Deep learning-based speech analysis for Alzheimer’s disease detection: a literature review

Abstract

Background Alzheimer’s disease has become one of the most common neurodegenerative diseases worldwide, which seriously affects the health of the elderly. Early detection and intervention are the most effective prevention methods currently. Compared with traditional detection methods such as traditional scale tests, electroencephalograms, and magnetic resonance imaging, speech analysis is more convenient for automatic large-scale Alzheimer’s disease detection and has attracted extensive attention from researchers. In particular, deep learning-based speech analysis and language processing techniques for Alzheimer’s disease detection have been studied and achieved impressive results. Methods To integrate the latest research progresses, hundreds of relevant papers from ACM, DBLP, IEEE, PubMed, Scopus, Web of Science electronic databases, and other sources were retrieved. We used these keywords for paper search: (Alzheimer OR dementia OR cognitive impairment) AND (speech OR voice OR audio) AND (deep learning OR neural network). Conclusions Fifty-two papers were finally retained after screening. We reviewed and presented the speech databases, deep learning methods, and model performances of these studies. In the end, we pointed out the mainstreams and limitations in the current studies and provided a direction for future research.

Artificial intelligence for dementia research methods optimization

“…Furthermore, CNN-related models have recently found increasing application in NLP tasks for AD detection, including sentence classification, search query retrieval, and semantic parsing. 82 RNNs are typically used on data that is sequential or time dependent in nature because their hidden state component (i.e., "memory cell") allows previous inputs to influence a given output. For example, Alam et al 79 applied a long short-term memory (LSTM)-based RNN to predict onset of physical agitation episodes in patients with dementia, using motion sequences obtained from smartwatches.…”

“…In speech-based AD detection, DL model architectures including RNN, LSTM, gated recurrent unit, and bidirectional LSTM, have been used to extract timing information from audio data. 82 DBNs have been used to explore dementia-related factors using genetic data 77 while GCNs have been applied primarily to neuroimaging data to capture brain network information for dementia classification. 78…”

“…These results demonstrate that DL models incorporating multiple data modalities may become vital to fully use the wealth of information available for dementia research. Furthermore, CNN‐related models have recently found increasing application in NLP tasks for AD detection, including sentence classification, search query retrieval, and semantic parsing 82 …”

“…The LSTM–RNN model achieved a higher F1‐score (0.85) than traditional ML methods, such as KNNclass (0.69), SVM (0.67), and AdaBoost (0.71), highlighting the potential of using such models in sensing‐based behavior inference. In speech‐based AD detection, DL model architectures including RNN, LSTM, gated recurrent unit, and bidirectional LSTM, have been used to extract timing information from audio data 82 …”

GPT-4 and Neurologists in Screening for Mild Cognitive Impairment in the Elderly: A Comparative Analysis Study

“…AI’s role in healthcare is expanding, particularly in diagnostics and decision-making [14]. Current early screening approaches for cognitive impairments include using digital tools like tablet computers [11,12,15], virtual reality [16–18], and machine interactions with robots. These methods contrast with the more complex procedures typically conducted by physicians [19].…”