{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,23]],"date-time":"2026-07-23T20:15:50Z","timestamp":1784837750067,"version":"3.55.0"},"reference-count":32,"publisher":"MDPI AG","issue":"7","license":[{"start":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T00:00:00Z","timestamp":1782950400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Ministry of Economic Development of the Russian Federation","award":["139-15-2025-011"],"award-info":[{"award-number":["139-15-2025-011"]}]},{"name":"Ministry of Economic Development of the Russian Federation","award":["000000C313925P4G0002"],"award-info":[{"award-number":["000000C313925P4G0002"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Life"],"abstract":"<jats:p>Respiratory diseases are a major global health challenge. However, identification of respiratory diseases is often limited by subjectivity, environmental noise and inter-clinician variability. This study presents an explainable multimodal deep learning framework for recording-level multiclass classification of respiratory audio signals. The proposed system integrates two complementary representations\u2014a spectro-temporal encoder based on a CNN\u2013BiLSTM-attention architecture and a handcrafted acoustic-feature encoder capturing acoustic descriptors commonly used in respiratory-audio analysis, including MFCCs, zero-crossing rate, spectral centroid, spectral bandwidth, chroma, RMS energy, and spectral rolloff features. These branches are combined through late-stage fusion to leverage both data-driven representation learning and domain-informed acoustic cues. The proposed model was trained and internally evaluated on the Asthma Detection Dataset Version 2, comprising five respiratory categories: bronchial disease, asthma, COPD, healthy, and pneumonia. Mono conversion, resampling to 16 kHz, 100\u20132000 Hz band-pass filtering, amplitude normalisation, fixed 4 s trimming or zero-padding, training-only augmentation, handcrafted-feature extraction, mel-spectrogram generation, quality control auditing, and stratified recording-level partitioning have been applied in the pre-processing steps. Across five repeated experiments with different random seeds, the proposed hybrid model achieved a mean held-out recording-level test accuracy of 0.9099\u00b10.0163, balanced accuracy of 0.8936\u00b10.0152, macro F1-score of 0.8937\u00b10.0177, macro ROC\u2013AUC of 0.9867\u00b10.0010, and macro PR\u2013AUC of 0.9489\u00b10.0044. Conventional machine learning baseline comparisons showed that the proposed model achieved stronger internal accuracy, balanced accuracy, macro recall, macro F1-score, and macro ROC\u2013AUC than classical machine learning algorithms trained on handcrafted acoustic features, although Random Forest remained competitive in macro PR\u2013AUC. Ablation analysis shows that the deep spectro-temporal branch was the primary contributor to predictive performance, while the handcrafted branch provided complementary interpretable acoustic information rather than consistently improving all classification metrics. Explainability was incorporated using Grad-CAM and Integrated Gradients for spectrogram-based interpretation and SHAP for handcrafted-feature attribution. Domain-shift evaluation on the ICBHI Respiratory Sound Database and a COPD-focused cohort revealed substantial dataset shift effects, including poor healthy-case recognition on ICBHI and seed-dependent COPD recognition in the COPD-focused cohort. Identifier-aware sensitivity analyses showed lower performance than the main recording-level split, suggesting that subject-like or source-level overlap may inflate internal performance estimates. The findings should be interpreted as promising internal held-out recording-level algorithmic performance with limited external transfer, rather than evidence of readiness for clinical use.<\/jats:p>","DOI":"10.3390\/life16071108","type":"journal-article","created":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T14:51:25Z","timestamp":1783003885000},"page":"1108","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Explainable Multi-Modal Deep Learning for Recording-Level Classification of Respiratory Audio Signals Under Internal and Domain-Shift Evaluation"],"prefix":"10.3390","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-2046-7177","authenticated-orcid":false,"given":"S M Asiful Islam","family":"Saky","sequence":"first","affiliation":[{"name":"School of Computing and Informatics, Albukhary International University, Alor Setar 05200, Kedah, Malaysia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-6112-7673","authenticated-orcid":false,"given":"Md Saiful","family":"Arefin","sequence":"additional","affiliation":[{"name":"School of Computing and Informatics, Albukhary International University, Alor Setar 05200, Kedah, Malaysia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-7593-1028","authenticated-orcid":false,"given":"Md Rashidul","family":"Islam","sequence":"additional","affiliation":[{"name":"School of Computing and Informatics, Albukhary International University, Alor Setar 05200, Kedah, Malaysia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-1405-2892","authenticated-orcid":false,"given":"Mohammad Saiful","family":"Islam","sequence":"additional","affiliation":[{"name":"School of Computing and Informatics, Albukhary International University, Alor Setar 05200, Kedah, Malaysia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3610-2562","authenticated-orcid":false,"given":"Rashadul Islam","family":"Sumon","sequence":"additional","affiliation":[{"name":"Institute of Digital Anti-Aging Healthcare, Inje University, Gimhae 50834, Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-2510-5167","authenticated-orcid":false,"given":"Md Mostafizur Rahman","family":"Masud","sequence":"additional","affiliation":[{"name":"Faculty of Computing, Universiti Teknologi Malaysia, Johor Bahru 81310, Johor, Malaysia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8117-9142","authenticated-orcid":false,"given":"Maria","family":"Lapina","sequence":"additional","affiliation":[{"name":"Trusted AI Research Center, Russian Academy of Sciences, 109004 Moscow, Russia"},{"name":"Department of Computational Mathematics and Cybernetics, North-Caucasus Federal University, 355017 Stavropol, Russia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7066-0061","authenticated-orcid":false,"given":"Mikhail","family":"Babenko","sequence":"additional","affiliation":[{"name":"Trusted AI Research Center, Russian Academy of Sciences, 109004 Moscow, Russia"},{"name":"Department of Computational Mathematics and Cybernetics, North-Caucasus Federal University, 355017 Stavropol, Russia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1165-7812","authenticated-orcid":false,"given":"Mohammed","family":"Muthanna","sequence":"additional","affiliation":[{"name":"Faculty of Computing and IT, Sohar University, Sohar 311, Oman"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,7,2]]},"reference":[{"key":"ref_1","unstructured":"World Health Organization (2025, November 16). Chronic Respiratory Diseases: More Than 80 Million Affected and Many More Undiagnosed, Warns New WHO and European Respiratory Society Report. Available online: https:\/\/www.who.int\/europe\/news\/item\/12-06-2025-chronic-respiratory-diseases--more-than-80-million-affected-and-many-more-undiagnosed--warns-new-who-andeuropean-respiratory-society-report."},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Kapetanidis, P., Kalioras, F., Tsakonas, C., Tzamalis, P., Kontogiannis, G., Karamanidou, T., Stavropoulos, T.G., and Nikoletseas, S. (2024). Respiratory diseases diagnosis using audio analysis and artificial intelligence: A systematic review. Sensors, 24.","DOI":"10.3390\/s24041173"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"012036","DOI":"10.1088\/1742-6596\/1937\/1\/012036","article-title":"Classification of lung disorders based on lung sounds using deep learning","volume":"1937","author":"Joy","year":"2021","journal-title":"J. Phys. Conf. Ser."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"17186","DOI":"10.1038\/s41598-021-96724-7","article-title":"Respiratory sound classification for crackles, wheezes, and rhonchi in the clinical field using deep learning","volume":"11","author":"Kim","year":"2021","journal-title":"Sci. Rep."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"13405","DOI":"10.1007\/s00500-022-07499-6","article-title":"Deep learning models for detecting respiratory pathologies from raw lung auscultation sounds","volume":"26","author":"Alqudah","year":"2022","journal-title":"Soft Comput."},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Yadav, S., Agarwal, P., and Rizvi, S.A.M. (2024). Automated Multiclass Classification of Lung Diseases with Deep Learning: A Hybrid CNN\u2013LSTM Approach. Proceedings of the 2024 4th International Conference on Technological Advancements in Computational Sciences (ICTACS), IEEE.","DOI":"10.1109\/ICTACS62700.2024.10841124"},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Gopi, G., and Varghese, A. (2025). Deep Learning for Respiratory Sound-Based Disease Classification. Proceedings of the 2025 International Conference on Advancements in Power, Communication and Intelligent Systems (APCI), IEEE.","DOI":"10.1109\/APCI65531.2025.11136918"},{"key":"ref_8","unstructured":"Sanjana, J., Naik, P.P., Padukudru, M.A., Koolagudi, S.G., and Rajan, J. (2023). Attention-Based CRNN Models for Identification of Respiratory Diseases from Lung Sounds. Proceedings of the 2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT), IEEE."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"105570","DOI":"10.1016\/j.bspc.2023.105570","article-title":"Respiratory sound-base disease classification and characterization with deep\/machine learning techniques","volume":"87","author":"Karaarslan","year":"2024","journal-title":"Biomed. Signal Process. Control"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"11667","DOI":"10.1007\/s00500-024-09866-x","article-title":"Deep learning and feature fusion-based lung sound recognition model to diagnoses the respiratory diseases","volume":"28","author":"Shehab","year":"2024","journal-title":"Soft Comput."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Vasava, R.P., and Joshiara, H.A. (2023). Different Respiratory Lung Sounds Prediction using Deep Learning. Proceedings of the 2023 4th International Conference on Electronics and Sustainable Communication Systems (ICESC), IEEE.","DOI":"10.1109\/ICESC57686.2023.10193040"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Jung, S.Y., Liao, C.H., Wu, Y.S., Yuan, S.M., and Sun, C.T. (2021). Efficiently classifying lung sounds through depthwise separable CNN models with fused STFT and MFCC features. Diagnostics, 11.","DOI":"10.3390\/diagnostics11040732"},{"key":"ref_13","unstructured":"Wang, Q., Bu, Z., Mao, J., Zhu, W., Zhao, J., Du, W., Shi, G., Zhou, M., Chen, S., and Qu, J. (2024). Towards reliable respiratory disease diagnosis based on cough sounds and vision transformers. arXiv."},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Petmezas, G., Cheimariotis, G.A., Stefanopoulos, L., Rocha, B., Paiva, R.P., Katsaggelos, A.K., and Maglaveras, N. (2022). Automated lung sound classification using a hybrid CNN\u2013LSTM network and focal loss function. Sensors, 22.","DOI":"10.3390\/s22031232"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"18","DOI":"10.1109\/MMUL.2024.3422022","article-title":"Multimodal Integration of an Enhanced Novel Pulmonary Auscultation Real-Time Diagnostic System","volume":"31","author":"Abhishek","year":"2024","journal-title":"IEEE Multimed."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Zhao, Z., Gong, Z., Niu, M., Ma, J., Wang, H., Zhang, Z., and Li, Y. (2022). Automatic respiratory sound classification via multi-branch temporal convolutional network. Proceedings of the ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE.","DOI":"10.1109\/ICASSP43922.2022.9746182"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"E15734056388107","DOI":"10.2174\/0115734056388107250710120917","article-title":"Evaluation of Deep Learning Methods for Pulmonary Disease Classification","volume":"21","author":"Singh","year":"2025","journal-title":"Curr. Med. Imaging"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"106257","DOI":"10.1016\/j.bspc.2024.106257","article-title":"Research on lung sound classification model based on dual-channel CNN-LSTM algorithm","volume":"94","author":"Zhang","year":"2024","journal-title":"Biomed. Signal Process. Control"},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"S, O., Sudhagar, G., and M, H. (2025, January 2\u201313). Deep Attention Fusion Framework on Lung Sound Classification by Multi-Spectral Features. Proceedings of the 2025 9th International Conference on Inventive Systems and Control (ICISC), Coimbatore, India.","DOI":"10.1109\/ICISC65841.2025.11187937"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Pham, L., McLoughlin, I., Phan, H., Tran, M., Nguyen, T., and Palaniappan, R. (2020). Robust deep learning framework for predicting respiratory anomalies and diseases. Proceedings of the 2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), IEEE.","DOI":"10.1109\/EMBC44109.2020.9175704"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Chamberlain, D., Kodgule, R., Ganelin, D., Miglani, V., and Fletcher, R. (2016). Application of Semi-Supervised Deep Learning to Lung Sound Analysis. Proceedings of the 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), IEEE.","DOI":"10.1109\/EMBC.2016.7590823"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Perna, D., and Tagarelli, A. (2019). Deep auscultation: Predicting respiratory anomalies and diseases via recurrent neural networks. Proceedings of the 2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS), IEEE.","DOI":"10.1109\/CBMS.2019.00020"},{"key":"ref_23","first-page":"54","article-title":"A Self-Attention Based Hybrid CNN-LSTM Architecture for Respiratory Sound Classification","volume":"18","author":"Bhushan","year":"2024","journal-title":"GMSARN Int. J."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"5236","DOI":"10.1109\/TCSVT.2025.3529473","article-title":"TM2SP: A Transformer-based Multi-Level Spatiotemporal Feature Pyramid Network for Video Saliency Prediction","volume":"35","author":"Li","year":"2025","journal-title":"IEEE Trans. Circuits Syst. Video Technol."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"2898","DOI":"10.1109\/JBHI.2022.3144314","article-title":"Robust and interpretable temporal convolution network for event detection in lung sound recordings","volume":"26","author":"Fernando","year":"2022","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"0074","DOI":"10.34133\/icomputing.0074","article-title":"Audio explainable artificial intelligence: A review","volume":"2","author":"Akman","year":"2024","journal-title":"Intell. Comput."},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Lo Giudice, M., Mammone, N., Ieracitano, C., Aguglia, U., Mandic, D., and Morabito, F.C. (2022). Explainable deep learning classification of respiratory sound for telemedicine applications. Proceedings of the International Conference on Applied Intelligence and Informatics, Springer.","DOI":"10.1007\/978-3-031-24801-6_28"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"84486","DOI":"10.1109\/ACCESS.2022.3197671","article-title":"Explainable AI for healthcare 5.0: Opportunities and challenges","volume":"10","author":"Saraswat","year":"2022","journal-title":"IEEE Access"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Amann, J., Blasimme, A., Vayena, E., Frey, D., Madai, V.I., and Consortium, P. (2020). Explainability for artificial intelligence in healthcare: A multidisciplinary perspective. Bmc Med. Inform. Decis. Mak., 20.","DOI":"10.1186\/s12911-020-01332-6"},{"key":"ref_30","unstructured":"Sun, Z. (2023). ICBHI 2017 Challenge, Harvard Dataverse. Version V1."},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Rocha, B.M., Filos, D., Mendes, L., Vogiatzis, I., Perantoni, E., Kaimakamis, E., Natsiavas, P., Oliveira, A., J\u00e1come, C., and Marques, A. (2017). A respiratory sound database for the development of automated classification. Proceedings of the International Conference on Biomedical and Health Informatics, Springer.","DOI":"10.1007\/978-981-10-7419-6_6"},{"key":"ref_32","unstructured":"Altan, G., and Kutlu, Y. (2020). RespiratoryDatabase@TR (COPD Severity Analysis), Mendeley Data. Version V1."}],"container-title":["Life"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2075-1729\/16\/7\/1108\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,2]],"date-time":"2026-07-02T15:32:01Z","timestamp":1783006321000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2075-1729\/16\/7\/1108"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,2]]},"references-count":32,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2026,7]]}},"alternative-id":["life16071108"],"URL":"https:\/\/doi.org\/10.3390\/life16071108","relation":{},"ISSN":["2075-1729"],"issn-type":[{"value":"2075-1729","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,2]]}}}