Figure 1 - uploaded by Kristina Lundholm Fors
Content may be subject to copyright.
The Cookie-theft picture.
Source publication
Cognitive and mental deterioration, such as difficulties with memory and language, are some of the typical phenotypes for most neurodegenerative diseases including Alzheimer's disease and other dementia forms. This paper describes the first phases of a project that aims at collecting various types of cognitive data, acquired from human subjects in...
Context in source publication
Context 1
... the acquisition of the audio signal we use the Cookie-theft picture 1 (see Figure 1) from the Boston Diagnostic Aphasia Examination (BDAE; Goodglass & Kaplan, 1983) which is often used to elicit speech from people with various mental and cognitive impairments. During the presentation of the Cookie-theft stimuli (which illustrates an event taking place in a kitchen) the subjects are asked to tell a story about the picture and describe everything that can be observed while the story is recorded. ...
Similar publications
The current study investigated a novel visual distracter task as a potential diagnostic marker for the detection of cognitive impairment and the extent to which this compares in healthy ageing across two cultures. The Inhibition of a Recent Distracter Effect (IRD) refers to the inhibition of a saccadic eye movement towards a target that is presente...
Citations
... Speech analysis is currently proposed as an essential and useful low-cost tool for diagnosis and monitoring dementia progression [15][16][17][18][19]. Different procedures have been used to elicit spontaneous speech in samples of persons with MCI and AD: the "Cookie Theft" picture description [16,18,20], oral descriptions of common objects [21], semi-structured interview format [15,22], and verbal picture descriptions [23]. ...
... -B -this index was included because, unlike TTR, it quantifies lexical richness without being sensitive to text length. Its range is [10,20], and the lower the value, the richer the speech. ...
Background
Based on recent researches on the use of natural language processing techniques for very early detection of cognitive decline and the benefits of cognitive stimulation for people with cognitive impairments, the Dementia Monitoring application (DMapp) is developed inside the Memento project.AimsThe aims of this work are: (1) to present DMapp; (2) to report the results of two preliminary studies on DMapp; (3) to describe the clinical and experimental potentiality of DMapp.Methods
Italian people with the diagnosis of mild cognitive impairment due to Alzheimer’s disease or dementia due to Alzheimer’s Disease with a Mini-Mental-State-Examination between 24 and 28 (inclusive) were involved in the DMapp development prototype during the Lab Trial (4 subjects) and Filed Trial (5 subjects) of the Memento project. Qualitative and quantitative data were collected to evaluate participants’ opinions, the DMapp ability to perform the automatic analysis of the speech and participants’ visible emotional state effective. Ad hoc interviews, the Observed Emotion Rating Scale and performance metrics to solve different tasks were used. The relation between cognitive measures (global cognitive measures) and linguistic indexes values was considered using Euclidean distances between the participants.ResultsLinguistic indexes were calculated and seemed to classify the participants’ performance as expected from cognitive measures. The DMapp was appreciated by people with cognitive impairment. Positive emotions were present.ConclusionDMapp seems an interesting approach to monitor dementia symptoms progression and stimulate memory. Possible developments and open questions are discussed.
... The Gothenburg MCI Study has been approved by the local ethical committee review board (reference number: L091-99, 1999; T479-11, 2011); while the currently described study has been approved with reference number 2016). A total of 31 MCI patients and 36 healthy controls (HC) were recruited according to detailed inclusion and exclusion criteria, the details of which can be found in Kokkinakis et al. (2017). Briefly, these criteria specified that all participants should be in the age range 50-79 years, speak Swedish as their first language and as the only language before the age of 5 years, and should have recent neuropsychological test results available as a result of their participation in the Gothenburg MCI Study. ...
Recent work has indicated the potential utility of automated language analysis for the detection of mild cognitive impairment (MCI). Most studies combining language processing and machine learning for the prediction of MCI focus on a single language task; here, we consider a cascaded approach to combine data from multiple language tasks. A cohort of 26 MCI participants and 29 healthy controls completed three language tasks: picture description, reading silently, and reading aloud. Information from each task is captured through different modes (audio, text, eye-tracking, and comprehension questions). Features are extracted from each mode, and used to train a series of cascaded classifiers which output predictions at the level of features, modes, tasks, and finally at the overall session level. The best classification result is achieved through combining the data at the task level (AUC = 0.88, accuracy = 0.83). This outperforms a classifier trained on neuropsychological test scores (AUC = 0.75, accuracy = 0.65) as well as the “early fusion” approach to multimodal classification (AUC = 0.79, accuracy = 0.70). By combining the predictions from the multimodal language classifier and the neuropsychological classifier, this result can be further improved to AUC = 0.90 and accuracy = 0.84. In a correlation analysis, language classifier predictions are found to be moderately correlated (ρ = 0.42) with participant scores on the Rey Auditory Verbal Learning Test (RAVLT). The cascaded approach for multimodal classification improves both system performance and interpretability. This modular architecture can be easily generalized to incorporate different types of classifiers as well as other heterogeneous sources of data (imaging, metabolic, etc.).
... Such visual impairments characteristic of vascular dementia are primarily rooted in psychomotor slowing, not a specific deficit to the search methods [38]. In terms of analysis, eye-tracking data, algorithms to detect fixations, saccades (rapid switching between two fixations) and backtracking (moving "backwards" to previous fixations) are of particular interest because this enumerates the participation's attention and thus best measure the underlying cognitive impairments associated with neurodegenerative disorders that may be present [39]. ...
Vascular Dementia is often Clinically diagnosed once the effects of the disease are prevalent in a person's daily living routines. However, previous research has shown various behavioral and physiological changes linked to the development of Vascular Dementia, with these changes beginning to present earlier than clinical diagnosis is currently possible. In this review, works focused on these early signs of Vascular Dementia are highlighted. However, recognizing these changes is difficult. Many computational systems have been proposed for the evaluation these early signs of Vascular Dementia. The chosen works have largely focused on utilizing sensors systems or algorithmic evaluation can be incorporated into a person's environment to measure behavioral, and phycological metrics. This raw data can then be computationally analyzed to draw conclusions about the patterns of change surrounding the onset of Vascular Dementia. This compilation of works presents current a framework for investigating the various behavioral and physiological metrics as well as potential avenues for further investigating of sensor system and algorithmic design with the goal of enabling earlier Vascular Dementia detection.
... All the participants in the current study on "Linguistic and extra-linguistic parameters for early detection of cognitive impairment" were recruited from the Gothenburg MCI study . All participants were speakers of Swedish, selected according to detailed inclusion and exclusion criteria (Kokkinakis et al., 2017). Data collection took place in a quiet lab environment where participants were fitted with a lapel microphone (Au-dioTechnica ATR3350) and digitally recorded with a Zoom H4n Pro recorder (44.1 kHz sampling rate; 16bit resolution). ...
The Semantic Verbal Fluency (SVF) task is a classical neuropsychological assessment where persons are asked to produce words belonging to a semantic category (e.g., animals) in a given time. This paper introduces a novel method of temporal analysis for SVF tasks utilizing time intervals and applies it to a corpus of elderly Swedish subjects (mild cognitive impairment, subjective cognitive impairment and healthy controls). A general decline in word count and lexical frequency over the course of the task is revealed, as well as an increase in word transition times. Persons with subjective cognitive impairment had a higher word count during the last intervals, but produced words of the same lexical frequencies. Persons with MCI had a steeper decline in both word count and lexical frequencies during the third interval. Additional correlations with neuropsychological scores suggest these findings are linked to a person's overall vocabulary size and processing speed, respectively. Classification results improved when adding the novel features (AUC = 0.72), supporting their diagnostic value.
... A total of 31 MCI patients and 36 healthy controls were included in the present study, according to detailed inclusion and exclusion criteria (Kokkinakis et al., 2017). The participants in the current study all provided written informed consent. ...
We analyze the information content of narrative speech samples from individuals with mild cognitive impairment (MCI), in both English and Swedish, using a combination of supervised and unsupervised learning techniques. We extract information units using topic models trained on word embeddings in monolingual and multilingual spaces, and find that the multilingual approach leads to significantly better classification accuracies than training on the target language alone. In many cases, we find that augmenting the topic model training corpus with additional clinical data from a different language is more effective than training on additional monolingual data from healthy controls. Ultimately we are able to distinguish MCI speakers from healthy older adults with accuracies of up to 63% (English) and 72% (Swedish) on the basis of information content alone. We also compare our method against previous results measuring information content in Alzheimer's disease, and report an improvement over other topic-modeling approaches. Furthermore, our results support the hypothesis that subtle differences in language can be detected in narrative speech, even at the very early stages of cognitive decline, when scores on screening tools such as the Mini-Mental State Exam are still in the “normal” range.
... The Gothenburg MCI Study is approved by the local ethical committee review board (reference number: L091-99, 1999; T479-11, 2011); while the currently described study is approved by the local ethical committee (decision 206-16, 2016). A total of 31 MCI patients and 36 healthy controls were included in the present study, according to detailed inclusion and exclusion criteria (Kokkinakis et al., 2017). One control participant was excluded from the current analysis because their MMSE score was not available. ...
The Mini-Mental State Exam (MMSE) is a screening tool for cognitive impairment. It has been extensively validated and is widely used, but has been criticized as not being effective in detecting mild cognitive impairment (MCI). In this study, we examine the utility of augmenting MMSE scores with automatically extracted linguistic information from a narrative speech task to better differentiate between individuals with MCI and healthy controls in a Swedish population. We find that with the addition of just four linguistic features, the AUC score (measuring a trade-off between sensitivity and specificity) is improved from 0.68 to 0.87 in logistic regression classification. These preliminary results suggest that the accuracy of traditional screening tools may be improved through the addition of computerized language analysis.
... The Swedish study will extend the speech analysis with eye movement analysis of the patients and other cognitive markers. The planned work is described in Kokkinakis et al. (2017). ...
This open access book describes the results of natural language processing and machine learning methods applied to clinical text from electronic patient records.
It is divided into twelve chapters. Chapters 1-4 discuss the history and background of the original paper-based patient records, their purpose, and how they are written and structured. These initial chapters do not require any technical or medical background knowledge. The remaining eight chapters are more technical in nature and describe various medical classifications and terminologies such as ICD diagnosis codes, SNOMED CT, MeSH, UMLS, and ATC. Chapters 5-10 cover basic tools for natural language processing and information retrieval, and how to apply them to clinical text. The difference between rule-based and machine learning-based methods, as well as between supervised and unsupervised machine learning methods, are also explained. Next, ethical concerns regarding the use of sensitive patient records for research purposes are discussed, including methods for de-identifying electronic patient records and safely storing patient records. The book’s closing chapters present a number of applications in clinical text mining and summarise the lessons learned from the previous chapters.
The book provides a comprehensive overview of technical issues arising in clinical text mining, and offers a valuable guide for advanced students in health informatics, computational linguistics, and information retrieval, and for researchers entering these fields.
