Abstract:To evaluate the speech characteristics in patients with amnestic mild cognitive impairment (aMCI) and early-stage Alzheimer’s disease (AD) assisted with computerized software for speech assessment. 17 AD patients, 16 aMCI patients and 20 healthy elderly controls were recruited from Shanghai Ruijin Hosptial. Each subject underwent the neuropsychological tests of mini-mental state examination (MMSE) and Addenbrooke’s cognitive examination (ACE-III). Spontaneous speech samples were recorded through the description task of Boston Diagnostic Aphasia Examination Cookie-Theft Picture. The voice recordings were processed and analyzed by the computerized software (developed by the authors) to measure 6 speech variables. Independent sample t- test was used to evaluate the differences of the speech variables and neuropsychological test scores between subject groups. Pearson correlation was used to identify the correlation between the speech variables and the neuropsychological results. The differences of ratio of total pause duration to total speech duration were significant in the AD vs control group (P<0.001) and aMCI vs control group(P<0.001), but not significant between the aMCI and AD group(P=0.658). The ratio of total pause duration to total speech duration was strongly associated with the Fluency score in ACE-III, and also significantly associated with the MMSE score, total ACE-III score, and scores of all the other items in ACE-III. The Normalized long pause count was larger in the aMCI group (P=0.007), while smaller in the AD group (P=0.013) when compared with the control group. The Normalized long pause was significantly associated with language score in ACE-III. The ratio of total pause duration was thought to be a valuable indicator of cognitive impairment severity. And the results of the Normalized long pause count implied that aMCI patients had more difficulty in words finding and had more frequent pauses, while AD patients showed more silent time at each pause and had fewer pause counts. Normalized long pause showed to be a very important indicator to help distinguish among healthy controls, aMCI, and AD. Computer assisted speech analysis showed to be a noninvasive and a promising assessment tool in the evaluation and differentiation diagnosis of cognitive impairment.

Efficient Pause Extraction and Encode Strategy for Alzheimer's Disease Detection Using Only Acoustic Features from Spontaneous Speech

Efficient Pause Extraction and Encode Strategy for Alzheimer’s Disease Detection Using Only Acoustic Features from Spontaneous Speech

Identification of Alzheimer's Disease Patients Based on Oral Speech Features

Leveraging Prompt Learning and Pause Encoding for Alzheimer's Disease Detection

Exploring Multimodal Approaches for Alzheimer's Disease Detection Using Patient Speech Transcript and Audio Data

Artificial Intelligence-Enabled End-To-End Detection and Assessment of Alzheimer's Disease Using Voice

Leveraging Pretrained Representations with Task-related Keywords for Alzheimer's Disease Detection

Temporal Integration of Text Transcripts and Acoustic Features for Alzheimer's Diagnosis Based on Spontaneous Speech

Leveraging Pretrained Representations with Task-Related Keywords for Alzheimer’s Disease Detection

Detecting Alzheimer's Disease Based on Acoustic Features Extracted from Pre-trained Models

End-to-End ASR-Enhanced Neural Network for Alzheimer’s Disease Diagnosis

Alzheimer's Dementia Recognition Using Acoustic, Lexical, Disfluency and Speech Pause Features Robust to Noisy Inputs

Alzheimer Disease Classification through ASR-based Transcriptions: Exploring the Impact of Punctuation and Pauses

Detection of Mild Cognitive Impairment From Non-Semantic, Acoustic Voice Features: The Framingham Heart Study

Combining Prosodic, Voice Quality and Lexical Features to Automatically Detect Alzheimer's Disease

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

Noninvasive automatic detection of Alzheimer's disease from spontaneous speech: a review

Detecting Alzheimer's Disease from Continuous Speech Using Language Models.

Using the Outputs of Different Automatic Speech Recognition Paradigms for Acoustic- and BERT-based Alzheimer's Dementia Detection through Spontaneous Speech

The Optimization of a Natural Language Processing Approach for the Automatic Detection of Alzheimer’s Disease Using GPT Embeddings

P1‐349: Computer Assisted Speech Analysis in Amnestic Mild Cognitive Impairment and Alzheimer’s Disease