By Franklin Delgado
Researchers from the University of Alicante (UA) and the Institute for Health and Biomedical Research (Spain) have achieved an accuracy close to 90% in the early detection of Alzheimer’s by analyzing voice recordings of just four minutes.
This tool represents a great advance in the preliminary evaluation of the disease.
The application listens What a patient says and how he says it to detect signs and have a preliminary evaluation, collected The Newspaper.
Speech analysis
The platform, called IAEAV, uses advanced technologies to analyze both the content and the way in which the patient expresses himself. Elements such as tone, pauses and linguistic structure are examined, facilitating a more effective identification of cognitive changes.
“We use the voice as if it were purely audio. We look to see if there are more pauses, if the tone is flatter or monotonous. And we also convert voice into text and look for recurring linguistic structures in Alzheimer’s patients,” explained Miguel Ángel Teruel Martínez, researcher in Computer Languages and Systems at the UA, professor and head of the technological development of the project.
Data collection is carried out through a mobile application, which allows users to record their voices in various contexts. This methodology facilitates access for older people in clinical and domestic settings, eliminating barriers to entry.
Implications and future potential
The team created a valuable data substandard composed of recordings from 223 volunteers. These collections not only support early diagnosis, but also provide field cloth for further research into the relationship between language and neurodegenerative changes.
The researchers stated that the technology not only seeks to optimize diagnosis, but also encourage the generation of data that can drive future discoveries in the treatment and management of Alzheimer’s.
Other diagnostic methods
New diagnostic methods for Alzheimer’s, such as biomarker-based blood tests (e.g. p-tau217) or artificial intelligence (AI) tools that analyze imaging tests or clinical data, are often compared in terms of accuracy, accessibility, cost and level of invasiveness with traditional methods such as clinical evaluation, lumbar puncture and positron emission tomography (amyloid PET).
Diagnostic accuracy
- Some blood tests for p-tau217 achieve greater than 90% accuracy in detecting Alzheimer’s in people with cognitive symptoms, the same as that obtained with lumbar puncture and higher than clinical evaluation without biomarkers.
- Amyloid PET and its visual evaluation also show sensitivity and specificity close to 97%, making it a very solid reference, although expensive and invasive.
Invasiveness and accessibility
- Blood biomarkers and AI tools are much less invasive than lumbar puncture and can be applied on a large scale, facilitating early screening and use in primary care.
- In contrast, PET and photon emission tomography (SPECT) are more expensive and less accessible, and lumbar puncture requires more complex procedures and may deter many patients.
Cost and efficiency
- It has been estimated that the use of blood biomarkers can reduce the costs associated with the diagnosis of Alzheimer’s by between 60% and 81% compared to current methods.
- AI tools that analyze brain images or medical records can speed diagnosis and increase accuracy, while integrating into common workflows without adding new invasive tests.
You may also be interested in:
· Early menopause carries a high risk of suffering heart attacks: what signs should alert you?
· AI can predict biological aging through the retina
· The microbiome: key to understanding the increase in colorectal cancer in young people
