ISABIAL and University of Alicante researchers develop a platform for the early detection of Alzheimer’s disease through voice analysis

  • Using advanced natural language processing (NLP) and deep‑learning technologies
  • Voice data will be collected through a simple and accessible mobile application that allows users to record their voices in various contexts

Alicante (28.04.2026).- Researchers from the Alicante Institute for Health and Biomedical Research (ISABIAL) and the University of Alicante (UA) have developed a platform for the early detection of Alzheimer’s disease through voice analysis, using advanced natural language processing (NLP) and deep‑learning technologies. The system collects voice data through a simple and accessible mobile application that enables users to record their speech in various contexts.

The project, funded by the Department of Innovation, Industry, Trade and Tourism of the Generalitat Valenciana through Next Generation funds and managed by the GVA (Project INREIA/2024/176), addresses the detection of Alzheimer’s disease. Since current treatments are more effective when administered in the early stages of the condition, early diagnosis becomes a key tool for slowing its progression and improving the quality of life of patients and their caregivers.

The research team, led by Miguel Ángel Teruel (Principal Investigator) from the University of Alicante (UA), and Ángel Pérez Sempere (Co-Principal Investigator) from ISABIAL, has included researchers from UA Álvaro Navarro, Javier Sanchis, Cristian Vera, Javier García, Bárbara Escalante, Luis Moreno Navarro and Lyan Montero Pardo, from ISABIAL. The Artificial Intelligence Platform for Early Detection of Alzheimer’s Disease through Voice (IAEAV) aims to identify patterns of cognitive decline based on the analysis of acoustic and linguistic signals of the human voice.

The researchers note that this innovative tool is based on previous studies showing that early neurological changes can manifest in language alterations, such as reduced syntactic complexity, prolonged pauses and grammatical errors.

To this end, they add, advanced natural language processing (NLP) and deep‑learning technologies are used. This platform seeks not only to facilitate early diagnosis, but also to offer a simple, low‑cost, easily accessible and non‑invasive solution.

In addition, voice data will be collected through a simple and accessible mobile application that allows users to record their voices in various contexts, such as reading texts, spontaneous narration or responses to standardised questions. The application is designed for use both in clinical settings and at home, which, as the team highlights, “reduces access barriers and facilitates data collection in populations with limited resources”.

The recordings collected through the mobile application are then processed to extract acoustic features such as pitch, intensity and pauses, as well as linguistic aspects such as semantic richness and errors in verbal fluency. These features are subsequently evaluated by deep‑learning models trained on representative datasets, enabling accurate and personalised detection. One of the most relevant aspects of this approach is its accessibility.
In this regard, Miguel Ángel Teruel explains that this technology “not only aims to improve clinical detection, but also to contribute to scientific progress by generating large volumes of voice data, which can enable deeper research into the relationship between linguistic alterations and neurodegenerative changes, fostering advances in the treatment and management of the disease”.
Back To Top
No results found...