Identification of Mild Cognitive Impairment From Speech in Swedish Using Deep Sequential Neural Networks

Abstract

While people with mild cognitive impairment (MCI) portray noticeably incipient memory difficulty in remembering events and situations along with problems in decision making, planning, and finding their way in familiar environments, detailed neuropsychological assessments also indicate deficits in language performance. To this day, there is no cure for dementia but early-stage treatment can delay the progression of MCI; thus, the development of valid tools for identifying early cognitive changes is of great importance. In this study, we provide an automated machine learning method, using Deep Neural Network Architectures, that aims to identify MCI. Speech materials were obtained using a reading task during evaluation sessions, as part of the Gothenburg MCI research study. Measures of vowel duration, vowel formants (F1 to F5), and fundamental frequency were calculated from speech signals. To learn the acoustic characteristics associated with MCI vs. healthy controls, we have trained and evaluated ten Deep Neural Network Architectures and measured how accurately they can diagnose participants that are unknown to the model. We evaluated the models using two evaluation tasks: a 5-fold crossvalidation and by splitting the data into 90% training and 10% evaluation set. The findings suggest first, that the acoustic features provide significant information for the identification of MCI; second, the best Deep Neural Network Architectures can classify MCI and healthy controls with high classification accuracy (M = 83%); and third, the model has the potential to offer higher accuracy than 84% if trained with more data (cf., SD≈15%). The Deep Neural Network Architecture proposed here constitutes a method that contributes to the early diagnosis of cognitive decline, quantify the progression of the condition, and enable suitable therapeutics.

Feature Selection and Combination of Information in the Functional Brain Connectome for Discrimination of Mild Cognitive Impairment and Analyses of Altered Brain Patterns

“…In recent years, machine learning approaches with datadriven algorithms have been used to combine and classify brain features. Some classifiers such as support vector machines (SVMs) (Prasad et al, 2015;Khazaee et al, 2016), Naïve Bayes (Zhuo et al, 2018) and deep neural networks (Themistocleous et al, 2018) are applied to discriminate normal controls from subjects with MCI. However, most of these methods focus on a single modality of imaging, the functional connectome, or graph theory attributes separately, resulting in relatively poor classification performance (Suk et al, 2014).…”

Neurodegeneration and the Gothenburg MCI study

“…The collaboration with Dr. Dimitrios Kokkinakis and colleagues within the project 'Linguistic and extra-linguistic parameters for early detection of cognitive impairment' (Kokkinakis et al, 2016) is an important next step to further refine the study of cognitive functioning in relation to dementia progression, using novel technical and analytical tools other than traditional neuropsychological testing; see for instance Fraser et al (2017); Themistocleous et al (2018) and Linz et al, (2019).…”

Analyze Informant-Based Questionnaire for The Early Diagnosis of Senile Dementia Using Deep Learning

“…In recent years, deep neural network (DNN) has garnered clinical interests in cognitive diagnostic applications due to its advantages in efficient classification. Moreover, many existing methods have been proposed, where some of them [4]- [7] combine DNN with neuroimaging markers, while other methods [8]- [10] combine DNN with neuropsychological assessments. Jain et al [5] proposed a transfer learning approach for accurately classifying brain sMRI slices amongst 3 different classes: Alzheimer's disease (AD), cognitively normal(CN) and mild cognitive impairment (MCI).…”

“…The experimental results showed that the model could accurately predict MCI and AD type dementia on a very sparse clinical language dataset. Themistocleous et al [10] provided an automated deep learning method using DNN architectures that identified individuals with MCI from healthy controls. However, there are still limitations in the current studies.…”