Improving Detection of Alzheimer's Disease Using Automatic Speech Recognition to Identify High-Quality Segments for More Robust Feature Extraction

  title={Improving Detection of Alzheimer's Disease Using Automatic Speech Recognition to Identify High-Quality Segments for More Robust Feature Extraction},
  author={Yilin Pan and Bahman Mirheidari and Markus Reuber and Annalena Venneri and Daniel J. Blackburn and Heidi Christensen},
Speech and language based automatic dementia detection is of interest due to it being non-invasive, low-cost and potentially able to aid diagnosis accuracy. The collected data are mostly audio recordings of spoken language and these can be used directly for acoustic-based analysis. To extract linguistic-based information, an automatic speech recognition (ASR) system is used to generate transcriptions. However, the extraction of reliable acoustic features is difficult when the acoustic quality… 

Figures and Tables from this paper

Acoustic Feature Extraction with Interpretable Deep Neural Network for Neurodegenerative Related Disorder Classification
This paper proposes and evaluates the Sinc-CLA (with SincNet, Convolutional, Long Short-Term Memory and Attention layers) as a task-driven acoustic feature extractor for classifying MCI, ND and healthy controls and finds the learned task- driven features achieve a superior classification accuracy.
Modular Multi-Modal Attention Network for Alzheimer's Disease Detection Using Patient Audio and Language Data
This work proposes a modular multi-modal architecture to automatically detect Alzheimer’s disease using the dataset provided in the ADReSSo challenge and shows that the C-Attention-Unified network with Linguistic features and X-Vector embeddings achieves the best accuracy.
Alzheimer's Disease Detection from Spontaneous Speech through Combining Linguistic Complexity and (Dis)Fluency Features with Pretrained Language Models
This paper combines linguistic complexity and (dis)fluency features with pretrained language models for the task of Alzheimer's disease detection of the 2021 ADReSSo (Alzheimer's Dementia Recognition through Spontaneous Speech) challenge and shows that the best-performing model that integrated component models using a stacking ensemble technique performed equally well on cross-validation and test data, indicating that it is robust against overfitting.
Articulatory Coordination for Speech Motor Tracking in Huntington Disease
An experiment is presented that uses Vocal Tract Coordination (VTC) features extracted from read speech to estimate a motor score and finds that VTC features significantly outperform other acoustic features across varied-length audio segments, which highlights the effectiveness of these features for both short and long-form reading tasks.


Automatic Hierarchical Attention Neural Network for Detecting AD
This paper proposes a novel model that encompasses both the hierarchical and sequential structure of the description and detect its informative units by attention mechanism, and demonstrates that the use of hierarchical models with attention mechanism improves the AD/HC discrimination performance.
Speech-based automatic and robust detection of very early dementia
There is evidence to the potential use of simple spoken tasks for automatic assessment of very early dementia, and a new data regularization technique to overcome data sparseness due to the limited data set size is presented.
Automatic detection of mild cognitive impairment from spontaneous speech using ASR
This work automates the extraction of the features of Mild Cognitive Impairment by applying automatic speech recognition (ASR), and uses machine learning methods to separate the subjects with MCI from the control group.
Evaluation of speech-based protocol for detection of early-stage dementia
This analysis is based on recordings of over 80 diagnosed subjects; it yields dementia and MCI detection equal-error-rate below 20%, and demonstrates the high value of using speech and voice analysis for automatic screening and status tracking of dementia from the very early stage of MCI.
Aided diagnosis of dementia type through computer-based analysis of spontaneous speech
This pilot study evaluates the ability of machined learned algorithms to assist with the differential diagnosis of dementia subtypes based on brief (< 10 min) spontaneous speech samples and suggests this clinical speech analytic approach offers promise as an additional, objective and easily obtained source of diagnostic information for clinicians.
Detecting Alzheimer's Disease Using Gated Convolutional Neural Network from Audio Data
We propose an automatic detection method of Alzheimer's diseases using a gated convolutional neural network (GCNN) from speech data. This GCNN can be trained with a relatively small amount of data
Linguistic Features Identify Alzheimer's Disease in Narrative Speech.
BACKGROUND Although memory impairment is the main symptom of Alzheimer's disease (AD), language impairment can be an important marker. Relatively few studies of language in AD quantify the
Longitudinal Monitoring and Detection of Alzheimer's Type Dementia from Spontaneous Speech Data
  • S. Luz
  • Computer Science
    2017 IEEE 30th International Symposium on Computer-Based Medical Systems (CBMS)
  • 2017
Tests of the proposed method on a data set of spontaneous speech recordings of Alzheimers patients and elderly controls show that accuracy of 68% can be achieved with a Bayesian classifier operating on features extracted through simple algorithms for voice activity detection and speech rate tracking.
Improving Cognitive Impairment Classification by Generative Neural Network-Based Feature Augmentation
This paper uses a variational autoencoder to augment data at the feature-level as opposed to the speech signal-level, and investigates whether this suits some feature types better than others.
Speaking in Alzheimer’s Disease, is That an Early Sign? Importance of Changes in Language Abilities in Alzheimer’s Disease
The purpose of this article is to summarize the relation between prodromal and manifest AD and language functions and language domains and claim that AD can be more sensitively detected with the help of a linguistic analysis than with other cognitive examinations.