Blog | Tartaglia, a network to exchange relevant information to research

Being able to identify those people that in the future will develop a disease is one of the most important scientific endeavours nowadays in health research.

At Ace Alzheimer Center Barcelona we are leaders in Alzheimer's research, where one of the main objectives is to identify speech parameters that allow us to identify people who are exposed to a higher risk of this disease.

This project is framed within a larger one called Tartaglia (Niccolò Tartaglia, an Italian mathematician and engineer of the 16th century), in which 16 healthcare and technological entities work, and which aims to harness the benefits of artificial intelligence to solve some of the challenges of medicine and speed up clinical research. Tartaglia’s aim is to design a collaborative network between different entities, in the form of a federated network, where clinical data is shared securely and where different professional profiles converge (healthcare professionals, biologists, computer scientists, data scientists, etc.).

But why are we interested in speech?

In 1996, a study was published in the prestigious medical journal JAMA involving a group of cloistered nuns in the United States. Various parameters were collected from autobiographical data extracted from their correspondence by letter. The first data collected date back to when the participants were in their early twenties.

Almost sixty years later, the cognitive capacity of those participants who were alive was assessed. In the case of the deceased, their brains could be studied in an attempt to identify indicators of Alzheimer's disease.

The results indicated that poorer language skills during youth were associated with poorer cognitive skills in old age, as well as a higher probability of developing Alzheimer’s.

More than 25 years have passed since the publication of this study and the technological capabilities we now have at our disposal to analyze language have changed substantially.

We know that memory impairments, which are the most striking in Alzheimer's disease, are also associated with changes in language production. The presence of signs of cognitive impairment that have a very early appearance, known as early onset signs, often require comprehensive neurological and neuropsychological examinations.

These examinations and the interpretation of the results require a high level of expertise and often several hours of exploration. This investment of resources, unfortunately, does not facilitate repeated assessments, let alone targeting individuals who do not live with the disease but who may already be showing some signs of it.

The traditional way of assessing language function has usually been based on the use of structured linguistic tests, inserted in more or less extensive evaluation batteries where other cognitive functions are often evaluated and which are popularly known as "pen and paper" tests. This would be the most frequent procedure expected when a thorough evaluation of every person is to be made.

However, the emergence of speech processing technologies based on artificial intelligence, coupled with the processing capacity of current computers, allow to complement and diversify the strategies available for the identification of people at risk. Many of them, as mentioned above, are people who do not present overtly relevant or worrisome symptoms, but to whom we should already be paying some attention.

The analysis of spontaneous speech is based on the analysis of conversations or on the spoken description of ordinary activities, often related to daily-life situations. One of these activities, which is often used, is the description of a picture where an action is taking place. These types of activities have the advantage that they are easy to use in the tests and representative of the linguistic abilities of most of our patients, since they are based on, as we said, activities carried out generally by most people.

In addition, and very importantly, they allow the use of digital devices, such as a tablet or a cell phone, and do not require more than a few minutes to be completed. Despite the simplicity of the evaluation, it is possible to generate a large number of variables and indicators that can be quantified and that may be of scientific interest once they have been properly processed and analyzed. Most importantly, nowadays all of this can be done automatically and almost immediately thanks to the use of electronic devices.

How do we do it at Ace Alzheimer Center Barcelona?

We ask our patients to collaborate by spending a few extra minutes to complete this new language protocol. They do so by taking advantage of the same visit they have already scheduled with us. It only takes 6-7 minutes more. What we ask them is to do some activities like describe an action presented in a picture or talk freely about a pleasant personal event. And we do so with the help of a tablet.

Soon we will be able to do these examinations of speech from home, without the patient having to come visit us in person. In fact, we already offer some of our patients and family members regular follow-ups and examinations from home, using mobile devices or computers.

To ensure the quality of our patients' speech processing and the use of the best artificial intelligence algorithms to identify the most relevant parameters, we work with the digital platform called Accexible, a technology company specialized in automatic speech processing, and which is also part of the Tartaglia consortium.

What else do we need in this project?

To continue to enrich the study sample. Let it grow.

We need more people with dementia, but also people who do not yet have the disease since they are extremely valuable in research when it comes to identifying variables or early onset symptoms.

The more people we evaluate and the more enriched they are in terms of disease manifestation (stage, severity, type of dementia...) the more options we will have to identify speech factors (sometimes very subtle!) that may have a great relevance in the detection of people with a higher risk of developing Alzheimer's disease.

We plan to evaluate more than 1500 people, some of them more than once, so that we can make a proper follow-up.

The Tartaglia project has enabled us to acquire some of the financial and professional resources needed to carry out a technological project of this magnitude.


Medical Psychology Head at Ace Alzheimer Center Barcelona