves.

The empirical case for the approach rests on extensive experiments on real-world stock market datasets, where the authors report that FAGNet significantly outperforms state-of-the-art baseline methods. The baselines against which such models are typically judged include temporal relational ranking systems, dynamic graph attention networks, integrated convolutional and recurrent architectures built on knowledge-incorporated graphs, and hybrid frameworks that combine graph attention with recurrent memory cells. The claim that a frequency-guided design beats these established approaches suggests that the multi-scale entanglement of risk contagion is not merely a theoretical nuisance but a genuine bottleneck limiting the accuracy of existing models. The authors argue that the results validate frequency domain decoupling as a new paradigm for financial risk modelling, one that is both more accurate and more dynamically interpretable than time-domain alternatives.

That word, interpretable, deserves emphasis, because it points to what may be the framework’s most valuable property for practitioners. Deep learning models in finance are often criticised as black boxes: they may predict well, but they offer little insight into why. FAGNet’s architecture is different in kind. Because each graph network operates on a well-defined frequency band, its learned contagion topology can be read as a statement about how risk propagates at that particular scale. The dynamic fusion weights, meanwhile, provide a running record of which scales the model considers most informative at any moment. In principle, an analyst could inspect these weights during a market stress event and see whether the model is relying on fast, panic-like components or slow, fundamental ones. This is a form of transparency that emerges naturally from the design, rather than being bolted on afterwards, and it aligns with a broader movement in the field toward models whose internal structure mirrors the phenomena they describe.

The study also situates itself within a rapidly growing literature on graph-based financial prediction. Graph neural networks have become the dominant approach for modelling risk contagion, with applications ranging from recommending profitable stocks via financial graph attention networks to predicting stock movements using hybrid-relational market knowledge graphs. Yet the authors identify a persistent limitation: the vast majority of deep graph models operate directly in the raw time domain, making it difficult for them to handle the multi-scale entanglement and dynamic variability of contagion. Related work by some of the same authors, including a frequency-domain graph learning framework for understanding risk propagation and a frequency-decoupled progressive graph learning approach for heterogeneous contagion, indicates that this research group has been systematically developing the ideas that culminate in FAGNet. The new paper consolidates that programme into a single adaptive architecture.

Caveats remain, as they always do in financial machine learning. The paper reports that data will be made available on request, and independent replication on different markets, asset classes and time periods will be needed to establish how general the gains are. Markets are adversarial environments where any published predictive edge tends to erode as it is arbitraged away, and no architecture, however elegant, escapes that dynamic. Still, the conceptual contribution stands on its own merits. By treating financial contagion as a multi-scale signal processing problem, partitioning the spectrum by energy content, and letting each scale have its own graph model with state-aware fusion, FAGNet offers a template that could plausibly extend beyond equities to other networked dynamical systems, from cryptocurrency markets to interbank lending networks. For a field searching for ways to make deep learning both sharper and more transparent, the message is that sometimes the answer is not a bigger network, but a better coordinate system in which to run it.

Subject of Research: Frequency-domain graph neural networks for financial risk contagion and stock market prediction

Article Title: Fagnet: an adaptive frequency-guided multi-scale graph network for financial market prediction

Article References: Zhou, C., Xiao, S., & Hu, C. (2026). Fagnet: an adaptive frequency-guided multi-scale graph network for financial market prediction. International Journal of Machine Learning and Cybernetics, 17(10), Article 485. https://doi.org/10.1007/s13042-026-03301-3

Image Credits: AI Generated

DOI: 10.1007/s13042-026-03301-3

Keywords: graph neural networks, stock market prediction, frequency domain analysis, risk contagion, Fourier transform, power spectral density, multi-scale learning, financial networks, deep learning, dynamic fusion, interpretable AI, machine learning

The scientific and legal momentum behind tools like TUNE stems largely from the European Union’s Marine Strategy Framework Directive, the landmark 2008 legislation that obliges member states to achieve Good Environmental Status in their marine waters. Under the Directive’s Descriptor 11, which covers the introduction of energy, including underwater sound, two distinct criteria apply. The first addresses low- and mid-frequency impulsive sounds, such as the hammering of pile drivers during offshore construction, while the second targets continuous low-frequency noise, dominated overwhelmingly by commercial shipping. Shipping noise is the most pervasive form of underwater noise pollution and dominates ambient sound levels below one kilohertz across much of the world’s oceans. Criterion D11C2 specifically requires monitoring programs to report sound pressure levels within decidecade frequency bands centered at 63 and 125 hertz, two narrow slices of the acoustic spectrum that serve as proxies for overall shipping-related noise exposure.

Meeting that regulatory requirement sounds straightforward, but in practice it has been anything but. Reliable estimation of these acoustic quantities depends on robust and reproducible signal processing workflows, and the software landscape has long been fragmented. Raven Pro, a mainstay of bioacoustics research, offers limited flexibility in calibration and requires a commercial license. MATLAB-based tools such as CHORUS provide advanced processing capabilities but are not distributed as freely available standalone applications. The open-source MANTA package has improved accessibility and standardization for long-term ocean soundscape analysis, but it produces outputs in hybrid millidecade bands rather than the decidecade bands adopted for Marine Strategy Framework Directive reporting. Many other packages remain proprietary or closed-source, preventing users from inspecting and verifying the algorithms that ultimately feed into official environmental assessments. TUNE was engineered to close this gap by combining calibration accuracy, usability and regulatory compliance in a single freely available package distributed under the GNU General Public License.

At its core, TUNE processes calibrated acoustic recordings in wav format and computes a suite of standard underwater noise metrics. These include broadband sound pressure level, decidecade sound pressure level, zero-to-peak sound pressure level and power spectral density estimates. The software embraces recent international standardization: it uses the term decidecade in place of one-third octave band, following the 2025 update of the relevant ISO standard, which considers the two metrics approximately equivalent, and it adopts the sound pressure symbol conventions of ISO 18405:2017 for the definitions and equations of all four metrics. Processing is carried out over user-defined time intervals called snapshots, a concept aligned with the temporal observation windows introduced in recent ISO documents. Each recording can be segmented into consecutive, skipped or overlapping snapshots, allowing fully time-resolved estimates of how noise levels rise and fall, for example as a vessel approaches and then recedes from a monitoring station.

A distinguishing technical feature of TUNE is its support for frequency-dependent calibration. Hydrophones, the underwater microphones that record ocean sound, do not respond uniformly across frequencies, and their sensitivity can vary substantially across the acoustic spectrum. Many existing tools accept only a single calibration factor, which introduces errors when the hydrophone response is uneven. TUNE instead accepts a two-column text file containing frequency values and corresponding hydrophone sensitivity, expressed either as voltage sensitivity in decibels relative to one volt per micropascal or as an equivalent calibration factor in pascals per volt, with a minimum of three frequency points required. Spectral estimates are corrected bin by bin using this sensitivity curve, improving the accuracy of the resulting sound pressure levels. The software also allows optional high-pass, low-pass or band-pass digital filtering before analysis, with filters designed to fixed passband ripple and stopband attenuation specifications, and it supports flexible FFT parameters including window type, overlap and transform length.

Internally, the software follows a modular architecture organized into five stages: user input, data import, calibration, signal processing and output generation. A single-window graphical user interface lets users configure directories, calibration files, time ranges, spectrogram settings and filtering options, and the entire parameter set can be saved and reloaded as a setup file to guarantee reproducibility across monitoring campaigns. Processing proceeds through three nested loops at the file, channel and snapshot levels, with multi-channel recordings handled channel by channel. For each snapshot, the software computes a series of Fast Fourier Transforms over signal subsegments and averages them into a single representative spectrum, then corrects that spectrum against the calibration curve. Decidecade integration is performed using the nearest available FFT bins rather than exact theoretical band edges, a pragmatic approximation that benchmarking described in the paper’s supplementary materials shows produces negligible discrepancies for standard monitoring applications, while dramatically accelerating computation. One creative addition is the option to overlay Knudsen curves for sea states zero through six, letting analysts compare measured ambient noise against classic empirical predictions of wind-driven sea-state noise.

To demonstrate the tool, the researchers analyzed a real recording of a ship passing through a harbor, captured with a Teledyne Reson TC4040 hydrophone and VP1000 preamplifier feeding a Tascam DR-100 recorder at 96 kilohertz sampling with 24-bit resolution. The ten-minute recording was processed in ten-second snapshots using 131072-point FFTs with a Hann window and fifty percent overlap, and the entire computation completed in just two minutes and thirty-five seconds. The results were vivid: mean sound pressure levels in the 63 and 125 hertz decidecade bands reached 111 and 117 decibels respectively, the mean broadband level was 123 decibels, and the maximum zero-to-peak level hit 156 decibels at 256 seconds into the recording, precisely the closest point of approach of the vessel. All four metrics rose clearly during the first half of the recording, matching the ship passage documented in the campaign logs, and every output was delivered both as figures and as numerical text files ready for statistical aggregation across recordings, sites and seasons.

The impact of the software is already visible in practice. By providing a consistent, transparent and reproducible framework for acoustic data processing, TUNE directly addresses a long-standing problem: differences in signal processing workflows can produce substantial discrepancies in reported sound pressure levels, undermining the comparability of environmental assessments across institutions and countries. Perhaps most significantly, the tool has been adopted by multiple Regional Environmental Protection Agencies in Italy, establishing a shared national framework for underwater noise monitoring and overcoming methodological fragmentation among regional programs. This software-driven harmonization improves the consistency of reported indicators, increases transparency in reporting and supports the coordinated implementation of national monitoring obligations under the Marine Strategy Framework Directive. Because TUNE can process long-term acoustic datasets and generate decidecade-band time series, it also supports the estimation of baseline conditions, trend analysis and threshold exceedance assessments in line with recent European Commission guidance on threshold values.

The developers see TUNE as a foundation rather than a finished product. Future versions will extend the range of computed metrics to include the Sound Exposure Level, which would allow the software to cover the Directive’s impulsive-sound criterion in addition to continuous noise, along with greater flexibility in temporal analysis. Beyond European regulatory monitoring, the authors note that the tool’s applicability extends to a broad range of marine and freshwater acoustic studies, from low-noise calibration sites to research linking low-frequency noise shifts to changes in fin whale song. In a field where proprietary licenses and closed algorithms have often stood between environmental agencies and their legal obligations, TUNE represents a quietly radical proposition: that the measurement of a pollutant threatening ocean life should be accurate, transparent and free for anyone who needs it.

Subject of Research: A free standalone software tool for standardized evaluation of underwater noise pollution under the Marine Strategy Framework Directive

Article Title: TUNE: a Tool for Underwater Noise Evaluation

Article References: Caradonna, V., Buogo, S., Azzellino, A., Vicinanza, D., & Borsani, J. F. (2026). TUNE: a Tool for Underwater Noise Evaluation. SoftwareX, 35, Article 103008. https://doi.org/10.1016/j.softx.2026.103008

Image Credits: AI Generated

DOI: 10.1016/j.softx.2026.103008

Keywords: underwater noise, TUNE software, Marine Strategy Framework Directive, shipping noise, hydrophone calibration, sound pressure level, decidecade bands, power spectral density, ocean acoustics, marine monitoring, MATLAB, environmental protection