Using machine learning for early detection of dementia

A new tool trained using machine learning can predict dementia much earlier than normal.

According to Tekna technology and technology news service, scientists have developed a machine learning model that uses speech patterns for diagnosis with the aim of early diagnosis of dementia. This device can evaluate the conditions in the early stages.

Zahra Shah, one of the researchers of this project and the researcher of the Alberta Research Institute, says: Our goal in this research is to pay attention to speech as a window to the mind. Speech as a biomarker is to be considered here. In this way, it is possible to identify patterns that are effective in diagnosing and monitoring diseases such as dementia. Usually, speech defects can be recognized through three characteristics: pauses in speech, word complexity, and comprehensibility.

Shah said in this regard: If a word has a longer pronunciation, it will be more complicated. The hypothesis of this research says that patients with dementia use less complexity in speech. The model developed by the researchers has between 70 and 75% accuracy. Shah added: This model is like a tool for clinical diagnosis, but we don’t use it like a tool.

But it can be a starting point for screening people at risk. Of course, this tool cannot yet be introduced as a clinical diagnosis. However, Azami can help people to assess the risk of contracting the disease with the help of a smartphone.