A hidden signal in the way you speak could predict Alzheimer’s

Diagnosing someone with Alzheimer’s is a long and difficult process, and only generally begins once they are showing signs of mild cognitive impairment (MCI).

Tests include interviews, brain imaging, and brain and spinal fluid analysis.

All of this takes time, delaying potential intervention to reduce the symptoms, which is more effective the earlier it used.

Now however, researchers from Boston University have turned to artificial intelligence to help spot whether someone with MCI will go on to develop Alzheimer’s – just from how they speak.

The team said their model can predict with 78.5% accuracy whether someone with a mild cognitive impairment is likely to remain stable over the next six years or develop the dementia associated with Alzheimer’s disease. 

Co-author Dr Ioannis Paschalidis said: ‘We wanted to predict what would happen in the next six years – and we found we can reasonably make that prediction with relatively good confidence and accuracy. 

‘It shows the power of AI.’

The team published their findings in the journal Alzheimer’s & Dementia.

The new algorithm was trained in transcribed audio recordings of 166 individuals with MCI, aged 63 to 97.

As the team already knew who had developed Alzheimer’s, a machine learning approach was used to find common signs in their transcribed speech that linked the 90 people whose cognitive function would decline into Alzheimer’s.

Then, the algorithm could be applied in reverse, but with factors such as age and self-reported sex added to produce the final predictive score.

The model pulled out content from the interviews, such as the words spoken and how they are structured – rather than enunciation or speed. The researchers then used a combination of speech recognition tools to spot connections between speech, demographics, diagnosis and disease progression.

Although there is no cure for Alzheimer’s, there are drugs to help reduce or control some cognitive and behavioural symptoms, and early detection will help provide more time with loved ones and provide more opportunities to study the disease and its progression. 

‘We hope, as everyone does, that there will be more and more Alzheimer’s treatments made available,’ Dr Paschalidis said. 

‘If you can predict what will happen, you have more of an opportunity and time window to intervene with drugs, and at least try to maintain the stability of the condition and prevent the transition to more severe forms of dementia.’