Automatic identification of Mild Cognitive Impairment through the analysis of Italian spontaneous speech productions

Automatic Identification of Mild Cognitive Impairment through the Analysis of Italian Spontaneous Speech Productions Daniela Beltrami1,2, Laura Calzà1, Gloria Gagliardi1, Enrico Ghidoni2, Norina Marcello2, Rema Rossini Favretti1, Fabio Tamburini1 1 Alma Mater Studiorum - Università di Bologna 2 Arcispedale S.Maria Nuova, Reggio Emilia E-mail: daniela.beltrami3@unibo.it, laura.calza@unibo.it, gloria.gagliardi2@unibo.it, Enrico.Ghidoni@asmn.re.it, Norina.Marcello@asmn.re.it, rema.rossini@unibo.it, fabio.tamburini@unibo.it Abstract This paper presents some preliminary results of the OPLON project. It aimed at identifying early linguistic symptoms of cognitive decline in the elderly. This pilot study was conducted on a corpus composed of spontaneous speech sample collected from 39 subjects, who underwent a neuropsychological screening for visuo-spatial abilities, memory, language, executive functions and attention. A rich set of linguistic features was extracted from the digitalised utterances (at phonetic, suprasegmental, lexical, morphological and syntactic levels) and the statistical significance in pinpointing the pathological process was measured. Our results show remarkable trends for what concerns both the linguistic traits selection and the automatic classifiers building. Keywords: Pathological language, Mild Cognitive Impairment, Linguistic Features, Automatic classifiers. 1. Background This research is part of the OPLON project (“OPportunities for active and healthy LONgevity”, Smart Cities and Communities – DD 391/RIC, co-funded by the Ministry of Education as part of the Contract “Smart Cities and Communities and Social Innovation”). The project intends to propose actions and methods to prevent fragility and decline and promote the health of the elderly, designing and developing tools and networks of early diagnosis and “care & cure”. Given this general project framework, the prevention of the various types of dementia appears to be one of the most challenging, nonetheless most pressing, tasks (Calzà et al., 2015). Individuals with preclinical dementia manifest alterations in various cognitive domains: a number of longitudinal retrospective studies have already demonstrated that linguistic features could act as a prodromic marker of cognitive dysfunctions: for example, the Nun study (Snowdon, 2003), the Iris Murdoch study (Garrard et al., 2005) or the Harold Wilson project (Garrard, 2009). Deficits are seen in verbal fluency, naming and semantic knowledge (Taler & Phillips, 2008); it is also well documented that discourse alterations may be one of the earliest signs of the pathology, often measurable years before other cognitive deficits become apparent (Caramelli et al., 1998). Looking at the literature on this topic, syntactic and phonological abilities seem to be relatively preserved, even though individuals produce semantically impoverished discourse that lacks in coherence. These linguistic complaints are definitely concomitant with neuropathological alterations and clinical manifestation, but also recognizable in the presymptomatic phases of the cognitive impairment. The investigation of this domain seems to be promising, both for early diagnosis and dementia large-scale screenings. During the last few years, the development of new sophisticated techniques from Natural Language Processing (NLP) have been used to analyse written texts, clinically elicited utterances and spontaneous production, in order to identify signs of psychiatric or neurological disorders and to extract automatically derived linguistic features for pathologies recognition, classification and description. Computational methods have been already successfully applied to the study of linguistic cues of cerebral …

Content truncated to 500 words