Longitudinal detection of dementia through Introduction Motivation - PowerPoint PPT Presentation

Le et al. (2011): Detection of Dementia Daniela Stier Longitudinal detection of dementia through Introduction Motivation & lexical and syntactic changes in writing: a Background Approach Language in Ageing & case study of three British novelists Dementia Analysis by Le Le et al. (2011) et al. (2011) Lexical Analysis Vocabulary size Lexical repetition Word specificity Word-class deficits Daniela Stier Fillers Syntactic Analysis MLU & MCU Parse tree depth University of T¨ ubingen D-Level Passive voice daniela.stier@student.uni-tuebingen.de Summary & Conclusion December 2, 2015 References 1 / 30

Le et al. (2011): Table of Contents Detection of Dementia Introduction Daniela Stier Motivation & Background Introduction Approach Motivation & Background Language in Ageing & Dementia Approach Language in Ageing & Dementia Analysis by Le et al. (2011) Analysis by Le et al. (2011) Lexical Analysis Lexical Analysis Vocabulary size Vocabulary size Lexical repetition Lexical repetition Word specificity Word-class deficits Word specificity Fillers Word-class deficits Syntactic Analysis MLU & MCU Fillers Parse tree depth D-Level Syntactic Analysis Passive voice MLU & MCU Summary & Conclusion Parse tree depth References D-Level Passive voice Summary & Conclusion 2 / 30

Le et al. (2011): Motivation Detection of Dementia Daniela Stier Introduction Motivation & Background I Alzheimer’s disease (AD) among most prevalent Approach Language in Ageing & geriatric conditions Dementia Analysis by Le I definite diagnosis only post mortem et al. (2011) Lexical Analysis I no proven cure for dementia Vocabulary size Lexical repetition Word specificity Word-class deficits Fillers Syntactic Analysis → correct, timely diagnosis is of great importance MLU & MCU Parse tree depth D-Level → su ffi ciently early diagnosis of AD may even make Passive voice prevention possible in future Summary & Conclusion References 3 / 30

Le et al. (2011): Background Detection of Dementia Daniela Stier The Alzheimer’s pathology likely begins many Introduction Motivation & years and perhaps decades before the onset of Background Approach symptoms; therefore, there is an opportunity for Language in Ageing & Dementia prevention once future advances make it possible Analysis by Le et al. (2011) to diagnose the disease through the use of Lexical Analysis Vocabulary size biomarkers before symptom onset. Lexical repetition Word specificity (Blazer et al., 2004, p. 249) Word-class deficits Fillers Syntactic Analysis I early diagnosis through linguistic analysis : MLU & MCU Parse tree depth D-Level I a ff ecting linguistic abilities in speech and writing Passive voice I possibility to develop techniques, e.g. looking for Summary & Conclusion diachronic changes in patients’ writings References I problem: how to get a lifelong corpus of writing? 4 / 30

Le et al. (2011): Approach Detection of Dementia Daniela Stier Introduction I study: large-scale longitudinal study of lexical and Motivation & Background syntactic changes in language in Alzheimer’s disease Approach Language in Ageing & Dementia (AD) Analysis by Le et al. (2011) I corpus: complete, fully parsed texts by three authors Lexical Analysis Vocabulary size Lexical repetition I hypothesis: Word specificity Word-class deficits Fillers → signs of dementia can be found in diachronic analyses of − Syntactic Analysis patients’ writings MLU & MCU Parse tree depth → lead to a new understanding of the work of the − D-Level Passive voice individual authors Summary & I related: Williams et al. (2003), Garrard et al. (2005) 1 Conclusion References 1 Please refer to Le (2010) for more details. 5 / 30

Le et al. (2011): Material Detection of Dementia Daniela Stier Introduction Motivation & Background Approach Language in Ageing & Dementia Analysis by Le et al. (2011) P. D. James Agatha Christie Iris Murdoch Lexical Analysis (*1920 - † 2014) 2 (*1890 - † 1976) 3 (*1919 - † 1999) 4 Vocabulary size Lexical repetition Word specificity Word-class deficits aged healthily suspected of died with Fillers Syntactic Analysis Alzheimer’s Alzheimer’s MLU & MCU Parse tree depth D-Level Passive voice Summary & Conclusion 2Extracted from http://www.independent.co.uk/news/people/ References pd-james-dead-detective-novelist-behind-death-comes-to-pemberley-and-children-of-men-dies-aged-94-9887573. html (last access: 11/30/2015). 3Extracted from http://www.niederdeutschebuehne.de/agatha-christie/ (last access: 11/30/2015). 4Extracted from http://www.theguardian.com/commentisfree/2009/jun/26/iris-murdoch (last access: 11/30/2015). 6 / 30

Le et al. (2011): Material Detection of Dementia Daniela Stier Introduction Motivation & P. D. James Agatha Christie Iris Murdoch Background Approach Language in Ageing & Dementia Expectation : Analysis by Le et al. (2011) James data will exhibit linguistic patterns of healthy ageing Lexical Analysis Vocabulary size Christie data will exhibit linguistic patterns similar to those of Lexical repetition Word specificity AD patients Word-class deficits Fillers Murdoch data will exhibit linguistic patterns of dementia patients Syntactic Analysis MLU & MCU Parse tree depth D-Level Passive voice Assumption: Summary & I no novel departs from the usual writing methodology of Conclusion References its author, belongs to an atypical genre or involves research to the degree that it should be judged an outlier 7 / 30

Le et al. (2011): Language in Ageing and Dementia Detection of Dementia I consensus: decline occurring in normal ageing is Daniela Stier accelerated in presence of AD Introduction Motivation & I AD: deficits in lexical features may be more prominent Background Approach than in syntactic ones Language in Ageing & Dementia Analysis by Le et al. (2011) Lexical Analysis Vocabulary size Lexical repetition Word specificity Word-class deficits Fillers Syntactic Analysis MLU & MCU Parse tree depth D-Level Passive voice Summary & Conclusion References Figure: Expected patterns of linguistic changes 8 / 30

Le et al. (2011): Lexical Analysis Detection of Dementia Daniela Stier Variety of measures for lexical markers : Introduction Text length a Vocabulary size Motivation & Background Approach measures sensitive to length: b Lexical repetition Language in Ageing & Dementia → threshold: 55.000 tokens c Word specificity Analysis by Le et al. (2011) d Word-class deficits Lexical Analysis for remaining measures: Vocabulary size Lexical repetition e Fillers complete text of all novels Word specificity Word-class deficits Fillers Syntactic Analysis MLU & MCU Changes over time: simple linear regression of the Parse tree depth D-Level Passive voice respective measure against the author’s age Summary & Conclusion Statistical significance: relationship between the author’s References age and the value of the respective measure Spearman correlation: correlation between measures ! 9 / 30

Le et al. (2011): Lexical Analysis Detection of Dementia a) Vocabulary size Daniela Stier Introduction I TTR: type/token ratio Motivation & Background I number of unique lemmatized word-types divided by Approach Language in Ageing & Dementia total number of word-tokens Analysis by Le et al. (2011) Lexical Analysis Vocabulary size Lexical repetition I M: rates drop - rise - Word specificity Word-class deficits drop, insignificant Fillers Syntactic Analysis MLU & MCU I C: rates vary, drop, Parse tree depth D-Level significant Passive voice I J: slight rising trend, Summary & Conclusion insignificant References Figure: Type/token ratio within the first 55.000 tokens 10 / 30

Le et al. (2011): Lexical Analysis Detection of Dementia a) Vocabulary size Daniela Stier Introduction I WTIR: word-type introduction rate Motivation & Background I cumulative number of unique lemmatized types, Approach Language in Ageing & Dementia computed at 10.000-token interval Analysis by Le et al. (2011) Lexical Analysis Vocabulary size Lexical repetition Word specificity Word-class deficits Fillers Syntactic Analysis MLU & MCU Parse tree depth D-Level Passive voice Summary & (a) I. Murdoch (b) A. Christie (c) P. D. James Conclusion References Figure: Word-type introduction rate up to the 70.000th token 11 / 30

Le et al. (2011): Lexical Analysis Detection of Dementia b) Lexical repetition Daniela Stier I global word n-gram repetitions Introduction Motivation & I i.e. 2-11 words, occurring at least twice Background Approach I maximals : longest repeating phrases in a text Language in Ageing & Dementia Analysis by Le I associates : substrings of maximals occurring more et al. (2011) frequently than maximals Lexical Analysis Vocabulary size Lexical repetition Word specificity Word-class deficits I M: rise and peak in last Fillers Syntactic Analysis novels, overall decrease, MLU & MCU Parse tree depth insignificant D-Level Passive voice I C: overall increase, Summary & Conclusion significant References I J: overall decrease, Figure: Maximal and associate insignificant phrasal repetitions (types) 12 / 30