Detecting Alzheimer’s Disease Using Interactional and Acoustic Features from Spontaneous Speech
The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Abstract
Alzheimer's Disease (AD) is a form of Dementia that manifests in cognitive decline including memory, language, and changes in behavior.Speech data has proven valuable for inferring cognitive status, used in many health assessment tasks, and can be easily elicited in natural settings.Much work focuses on analysis using linguistic features; here, we focus on non-linguistic features and their use in distinguishing AD patients from similar-age Non-AD patients with other health conditions in the Carolinas Conversation Collection (CCC) dataset.We used two types of features: patterns of interaction including pausing behaviour and floor control, and acoustic features including pitch, amplitude, energy, and cepstral coefficients.Fusion of the two kinds of features, combined with feature selection, obtains very promising classification results: classification accuracy of 90% using standard models such as support vector machines and logistic regression.We also obtain promising results using interactional features alone (87% accuracy), which can be easily extracted from natural conversations in daily life and thus have the potential for future implementation as a noninvasive method for AD diagnosis and monitoring.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“For example, support vector machines have demonstrated particular efficacy in identifying imaging biomarkers associated with frontotemporal dementia.75”
From Review of Artificial Intelligence for Clinical Use in Alzheimer's Disease and Related Dementias (2026) · cites this paper as [75]
Citations
Cites
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016 · cited 1×
“We used the recursive feature elimination (RFE) method, an iterative process that removes a specific number of features, and examines the effect on classification accuracy [26].”
MethodSagi, O., & Rokach, L. (2018) · WIREs Data Mining and Knowledge Discovery · doi:10.1002/widm.1249 · Cited by 3,103 · cited 1×
“The prediction scores of each classifier are then combined using a standard soft voting ensemble method [29]: soft voting computes the average probability for each class over each component classifier and eventually bases the final prediction on maximum average probability.”
MethodCurrent understanding of Alzheimer’s disease diagnosis and treatment
Weller, J., & Budson, A. (2018) · F1000Research · doi:10.12688/f1000research.14506.1 · Cited by 1,083 · cited 1×
“Currently, standard AD diagnosis methods include clinical assessments complemented with family history, neuropsychological tests (including the Mini-Mental State Examination (MMSE) [4] and many others), self-report questionnaires, MRI [5] and Positron Emission Tomography (PET) [6].”
MethodLinguistic Features Identify Alzheimer’s Disease in Narrative Speech
Fraser, K. C., Meltzer, J. A., & Rudzicz, F. (2015) · Journal of Alzheimer's Disease · doi:10.3233/jad-150520 · Cited by 797 · cited 1×
“Linguistic variables are used to describe the quantitative and qualitative aspects of language production, for example via the decline in lexical-semantic abilities, word comprehension, verbal fluency, and syntactic processing for particular kinds of tasks such as picture description [7, 8].”
MethodConnected Speech in Neurodegenerative Language Disorders: A Review
Boschi, V., Catricalà, E., Consonni, M., Chesi, C., Moro, A., & Cappa, S. F. (2017) · Frontiers in Psychology · doi:10.3389/fpsyg.2017.00269 · Cited by 362 · cited 1×
“Linguistic variables are used to describe the quantitative and qualitative aspects of language production, for example via the decline in lexical-semantic abilities, word comprehension, verbal fluency, and syntactic processing for particular kinds of tasks such as picture description [7, 8].”
MethodGarcía, S. D. L. F., Ritchie, C. W., & Luz, S. F. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.06047 · Cited by 265 · cited 1×
“This use of interaction cues has the potential to be more versatile in AD prediction and monitoring in more daily life settings than individual language tasks [20, 21].”
BackgroundForbes-McKay, K. E., & Venneri, A. (2005) · Neurological Sciences · doi:10.1007/s10072-005-0467-9 · Cited by 261 · cited 1×
“The most prominent associations with AD are disparity in language production, speech comprehension, impaired reasoning, and memory functions, resulting in reduced vocabulary, verbal fluency, and difficulty performing daily tasks related to semantic information [3].”
BackgroundBeltrami, D., Gagliardi, G., Rossini Favretti, R., Ghidoni, E., Tamburini, F., & Calzà, L. (2018) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2018.00369 · Cited by 175 · cited 2×
“Standardized phonation time (SPT) and transformed phonation rate (TPR) also show significant differences between the two groups, con-firming findings in the literature [12].”
BackgroundSpeech in Alzheimer's Disease: Can Temporal and Acoustic Parameters Discriminate Dementia?
Meilán, J. J. G., Martínez-Sánchez, F., Carro, J., López, D. E., Millian-Morell, L., & Arana, J. M. (2014) · Dementia and Geriatric Cognitive Disorders · doi:10.1159/000356726 · Cited by 152 · cited 1×
“…feature was LogHNR, known to be important in acoustic analysis for the diagnosis of pathological voices; loudness, raw fundamental frequency, variation in jitter, intensity, and LogHNR all positively correlate with AD and have been reported as useful features in literature for Dementia [31, 9].”
BackgroundElsey, C., Drew, P., Jones, D., Blackburn, D., Wakefield, S., Harkness, K., Venneri, A., & Reuber, M. (2015) · Patient Education and Counseling · doi:10.1016/j.pec.2015.05.021 · Cited by 114 · cited 1×
“Conversation Analysis (CA) studies show that dialogue with dementia has characteristic features that would be missed if analyzing only individual speech [15, 16], but these studies are generally qualitative and/or small-scale.”
Background
Cited by
Ouyang, X., Shuai, X., Li, Y., Pan, L., Zhang, X., Fu, H., Cheng, S., Wang, X., Cao, S., Xin, J., et al. (2024) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3636534.3649370 · Cited by 86
“First, existing work is focused on a particular type of digital biomarker, such as motor function [17, 29], speech features [35, 44], or driving habits [12], which lacks generalizability to subjects with various demographic and medical characteristics.”
BackgroundMultimodal fusion for alzheimer’s disease recognition
Ying, Y., Yang, T., & Zhou, H. (2022) · Applied Intelligence · doi:10.1007/s10489-022-04255-z · Cited by 68
“While in AD recognition, designing AD-specific characteristics such as disfluencies, pause time are more reasonable [18–20].”
BackgroundNoninvasive automatic detection of Alzheimer's disease from spontaneous speech: a review
Qi, X., Zhou, Q., Dong, J., & Bao, W. (2023) · Frontiers in Aging Neuroscience · doi:10.3389/fnagi.2023.1224723 · Cited by 63
“…features include pauses (Yuan et al., 2020; Rohanian et al., 2021; Zhu et al., 2021), disfluency (Sarawgi et al., 2020; Qiao et al., 2021; Rohanian et al., 2021), interactional features (Nasreen et al., 2021a), cognition features (Sarawgi et al., 2020), ADR features (Martinc et al., 2021).”
BackgroundHuang, G., Li, R., Bai, Q., & Alty, J. (2023) · Health Information Science and Systems · doi:10.1007/s13755-023-00231-0 · Cited by 46
Voice biomarkers as indicators of cognitive changes in middle and later adulthood
Mahon, E., & Lachman, M. E. (2022) · Neurobiology of Aging · doi:10.1016/j.neurobiolaging.2022.06.010 · Cited by 42
Li, S., Nair, R., & Naqvi, S. M. (2024) · IEEE Journal of Translational Engineering in Health and Medicine · doi:10.1109/jtehm.2024.3369764 · Cited by 35
“Nasreen and Rohanian et al. [32], [33] designed AD-specific characteristics such as pause time, and disflu-encies for AD detection.”
BackgroundA longitudinal multi-modal dataset for dementia monitoring and diagnosis
Gkoumas, D., Wang, B., Tsakalidis, A., Wolters, M. K., Purver, M., Zubiaga, A., & Liakata, M. (2024) · Language Resources and Evaluation · doi:10.1007/s10579-023-09718-4 · Cited by 25
“III) The corpus is collected in the participants’ own environment using a custom-built tablet application.”
MethodSelf-citationZhang, Z., Wang, T., Hu, Z., Yang, L.-Z., & Li, H. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3509620 · Cited by 17
“A study by Nasreen and colleagues [22] focused on AD recognition with acoustic features, using multiple machine learning models to achieve an accuracy of 90%.”
A Study on Voice Measures in Patients With Alzheimer's Disease
Xiu, N., Vaxelaire, B., Li, L., Ling, Z., Xu, X., Huang, L., Sun, B., Huang, L., & Sock, R. (2022) · Journal of Voice · doi:10.1016/j.jvoice.2022.08.010 · Cited by 17
Alzheimer's Dementia Detection through Spontaneous Dialogue with Proactive Robotic Listeners
Li, Y., Lai, C., Lala, D., Inoue, K., & Kawahara, T. (2022) · ACM/IEEE International Conference on Human-Robot Interaction (HRI) · doi:10.1109/hri53351.2022.9889375 · Cited by 18
“We define six turn-pairs as a dialogue, and extract the following interactional features according to prior knowledge [45]: turn length (number of words per turn), floor control ratio (time amount during the human speaker speaks to the total speech time of the dialogue), standardized pause rate…”
Automatic selection of spoken language biomarkers for dementia detection
Ke, X., Mak, M., & Meng, H. M. L. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.10.018 · Cited by 10
Review of Artificial Intelligence for Clinical Use in Alzheimer's Disease and Related Dementias
Breithaupt, A. G., Tang, A., Paolillo, E. W., Bibars, M., Johnson, E. C. B., Saloner, R., Possin, K. L., Windon, C. C., Hill-Jarrett, T. G., Giorgio, J., et al. (2026) · Seminars in Neurology · doi:10.1055/a-2744-9871 · Cited by 6
“See Table 1 for a summary of AI approaches across modalities in ADRD, highlighting function, and clinical readiness.”
A Novel Approach to Mimic Linguistic Features of Transcripts of Atypical Speech
Mirbagheri, M., Saghir, H., & Chau, T. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3603921 · Cited by 6
“Previous studies have high-lighted some of these features, including pauses [34], [35], [36], disfluency [35], [37], [38], interactional features [39], cognition features [37], filler prediction [40], semantic features [41], grammatical constituents [42], and lexical diversity [43].”
A Systematic Review of Alzheimer's disease detection based on speech and natural language processing
Sevcik, A., & Rusko, M. (2022) · International Conference Radioelektronika · doi:10.1109/radioelektronika54537.2022.9764938 · Cited by 8
“The [30] ADReSS 79 87 0 166 Nasreen [31] CCC 15 15 0 30 Pappagari [32] ADReSS 79 87 0 166 Fu [33] PITT 82 82 0 164 Guerrero [34] PITT 98 98 0 196 Pappagari [32] ADReSS 78 78 0 156 Balagopalan [35] ADReSS 78 78 0 156 Thapa [12] PITT 168 98 0 266 Martinc [36] ADReSS 78 78 0 156 Fraser [11] PITT 97…”
Li, S., Song, S., & Naqvi, S. M. (2025) · IEEE Transactions on Affective Computing · doi:10.1109/taffc.2025.3604562 · Cited by 7
“Nasreen and Rohanian et al. [25], [51] developed specific characteristics tailored for Alzheimer’s detection, including pause time and disfluencies.”
Voice biomarkers in middle and later adulthood as predictors of cognitive changes
Mahon, E., & Lachman, M. E. (2024) · Frontiers in Psychology · doi:10.3389/fpsyg.2024.1422376 · Cited by 6
“Key prosody measures such as pitch, pulse, voice breaks, jitter, shimmer, and amplitude have shown potential (Konig et al., 2018, 2019; Martínez-Sánchez et al., 2012; Meilán et al., 2014; Mirheidari et al., 2019; Nasreen et al., 2021; Nishikawa et al., 2022; Testa et al., 2001).”
BackgroundQian, C., Di, J., & Li, J. (2022) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm55620.2022.9995234 · Cited by 4
“[18] fed an early fusion of interactional and acoustic features to the SVM classifier, and the accuracy could reach 90.”
MethodEstimation of Speech Features Using a Wearable Inertial Sensor
Du, Z., Xu, Y., Yu, X., Wang, S., & Lin, X. (2024) · Journal of Voice · doi:10.1016/j.jvoice.2024.09.012 · Cited by 3
Autofusion: Fusing Multi-Modalities with Interactions
Nguyen, T.-T., Deligianni, F., & Nguyen, H. D. (2024) · Procedia Computer Science · doi:10.1016/j.procs.2024.09.522 · Cited by 1
Lakshmi, S. G., Kumar, C., Kumar, P., Kunwar, F. B., & Singh, S. K. (2026) · Applied Computational Intelligence and Soft Computing · doi:10.1155/acis/8280381 · Cited by 1
Survey of Identification of Alzheimer’s Disease Using MRI, Speech and MMSE
Bhanusree, Y., Bulusu, D., Chinni, D., Narahari, A., Simhadri, S. S., & Bommareddy, V. (2022) · Lecture notes in networks and systems · doi:10.1007/978-981-19-4990-6_23
Multilingual early fusion approach for Alzheimer’s dementia detection, including Urdu
Hanif, A., Saleem, K., Khan, A. U., Wali, A., & Javaid, A. (2026) · International Journal of Speech Technology · doi:10.1007/s10772-026-10259-8
“In another relevant study, Nasreen et al. Nasreen et al. (2021) demonstrated the utility of linguistic features (LF) for AD detection.”
Nouveaux biomarqueurs vocaux pour la détection automatique de la somnolence
Martin, V. (2022) · theses.fr (ABES) · doi:10.70675/e8669cefz980fz4fc6zbc44z07ad3ca56810
Metadata
- Authors
- Nasreen, S., Hough, J., & Purver, M.
- Affiliations
- Queen Mary University of London (GB)
- Venue
- Interspeech 2021
- Publisher
- ISCA
- Published
- 2021-08-30
- Pages
- 1962-1966
- ISSN
- 2308-457X (ISSN-L)
- OpenAlex
- W3197699225
- Cited by
- 22
- References
- 26
- Citations by year
- 2026: 2 · 2025: 3 · 2024: 7 · 2023: 3 · 2022: 7
- FWCI
- 9.24
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Interspeech (USB) (publishedVersion)
- Cronfa (Swansea University) (submittedVersion)
- Queen Mary Research Online (Queen Mary University of London) (submittedVersion)
- Funders
- European Commission (HORIZON2020, 825153, 769661)
- Engineering and Physical Sciences Research Council (EP/S033564/1)
- Topics
- Health Sciences › Medicine › Psychiatry and Mental health › Dementia and Cognitive Impairment Research
- Physical Sciences › Computer Science › Artificial Intelligence › Emotion and Mood Recognition
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Keywords
- Alzheimer's disease detection, dementia, spontaneous speech, interactional features, acoustic features, floor control, pitch, cepstral coefficients, feature fusion, feature selection, speech-based diagnosis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.