Raum:
Saal New York 2
Topic:
Wissenschaftliches Programm
Topic 15: Diagnostik und Klassifikation
Topic 01: Neurokognitive Erkrankungen, organische psychische Störungen, Demenz, F0
English programme
Format:
Symposium
Dauer:
90 Minuten
Besonderheiten:
Q&A-Funktion
Automated speech analysis enables a valid and reliable investigation of speech disturbances, which present as a common though challenging clinical problem, particularly in older age. Disrupting the ability to communicate affects a person’s interaction with the social environment, leading to alterations in emotional stability as well as quality of life. An early and accurate diagnosis of speech deficiencies is of clinical relevance, particularly when planning possible interventions and monitoring their outcome. To date, however, clinicians face insufficiently objective manual techniques when analysing tests of speech and language functions. Besides its lack of objectivity, manual annotation is time-consuming, less sensitive for subtle impairment, and not able to cover the emotional character of speech. This session will introduce a more feasible approach for the detection of cognitive and affective alterations in several clinical populations.
Niklas Linz will give an overview on how automated speech analysis works. Jessica Peter will present a more fine-grained analysis of naming errors through automated speech analysis in patients with mild cognitive impairment. Alexandra König will talk about detecting apathy with speech-based analyses in clinical samples of older adults. Finally, Caroline Kuhn (PhD) will present speech-based diagnostics of cognitive changes in patients with Multiple Sclerosis.
Detecting apathy in older adults with cognitive disorders using automated speech analysis
Radia Zeghari, Nizza (France)
Details anzeigen
Autor:in:
Radia Zeghari, Nizza (France)
Apathy is frequent in several psychiatric and neurological conditions and has been found to have a severe negative effect on disease progression. In older people, it can be a predictor of increased dementia risk. Current assessment methods seem insufficiently objective and sensitive, thus new diagnostic tools and broad-scale screening technologies are needed. This study is the first of its kind aiming to investigate whether automatic speech analysis and computational linguistic could be used for characterization and detection of apathy. For this, apathetic and non-apathetic patients (n = 60) with mild to moderate neurocognitive disorder were recorded while performing two short narrative speech tasks. Paralinguistic features relating to prosodic, formant, source and temporal qualities of speech as well as linguistic features such as semantic content and affective valence are automatically extracted, examined between the groups and compared to healthy control subjects. Machine learning experiments are carried out to validate the diagnostic power of different speech and language markers. Correlations between apathy sub-scales and features revealed a relation between temporal aspects of speech and the subdomains of reduction in interest and initiative, as well as between prosody features and the affective domain. Group differences were found to vary for males and females, depending on the task. Differences in temporal aspects of speech were found to be the most consistent difference between apathetic and non-apathetic patients. Machine learning models trained on speech features achieved top performances of AUC = 0.88 for males and AUC = 0.77 for females. Significant reductions in the number of words, the magnitude of sentiment, the overall sentiment and the range between sentiment in the positive and negative question are found to differ for the apathetic population. These findings reinforce the usability of speech as a reliable biomarker in the detection and assessment
Automated qualitative analysis of errors during confrontational naming in mild cognitive impairment
Jessica Peter, Bern (Switzerland)
Details anzeigen
Autor:in:
Jessica Peter, Bern (Switzerland)
Errors during confrontational naming are common in dementia due to Alzheimer’s disease as well as its precursor amnestic mild cognitive impairment (aMCI) and may indicate dysfunction of visual perception, semantic processing, or word retrieval. We investigated whether automated analysis of these errors is superior to standard manual analysis in revealing differences between patients with aMCI and healthy controls (HC).
We administered a German version of the Graded Naming Test (30-Items) in 25 patients with aMCI and 25 HC. We automatically transcribed responses using automated speech recognition (ASR) and compared the results to standard manual transcription. For the computational approach, we computed word frequencies of missed target words on the basis of large reference corpora. We will additionally calculate the visual, phonetic, and semantic distance of the error word to the target word.
We found no significant difference between patients with aMCI and HC for the total error count as revealed by manual analysis. With computational analysis, we found that missed target objects had a significantly higher word frequency in patients with aMCI compared to HC. The analyses of visual, semantic, and phonetic distances are still ongoing but we expect to find significant differences, too.
Speech based diagnostics of cognitive changes in Multiple Sclerosis: a clinical approach
Caroline Kuhn, Saarbrücken (Germany)
Details anzeigen
Autor:in:
Caroline Kuhn, Saarbrücken (Germany)
Cognitive changes occur in the vast majority of multiple sclerosis patients. Even if individuals differ from one another on their disease courses, changes within the range of high-level brain functions are what they share. Information processing, attention and concentration, memory and executive functions and in particular verbal fluency are the most vulnerable neuropsychological functions.
Although patients suffer from persistent cognitive dysfunctions there is often a lack of corresponding lesion areas on MRI. As verbal fluency is a key function reflecting cognitive deficits in real time, our psychometric assessment uses speech based neuropsychological tasks in order to ensure an early recognition of cognitive changes.