Dementia classification from spontaneous speech using wrapper-based feature selection

Abstract

Dementia encompasses a group of syndromes that impair cognitive functions such as memory, reasoning, and the ability to perform daily activities. As populations globally age, nearly 10 million new dementia cases occur annually. Clinical diagnosis remains challenging because symptoms overlap with other conditions and require comprehensive cognitive assessment, highlighting the need for feasible and accurate detection methods. Recent advances in machine learning have highlighted spontaneous speech as a promising noninvasive, cost-effective, and scalable biomarker for dementia detection. In this study, spontaneous speech recordings from the ADReSS dataset and the extended Pitt Corpus were analyzed, consisting of picture description tasks performed by cognitively healthy individuals and participants with Alzheimer’s disease or dementia. Unlike many prior approaches relying on speech-active segments, acoustic features were extracted from entire recordings with the openSMILE toolkit. This recording-level representation reduces the number of feature vectors and provides a computationally efficient framework for dementia classification, while indirectly incorporating pause- and hesitation-related information. Classification models with classifier-based wrapper feature selection were employed to estimate feature importance and identify diagnostically relevant acoustic characteristics. Among the evaluated classifiers, the extreme minimal learning machine emerged as the most computationally efficient method, providing competitive classification accuracy with substantially lower training time in repeated leave-one-subject-out validation. The results demonstrated that the proposed framework is computationally efficient, interpretable, and well-suited as a supportive tool for speech-based dementia assessment.

Keywords

dementia / spontaneous speech / acoustic features / feature extraction / classification

References

[1] F. Pasquier, Early diagnosis of dementia: neuropsychology, J. Neurol., 246 (1999), 6-15. https://doi.org/10.1007/s004150050299

[2] World Health Organization, Global action plan on the public health response to dementia 2017-2025, Geneva: WHO Document Production Services, 2017.

[3] Y. Wang, M. L. Haaksma, I. H. Ramakers, F. R. Verhey, W. M. van de Flier, P. Scheltens, et al., Cognitive and functional progression of dementia in two longitudinal studies, Int. J. Geriatr. Psych., 34 (2019), 1623-1632. https://doi.org/10.1002/gps.5175

[4] M. F. Folstein, S. E. Folstein, P. R. McHugh, “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician, J. Psychiatr. Res., 12 (1975), 189-198. https://doi.org/10.1016/0022-3956(75)90026-6

[5] Z. S. Nasreddine, N. A. Phillips, V. Bédirian, S. Charbonneau, V. Whitehead, I. Collin, et al., The montreal cognitive assessment, moca: a brief screening tool for mild cognitive impairment, J. Amer. Geriatr. Soc., 53 (2005), 695-699. https://doi.org/10.1111/j.1532-5415.2005.53221.x

[6] K. A. Welsh, N. Butters, R. C. Mohs, D. Beekly, S. Edland, G. Fillenbaum et al., The consortium to establish a registry for alzheimer’s disease (cerad), part v: a normative study of the neuropsychological battery, Neurology, 44 (1994), 609- 609. https://doi.org/10.1212/WNL.44.4.609

[7] G. Zorluoglu, M. E. Kamasak, L. Tavacioglu, P. O. Ozanar, A mobile application for cognitive screening of dementia, Comput. Meth. Prog. Bio., 118 (2015), 252-262. https://doi.org/10.1016/j.cmpb.2014.11.004

[8] C. R. Jack, D. A. Bennett, K. Blennow, M. C. Carrillo, B. Dunn, S. B. Haeberlein, et al., Nia-aa research framework: toward a biological definition of alzheimer’s disease, Alzh. Dement., 14 (2018), 535-562. https://doi.org/10.1016/j.jalz.2018.02.018

[9] H. Hanyu, T. Asano, T. Iwamoto, M. Takasaki, H. Shindo, K. Abe, Magnetization transfer measurements of the hippocampus in patients with alzheimer’s disease, vascular dementia, and other types of dementia, Amer. J. Neuroradiol., 21 (2000), 1235-1242.

[10] M. A. Ebrahimighahnavieh, S. Luo, R. Chiong, Deep learning to detect alzheimer’s disease from neuroimaging: a systematic literature review, Comput. Meth. Prog. Bio., 187 (2020), 105242. https://doi.org/10.1016/j.cmpb.2019.105242

[11] G. Mirzaei, H. Adeli, Machine learning techniques for diagnosis of alzheimer disease, mild cognitive disorder, and other types of dementia, Biomed. Signal Proces., 72 (2022), 103293. https://doi.org/10.1016/j.bspc.2021.103293

[12] P. N. Young, M. Estarellas, E. Coomans, M. Srikrishna, H. Beaumont, A. Maass, et al., Imaging biomarkers in neurodegeneration: current and future practices, Alz. Res. Therapy, 12 (2020), 49. https://doi.org/10.1186/s13195-020-00612-7

[13] J. Reilly, J. E. Peelle, S. M. Antonucci, M. Grossman, Anomia as a marker of distinct semantic memory impairments in alzheimer’s disease and semantic dementia, Neuropsychology, 25 (2011), 413-426. https://doi.org/10.1037/a0022738

[14] O. Ivanova, I. Martínez-Nicolás, E. García-Piñuela, J. J. G. Meilán, Defying syntactic preservation in alzheimer’s disease: what type of impairment predicts syntactic change in dementia (if it does) and why? Front. Lang. Sci., 2 (2023), 1199107. https://doi.org/10.3389/flang.2023.1199107

[15] K. C. Fraser, J. A. Meltzer, F. Rudzicz, Linguistic features identify alzheimer’s disease in narrative speech, Journal of Alzheimer’s Disease, 49 (2015), 407-422. https://doi.org/10.3233/JAD-150520

[16] S. De la Fuente Garcia, C. W. Ritchie, S. Luz, Artificial intelligence, speech, and language processing approaches to monitoring alzheimer’s disease: a systematic review, Journal of Alzheimer’s Disease, 78 (2020), 1547-1574. https://doi.org/10.3233/JAD-200888

[17] H. Zhen, Y. Shi, J. J. Yang, J. M. Vehni, Co-supervised learning paradigm with conditional generative adversarial networks for sample-efficient classification, Applied Computing and Intelligence, 3 (2023), 13-26. https://doi.org/10.3934/aci.2023002

[18] M. Alsuhaibani, A. Pourramezan Fard, J. Sun, F. Far Poor, P. S. Pressman, M. H. Mahoor, A review of machine learning approaches for non-invasive cognitive impairment detection, IEEE Access, 13 (2025), 56355-56384. https://doi.org/10.1109/ACCESS.2025.3555176

[19] D. Ortiz-Perez, P. Ruiz-Ponce, D. Tomás, J. Garcia-Rodriguez, M. F. Vizcaya-Moreno, M. Leo, A deep learning-based multimodal architecture to predict signs of dementia, Neurocomputing, 548 (2023), 126413. https://doi.org/10.1016/j.neucom.2023.126413

[20] E. ŞAHİN, N. N. Arslan, D. Özdemir, Unlocking the black box: an in-depth review on interpretability, explainability, and reliability in deep learning, Neural Comput. Appl., 37 (2025), 859-965. https://doi.org/10.1007/s00521-024-10437-2

[21] J. Linja, J. Hämäläinen, P. Nieminen, T. Kärkkäinen, Feature selection for distance-based regression: An umbrella review and a one-shot wrapper, Neurocomputing, 518 (2023), 344-359. https://doi.org/10.1016/j.neucom.2022.11.023

[22] M. R. Kumar, S. Vekkot, S. Lalitha, D. Gupta, V. J. Govindraj, K. Shaukat, et al., Dementia detection from speech using machine learning and deep learning architectures, Sensors, 22 (2022), 9311. https://doi.org/10.3390/s22239311

[23] S. Luz, F. Haider, S. de la Fuente Garcia, D. Fromm, B. MacWhinney, Alzheimer’s dementia recognition through spontaneous speech, Front. Comput. Sci., 3 (2021), 780169. https://doi.org/10.3389/fcomp.2021.780169

[24] C. Cortes, V. Vapnik, Support-vector networks, Mach. Learn., 20 (1995), 273-297. https://doi.org/10.1007/BF00994018

[25] T. Hastie, R. Tibshirani, J. Friedman, The elements of statistical learning: data mining, inference, and prediction, 2 Eds., New York: Springer, 2009. https://doi.org/10.1007/978-0-387-84858-7

[26] T. Kärkkäinen, Extreme minimal learning machine: ridge regression with distance-based basis, Neurocomputing, 342 (2019), 33-48. https://doi.org/10.1016/j.neucom.2018.12.078

[27] J. J. G. Meilán, F. Martínez-Sánchez, J. Carro, D. E. López, L. Millian-Morell, J. M. Arana, Speech in alzheimer’s disease: can temporal and acoustic parameters discriminate dementia? Dement. Geriatr. Cogn., 37 (2014), 327-334. https://doi.org/10.1159/000356726

[28] I. Martínez-Nicolás, T. E. Llorente, F. Martínez-Sánchez, J. J. G. Meilán, Ten years of research on automatic voice and speech analysis of people with alzheimer’s disease and mild cognitive impairment: a systematic review article, Front. Psychol., 12 (2021), 620251. https://doi.org/10.3389/fpsyg.2021.620251

[29] K. R. Scherer, T. Johnstone, G. Klasmeyer, Vocal expression of emotion, In: Handbook of affective sciences, Oxford: Oxford University Press, 2003, 433-456. https://doi.org/10.1093/oso/9780195126013.003.0023

[30] A. Pistono, M. Jucla, E. J. Barbeau, L. Saint-Aubert, B. Lemesle, B. Calvet, et al., Pauses during autobiographical discourse reflect episodic memory processes in early alzheimer’s disease, Journal of Alzheimer’s Disease, 50 (2016), 687-698. https://doi.org/10.3233/JAD-150408

[31] R. Pappagari, J. Cho, L. Moro-Velázquez, N. Dehak, Using state of the art speaker recognition and natural language processing technologies to detect alzheimer’s disease and assess its severity, Proceedings of Interspeech 2020, 2020, 2177-2181. https://doi.org/10.21437/Interspeech.2020-2587

[32] A. Meghanani, C. S. Anoop, A. Ramakrishnan, An exploration of log-mel spectrogram and mfcc features for alzheimer’s dementia recognition from spontaneous speech, Proceedings of IEEE Spoken Language Technology Workshop (SLT), 2021, 670-677. https://doi.org/10.1109/SLT48900.2021.9383491

[33] M. M. Parlak, G. Saylam, M. A. Babademez, Ö. B. Munis, S. A. Tokgöz, Voice analysis results in individuals with alzheimer’s disease: How do age and cognitive status affect voice parameters? Brain Behav., 13 (2023), e3271. https://doi.org/10.1002/brb3.3271

[34] F. Eyben, K. R. Scherer, B. W. Schuller, J. Sundberg, E. André, C. Busso, et al., The geneva minimalistic acoustic parameter set (gemaps) for voice research and affective computing, IEEE Trans. Affect. Comput., 7 (2015), 190-202. https://doi.org/10.1109/TAFFC.2015.2457417

[35] F. Eyben, Real-time speech and music classification by large audio feature space extraction, Cham: Springer, 2015. https://doi.org/10.1007/978-3-319-27299-3

[36] G. Degottex, J. Kane, T. Drugman, T. Raitio, S. Scherer, Covarep-a collaborative voice analysis repository for speech technologies, Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing, 2014, 960-964. https://doi.org/10.1109/ICASSP.2014.6853739

[37] D. Snyder, D. Garcia-Romero, G. Sell, D. Povey, S. Khudanpur, X-vectors: robust dnn embeddings for speaker recognition, Proceedings of IEEE international conference on acoustics, speech and signal processing (ICASSP), 2018, 5329-5333. https://doi.org/10.1109/ICASSP.2018.8461375

[38] J. F. Gemmeke, D. P. Ellis, D. Freedman, A. Jansen, W. Lawrence, R. C. Moore, et al., Audio set: an ontology and human-labeled dataset for audio events, Proceedings of IEEE international conference on acoustics, speech and signal processing (ICASSP), 2017, 776-780. https://doi.org/10.1109/ICASSP.2017.7952261

[39] N. Cummins, Y. Pan, Z. Ren, J. Fritsch, V. S. Nallanthighal, H. Christensen, et al., A comparison of acoustic and linguistics methodologies for alzheimer’s dementia recognition, Proceedings of Interspeech 2020, 2020, 2182-2186. https://doi.org/10.21437/Interspeech.2020-2635

[40] A. Baevski, Y. Zhou, A. Mohamed, M. Auli, Wav2vec 2.0: a framework for self-supervised learning of speech representations, Proceedings of the 34th International Conference on Neural Information Processing Systems, 2020, 12449-12460.

[41] A. Balagopalan, B. Eyre, J. Robin, F. Rudzicz, J. Novikova, Comparing pre-trained and feature-based models for prediction of Alzheimer’s disease based on speech, Front. Aging Neurosci., 13 (2021), 635945. https://doi.org/10.3389/fnagi.2021.635945

[42] N. Dehak, P. J. Kenny, R. Dehak, P. Dumouchel, P. Ouellet, Front-end factor analysis for speaker verification, IEEE Trans. Audio Speech., 19 (2010), 788-798. https://doi.org/10.1109/TASL.2010.2064307

[43] M. Schmitt, B. Schuller, openXBOW-Introducing the Passau open-source crossmodal Bag-of-Words toolkit, J. Mach. Learn. Res., 18 (2017), 1-5.

[44] M. S. S. Syed, Z. S. Syed, M. Lech, E. Pirogova, Automated screening for alzheimer’s dementia through spontaneous speech, Proceedings of Interspeech 2020, 2020, 2222-2226. https://doi.org/10.21437/Interspeech.2020-3158

[45] F. Haider, S. De La Fuente, S. Luz, An assessment of paralinguistic acoustic features for detection of alzheimer’s dementia in spontaneous speech, IEEE J.-STSP, 14 (2019), 272-281. https://doi.org/10.1109/JSTSP.2019.2955022

[46] M. Niemela, M. von Bonsdorff, S. Ayrämo, T. Kärkkäinen, Classification of dementia from spoken speech using feature selection and the bag of acoustic words model, Applied Computing and Intelligence, 4 (2024), 45-65. https://doi.org/10.3934/aci.2024004

[47] M. Martinc, S. Pollak, Tackling the adress challenge: a multimodal approach to the automated recognition of alzheimer’s dementia, Proceedings of Interspeech 2020, 2020, 2157-2161. https://doi.org/10.21437/Interspeech.2020-2202

[48] M. Rohanian, J. Hough, M. Purver, Multi-modal fusion with gating using audio, lexical and disfluency features for alzheimer’s dementia recognition from spontaneous speech, Proceedings of Interspeech 2020, 2020, 2187-2191. https://doi.org/10.21437/Interspeech.2020-2721

[49] J. T. Becker, F. Boiler, O. L. Lopez, J. Saxton, K. L. McGonigle, The natural history of alzheimer’s disease: description of study cohort and accuracy of diagnosis, Arch. Neurol., 51 (1994), 585-594. https://doi.org/10.1001/ARCHNEUR.1994.00540180063015

[50] M. W. Fong, R. Van Patten, R. P. Fucetola, The factor structure of the boston diagnostic aphasia examination, J. Int. Neuropsych. Soc., 25 (2019), 772-776. https://doi.org/10.1017/S1355617719000237

[51] B. MacWhinney, The CHILDES project: Tools for analyzing talk, volume I: transcription format and programs, New York: Psychology Press, 2014. https://doi.org/10.4324/9781315805672

[52] R. EBU-Recommendation, Loudness normalisation and permitted maximum level of audio signals, European Broadcasting Union, 2011. Available from: https://tech.ebu.ch/publications/r128.

[53] International Telecommunication Union, Subjective test methodology for evaluating speech communication systems that include noise suppression algorithm, ITU-T Publications, 2003. Available from: https://www.itu.int/rec/T-REC-P.835/en.

[54] F. Eyben, M. Wöllmer, B. Schuller, Opensmile: the munich versatile and fast open-source audio feature extractor, Proceedings of the 18th ACM International Conference on Multimedia, 2010, 1459-1462. https://doi.org/10.1145/1873951.1874246

[55] J. Hämäläinen, A. S. C. Alencar, T. Kärkkäinen, C. L. C. Mattos, H. S. Amauri, J. P. P. Gomes, Minimal learning machine: theoretical results and clustering-based reference point selection, J. Mach. Learn. Res., 21 (2020), 1-29.

[56] T. Kärkkäinen, Model selection for extreme minimal learning machine using sampling, Proceedings of the 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, 2019, 391-396.

[57] T. Kärkkäinen, S. Ayrämo, On computation of spatial median for robust data mining, Proceedings of Evolutionary and Deterministic Methods for Design, Optimization and Control with Applications to Industrial and Societal Problems, 2005, 1-14.

[58] J. Qian, T. Hastie, J. Friedman, R. Tibshirani, N. Simon, Glmnet for matlab, Jemdoc, 2014. Available from: https://hastie.su.domains/glmnet_matlab/.

[59] L. Breiman, Random forests, Mach. Learn., 45 (2001), 5-32. https://doi.org/10.1023/A:1010933404324

[60] J. D. Gibbons, S. Chakraborti, Nonparametric statistical inference, New York: Chapman and Hall/CRC, 2010. https://doi.org/10.1201/9781439896129

[61] S. Jauhiainen, J. P. Kauppi, M. Leppänen, K. Pasanen, J. Parkkari, T. Vasankari, et al., New machine learning approach for detection of injury risk factors in young team sport athletes, Int. J. Sports Med., 42 (2021), 175-182. https://doi.org/10.1055/a-1231-5304

[62] M. Saarela, S. Jauhiainen, Comparison of feature importance measures as explanations for classification models, SN Appl. Sci., 3 (2021), 272. https://doi.org/10.1007/s42452-021-04148-9

[63] S. Luz, F. Haider, S. De la Fuente, D. Fromm, B. MacWhinney, Detecting cognitive decline using speech only: the adresso challenge, arXiv: 2104.09356. https://doi.org/10.48550/arXiv.2104.09356

[64] J. Koo, J. H. Lee, J. Pyo, Y. Jo, K. Lee, Exploiting multi-modal features from pre-trained networks for alzheimer’s dementia recognition, arXiv: 2009.04070. https://doi.org/10.48550/arXiv.2009.04070

[65] A. Pompili, T. Rolland, A. Abad, The inesc-id multi-modal system for the adress 2020 challenge, arXiv: 2005.14646. https://doi.org/10.48550/arXiv.2005.14646

[66] E. Edwards, C. Dognin, B. Bollepalli, M. K. Singh, V. Analytics, Multiscale system for alzheimer’s dementia recognition through spontaneous speech, Proceedings of Interspeech 2020, 2020, 2197-2201. https://doi.org/10.21437/Interspeech.2020-2781

[67] M. Kappen, G. Vanhollebeke, J. Van Der Donckt, S. Van Hoecke, M. A. Vanderhasselt, Acoustic and prosodic speech features reflect physiological stress but not isolated negative affect: a multi-paradigm study on psychosocial stressors, Sci. Rep., 14 (2024), 5515. https://doi.org/10.1038/s41598-024-55550-3

[68] S. Saeedi, S. Hetjens, M. Grimm, B. B. v. Latoszek, Acoustic speech analysis in alzheimer’s disease: a systematic review and meta-analysis, The Journal of Prevention of Alzheimer’s Disease, 11 (2024), 1789-1797. https://doi.org/10.14283/jpad.2024.132

[69] Z. Jafari, M. K. Andrew, K. J. Rockwood, Diagnostic utility of speech-based biomarkers in mild cognitive impairment: a systematic review and meta-analysis, Age Ageing, 54 (2025), afaf316. https://doi.org/10.1093/ageing/afaf316

[70] P. Herd, D. Carr, C. Roan, Cohort profile: Wisconsin longitudinal study (wls), Int. J. Epidemiol., 43 (2014), 34-41. https://doi.org/10.1093/ije/dys194

[71] S. Luz, S. D. L. F. Garcia, F. Haider, D. Fromm, B. MacWhinney, A. Lanzi, et al., Connected speech-based cognitive assessment in chinese and english, arXiv: 2406.10272. https://doi.org/10.48550/arXiv.2406.10272