Linguistic Parameters of Spontaneous Speech for Identifying Mild Cognitive Impairment and Alzheimer Disease

Abstract

In this paper, we seek to automatically identify Hungarian patients suffering from mild cognitive impairment (MCI) or mild Alzheimer’s Disease (mAD) based on their speech transcripts, focusing only on linguistic features. In addition to the features examined in our earlier study, we introduce syntactic, semantic and pragmatic features of spontaneous speech that might affect the detection of dementia. In order to ascertain the most useful features for distinguishing healthy controls, MCI patients and mAD patients, we will carry out a statistical analysis of the data and investigate the significance level of the extracted features among various speaker group pairs and for various speaking tasks. In the second part of the paper, we use this rich feature set as a basis for an effective discrimination among the three speaker groups. In our machine learning experiments, we will analyze the efficacy of each feature group separately. Our model which uses all the features achieves competitive scores, either with or without demographic information (3-class accuracy values: 68–70%, 2-class accuracy values: 77.3–80%). We also analyze how different data recording scenarios affect linguistic features and how they can be productively used when distinguishing MCI patients from healthy controls.

Automated Evaluation of Standardized Dementia Screening Tests

“…Free recall tasks of visual material, such as the CTPD, have the advantage that spontaneous speech is elicited and retrieved selectively, making it more contained and thus easier to process. The same is true for elicited speech based on free recall tasks from moving images, such as short films [13]. Most work uses either fluency [14,15] or deep speech markers [7] for classification, as these show high variance in distinguishing patients with cognitive impairment from healthy controls.…”

SpEx: a German-language dataset of speech and executive function performance

“…Such formal tests tend to be of a controlled experimental nature, allowing for an easy extraction of variables of interest, focusing mostly on number of correct responses or errors, and reaction times. However, within the last few years qualitative analyses and related objective parameters were shown to provide deeper insights into the complex involvement of cognitive processes 19 , 20 . Such parameters can be extracted from tasks that allow participants more freedom in the speech that they produce due to less experimental control.…”