A Spoken Language Database for Research on Moderate Cognitive Impairment: Design and Preliminary Analysis
- 713 Downloads
This paper addresses the use of spoken language technologies to identify cognitive impairment through the degree of speech deficits. We present the design of a spoken language database where patients’ voices are collected during regular clinical screening tests for cognitive impairment. Three different speaking styles are recorded: dialogues during structured interviews, readings of a short-passage and verbal picture descriptions. We hope these different spoken materials will help promoting the research on a wide range of spoken language technologies in assessing Moderate Cognitive Impairment (MCI). To illustrate this, a preliminary analysis on the speech recorded from a small group of MCI patients and healthy elder controls is also presented. A Random Forest classifier working on seven prosodic measures extracted from the reading task achieved 78.9% accuracy for MCI detection when compared with a control group, suggesting that these measures can offer a sensitive method of assessing speech output in MCI. This experimental framework shows the potential of the presented spoken language database for the research on automatic and objective identification of early symptoms of MCI in elderly adults.
Keywordsmoderate cognitive impairment spoken language database prosodic analysis
Unable to display preview. Download preview PDF.
- 1.Llibre, J.J.: Aging and dementia: implications for the scientist community, public health and Cuban society. Rev. Academia de Ciencias de Cuba 2(2), 36–54 (2012)Google Scholar
- 3.De Leon, M.J., De Santi, S., Zinkowski, R., Mehta, P.D., Pratico, D., Segal, S., Clark, C., Kerkman, D., De Bernardis, J., Li, J., Lair, L., Reisberg, B., Tsui, W., Rusinek, H.: MRI and CSF studies in the early diagnosis of Alzheimer’s disease. Journal of Internal Medicine 256, 205–223 (2004)CrossRefGoogle Scholar
- 10.Lehr, M., Prud’hommeaux, E., Shafran, I., Roark, B.: Fully automated neuropsychological assessment for detecting mild cognitive impairment. In: Interspeech 2012 (2012)Google Scholar
- 11.Hakkani-Tür, D., Vergyri, D., Tür, G.: Speech-based automated cognitive status assessment. In: INTERSPEECH, pp. 258–261 (2010)Google Scholar
- 12.Kato, S., Endo, H., Homma, A., Sakuma, T., Watanabe, K.: Early detection of cognitive impairment in the elderly based on Bayesian mining using speech prosody and cerebral blood flow activation. In: 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 5813–5816 (2013)Google Scholar
- 13.Thomas, C., Keselj, V., Cercone, N., Rockwood, K., Asp, E.: Automatic detection and rating of dementia of Alzheimer type through lexical analysis of spontaneous speech. In: 2005 IEEE International Conference on Mechatronics and Automation, vol. 3, pp. 1569–1574 (2005)Google Scholar
- 15.Rochford, I., Rapcan, V., D’Arcy, S., Reilly, R.B.: Dynamic minimum pause threshold estimation for speech analysis in studies of cognitive function in ageing. In: 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 3700–3703 (2012)Google Scholar
- 17.Darley, F.L., Aronson, A.E., Brown, J.R.: Motor speech disorders, 3rd edn. W.B. Saunders Company, Philadelphia (1975)Google Scholar
- 19.Mertens, P.: Automatic segmentation of speech into syllables. In: Laver, J., Jack, M. (eds.) Proceedings of the European Conference on Speech Technology, Edinburgh, vol. 2, pp. 9–12 (1987)Google Scholar