The team behind the system, Shimaa Nagro and Maha Helal of the College of Computing and Informatics, set out to close both gaps at once. Their approach begins with the raw audio of cardiopulmonary sounds, the acoustic signatures produced by the mechanical activity of the heart and the movement of air through the respiratory tract. Rather than choosing between the two dominant ways of turning sound into something a neural network can digest, XCardioPulmoNet uses both. The first representation is the Mel spectrogram, a time-frequency image that mimics aspects of human hearing by compressing frequency onto a scale that emphasizes the ranges most relevant to perception. The second is the Wavelet transform, which decomposes a signal into components at multiple scales and is particularly good at capturing brief, transient events, the sharp snap of a heart murmur or the fleeting crackle of an infected lung.

Each of these representations has complementary strengths. Mel spectrograms excel at encoding sustained spectral patterns across time, while wavelet-based images preserve fine temporal detail that can be smeared away by conventional Fourier-style analysis. XCardioPulmoNet therefore merges the information from both into a single hybrid image, a composite visual fingerprint of the sound that carries time-domain and frequency-domain cues side by side. This image-based strategy transforms the diagnostic problem into one that computer vision has already largely solved: classifying pictures. Any advance in image recognition architecture can, in principle, be brought to bear on the sounds of the body.

And that is precisely what the researchers did. At the heart of the framework sits EAGR_Net, a residual attention-guided convolutional neural network designed to squeeze maximal discriminative power out of the hybrid images. Residual connections, a now-classic deep learning innovation, allow very deep networks to train stably by giving signals shortcut paths around layers, so that each block of the network only needs to learn what remains to be corrected rather than relearning everything from scratch. Attention mechanisms then act like a spotlight within the network, weighting the feature channels and spatial locations that matter most for distinguishing, say, a normal heartbeat from one marred by valvular disease, while suppressing irrelevant background noise and artifacts.

The results reported in the paper are striking. Evaluated with five-fold cross-validation, a rigorous protocol in which the data are repeatedly split so that every recording serves in both training and independent testing, EAGR_Net outperformed both a baseline convolutional neural network and an attention-equipped CNN. On the heart sound dataset, the model achieved a ROC-AUC of 0.9851, a measure of how well it separates diseased from healthy cases across all decision thresholds, and a PR-AUC of 0.9558, which is especially informative when positive disease cases are relatively rare. On the lung sound dataset the performance climbed even higher, reaching a ROC-AUC of 0.9978 and a PR-AUC of 0.9917, figures that approach the practical ceiling of what a classifier can achieve.

But raw accuracy is only half the story, and it is arguably the less important half for medicine. Deep learning models have earned a reputation as black boxes: they produce answers without revealing their reasoning, a property that has slowed their adoption in clinical settings where doctors must justify every diagnosis. XCardioPulmoNet addresses this directly with Gradient-Weighted Class Activation Mapping, or Grad-CAM, a technique that produces heatmaps showing which regions of the input image contributed most strongly to the model’s prediction. In this framework, those heatmaps translate back to specific moments and frequency bands in the original sound, allowing a clinician to see, for example, that the network flagged a particular systolic interval as the telltale evidence for its classification.

This built-in explainability matters for more than just trust. By visualizing the discriminative regions, researchers can verify that the model is attending to physiologically meaningful features rather than exploiting spurious correlations, such as recording artifacts or sensor noise, that would not generalize to new patients and new devices. The interpretability layer thus doubles as a diagnostic tool for the diagnostic tool itself, exposing potential failure modes before the system ever encounters a real patient. In multiclass settings, where the model must distinguish among several categories of cardiopulmonary abnormality rather than simply flagging disease versus health, knowing which acoustic features drive each decision becomes essential for clinical credibility.

The unified treatment of heart and lung sounds is another notable departure from convention. Cardiology and pulmonology have traditionally run on parallel computational tracks, with separate models trained separately on phonocardiograms and lung sound recordings. By designing a single pipeline that handles both modalities, XCardioPulmoNet points toward integrated screening tools that could assess multiple organ systems from one examination. Because auscultation is inexpensive, noninvasive, and already embedded in routine clinical practice almost everywhere in the world, a reliable automated interpreter could extend specialist-level screening to clinics, rural health posts, and telemedicine consultations where cardiology and pulmonology expertise is scarce.

The implications reach into several of medicine’s most pressing challenges. Cardiovascular and respiratory diseases remain leading causes of death globally, and early detection is one of the most effective levers for improving outcomes. Heart murmurs, arrhythmia-related acoustic irregularities, wheezes, crackles, and rubs all leave characteristic traces in recorded sound, and a system that reliably characterizes them could triage patients, flag deteriorating conditions in hospital monitoring, or support large-scale screening programs. The authors suggest that the framework’s robustness, interpretability, and diagnostic potential make it a candidate for exactly these kinds of automated multiclass disease detection tasks, though translating laboratory benchmark performance into validated clinical deployment will require the usual gauntlet of prospective testing on diverse patient populations, varied recording hardware, and real-world noise conditions.

Published as open access on 9 October 2026, with the work received in June and accepted that September, the study arrives as part of a broader wave of research applying deep learning to medical acoustics, a field where machine learning techniques are reshaping everything from speech analysis to ultrasound. What distinguishes this contribution is its combination of complementary signal representations, a purpose-built attention-guided residual architecture, and a transparency mechanism woven into the pipeline from the start. The research, which the authors report received no external funding, offers a template for how medical AI systems might be built to earn the confidence of the clinicians who would actually use them: not just by being right more often, but by showing their work. If that combination proves durable in clinical validation, the stethoscope’s next two centuries may belong to a partnership between human ears and algorithms that can explain what they hear.

Subject of Research: Explainable deep learning classification of heart and lung sounds for cardiopulmonary disease detection

Article Title: XCardioPulmoNet: an explainable attention-guided residual framework for cardiopulmonary disease detection using sound classification

Article References: Nagro, S., & Helal, M. (2026). XCardioPulmoNet: an explainable attention-guided residual framework for cardiopulmonary disease detection using sound classification. Complex & Intelligent Systems. https://doi.org/10.1007/s40747-026-02539-2

Image Credits: AI Generated

DOI: 10.1007/s40747-026-02539-2

Keywords: deep learning, convolutional neural networks, heart sounds, lung sounds, Mel spectrogram, wavelet transform, attention mechanism, Grad-CAM, medical acoustics, cardiopulmonary disease, computer-aided diagnosis, explainable AI

The system, which the researchers call FiWAS, rests on two core fiber technologies developed at the National University of Singapore. The first is a contact microphone drawn as a fiber roughly 150 micrometers in diameter, comparable in thickness to a coarse human hair. The fiber is built from piezoelectric materials, meaning that mechanical pressure from sound vibrations generates an electrical signal within the fiber itself. Because the fiber is sewn, knitted, or woven directly into fabric and sits in intimate contact with the skin, it picks up vibrations traveling through the body rather than sound waves propagating through the air. This contact-based approach is critical: it allows the fiber to overcome the acoustic impedance mismatch that normally reflects most body-borne sound energy back into the tissue, and it inherently rejects airborne environmental noise, one of the biggest obstacles to reliable auscultation outside a quiet clinic room.

The second key component is a knitted fiber computer approximately 500 micrometers in diameter. This is not a conventional rigid chip attached to clothing but a functional digital device in fiber form, capable of real-time signal processing of the acoustic data streaming in from the microphone fibers. The fiber computer also handles wireless communication, transmitting processed cardiopulmonary sounds over Bluetooth to a mobile device. By embedding both the sensing and the computing elements as fibers, the researchers avoided the bulky rigid modules, wires, and adhesive patches that have plagued earlier attempts at wearable acoustic monitoring. The result is a garment that looks and behaves like ordinary clothing while functioning as a distributed, multi-point listening array across the torso.

Acoustic sensing of the body is a deceptively hard engineering problem. Heart sounds occupy low frequencies, and pathological features such as murmurs and gallops can be subtle. Lung sounds, including the crackles that suggest fluid accumulation, are transient and easily masked by motion artifacts, ambient noise, and the rustle of clothing itself. Conventional electronic stethoscopes address some of these issues with rigid chest pieces and careful manual placement, but they capture sound at a single point for a limited time. The fiber approach inverts this logic: instead of one high-quality microphone applied briefly, the garment deploys an array of modest but robust contact microphones distributed across the chest, capturing sound continuously and from multiple anatomical locations simultaneously.

The clinical validation of FiWAS involved two very different patient populations. In one study, the team monitored 10 adult patients with heart failure, a condition in which the failing heart often produces characteristic extra sounds, including the S3 gallop, and in which fluid backing up into the lungs generates crackles on auscultation. In a second study, the garment was used on 20 neonates, an especially demanding application because newborns are tiny, fragile, and impossible to instruct. The researchers report that the system captured cardiopulmonary sounds in both groups, including qualitative examples of pathological acoustic features such as cardiac murmurs, S3 gallops, and pulmonary crackles. Comparative testing against a digital stethoscope confirmed that the garment could suppress environmental noise and achieve high-fidelity sound capture, suggesting that clothing-based auscultation can match the diagnostic quality of a handheld clinical instrument.

One of the most striking capabilities demonstrated in the study is spatial sound mapping. Because the microphone fibers are distributed across the garment as an array, the system does not merely record a single sound channel; it can localize where within the chest a particular acoustic event originates. This matters clinically. Cardiac murmurs, for example, radiate from specific valve locations, and abnormal lung sounds concentrate in particular lobes or regions affected by disease. By mapping the intensity and timing of sounds across the fabric array, the garment can provide localized diagnostic information, essentially producing an acoustic image of cardiac and pulmonary function that a conventional stethoscope, applied point by point, can only approximate through the clinician’s memory and skill.

The significance of the work is amplified by the epidemiological backdrop. Cardiopulmonary diseases remain the leading causes of death and disability worldwide, and heart failure alone affects tens of millions of people, with outcomes that studies have shown can be worse than those of many common cancers. Early detection and continuous monitoring are known to improve outcomes, but current tools are episodic. A patient with worsening heart failure may develop increasing crackles and a new S3 gallop days before symptoms become severe enough to prompt a hospital visit. A garment that listens around the clock could flag these acoustic warning signs remotely, allowing clinicians to intervene before decompensation, potentially reducing hospitalizations and saving lives.

The neonatal application deserves particular attention. Premature and newborn infants in intensive care require constant monitoring, yet the devices attached to them, from electrodes to pulse oximeter probes, are adhesive, rigid, and can injure delicate skin. A soft fabric garment with embedded fiber sensors offers a gentler alternative, capturing heart and breath sounds continuously without any adhesive contact. The study’s demonstration of wireless neonatal monitoring suggests a path toward monitoring that is both more comfortable for the infant and richer in information for the clinical team, complementing existing physiological measurements with continuous acoustic data.

The broader context is the rapidly maturing field of functional fibers and textile electronics. Over the past two decades, researchers have learned to draw multimaterial fibers that can sense, compute, store energy, and communicate, and recent advances have produced fiber batteries, fiber displays, and even single-fiber computers capable of machine-learning inference within fabric networks. The FiWAS work represents a translation of these materials-science advances into a concrete clinical application with human validation data, a step that many textile electronics concepts have not yet taken. The researchers have also made their analysis code publicly available through GitHub and their raw data through Figshare, supporting reproducibility and enabling other groups to build on the platform.

Challenges remain before such garments become routine clinical or consumer products. Long-term durability through washing and daily wear, regulatory approval as a medical device, reimbursement pathways, and the development of robust algorithms that can automatically interpret continuous streams of heart and lung sounds all lie ahead. The current study reports qualitative examples of pathological sounds rather than establishing diagnostic performance metrics for disease detection, so larger and longer trials will be needed to define exactly how the technology should be used in practice. Nevertheless, the demonstration that a garment can capture clinically meaningful cardiopulmonary sounds, suppress environmental noise, map sounds spatially, and work in everyone from heart failure patients to newborns marks a genuine milestone. The humble stethoscope, invented in 1816, may finally be evolving into something that never needs to be picked up at all, because the shirt you are already wearing is doing the listening.

Subject of Research: An all-fiber wearable auscultation garment for continuous wireless monitoring of cardiopulmonary sounds

Article Title: All-fiber wireless auscultation garment for cardiopulmonary diagnostics

Article References: Li, Q., Li, Z., Zhang, X., Akgun, H., Chaturvedi, S., Teo, H. K., Anbalakan, K., Sim, D., Chew Yun Chi, K. M., Yeo, K. T., Hausenloy, D. J., & Khudiyev, T. (2026). All-fiber wireless auscultation garment for cardiopulmonary diagnostics. Nature Cardiovascular Research. https://doi.org/10.1038/s44161-026-00882-8

Image Credits: AI Generated

DOI: 10.1038/s44161-026-00882-8

Keywords: wearable sensors, fiber electronics, auscultation, heart failure, cardiopulmonary monitoring, smart textiles, piezoelectric fibers, neonatal monitoring, heart sounds, digital health, acoustic sensing, fiber computer