Detect if a person suffers mild cognitive impairment or dementia in early stages through the speech analysis It is already possible thanks to the acceXible model. Identifies vocal biomarkersspecific characteristics of the voice that indicate various health conditionsfor the diagnosis and monitoring of diseases.
Throughout 2019, a clinical trial with 150 participants of the Cognitive impairment center of Madrid and the Oroitu Center, located in Getxo (Bizkaia). In this clinical trial, data have been collected to create a model based on «machine learning” that learn how to detect mild cognitive impairment and dementia just like a doctor would. They are currently perfecting the acceXible model algorithm, which is why they are developing a series of pilots in Bilbao, Madrid and Barcelona.
Presentation of the Accessible model at the INNOLAB facilities in Bilbao
At the end of last February, this model was presented at the Bilbao innovation platform INNOLAB. During the presentation, the doctor in telecommunications engineering, Carmen Garcia Mateoexplained how different diseases can be detected through voice analysis, from the cognitive impairment to depression or Parkinson’s.
For his part, neurologist Manuel Fernández detailed how tests are traditionally performed for the diagnosis of diseases related to deterioration and the obstacles neurologists encounter nowadays.
And he explained how, based on these tests, the model based on oral description: from an image, list of words from the cerad, 2 tests of semantic verbal fluency and one phonetics. These tests are recorded on any device with internet access (ipad), and are sent to the acceXible model hosted in the cloudwhich returns, in a couple of minutes, the result whether the patient suffers from mild cognitive impairment or is otherwise healthy.
Susana Perez de las Heras, director of the Oroitu Center, shared her experience of the deployment of Accessible in her Center. She stressed that the tool saves time of 25 minutes, both for screening in detection and for follow-up in the evolution of the disease. He pointed the Easy to use of the tool and even the possibility of self-administration. In addition, he stressed the importance of early detection that allows the model, which reaches the 91% of people suffering from impairment or dementia in early stages.
Reliability greater than 80%
Almost half of the people who suffer That is why it is so important early detection. A detection of the deterioration allows greater possibilities of early treatment of the disease.
it’s a non-invasive alert systemeasily scalable and allow early detection. It is a solution that allows diagnosing cognitive impairment with a reliability greater than 80%. acceXible has participated in other events, such as the #BarcelonaTechSpiritconference organized on ai4health, and at the beginning of March he presented his model at the Pier01 from Barcelona Tech City.