It takes a lot of time and money to diagnose Alzheimer's disease. After lengthy face-to-face neuropsychological examinations, doctors have to decipher, review and analyze each response in detail.

But researchers at Boston University have developed a new tool that could automate the process and eventually bring it online. Their computational model, based on machine learning, can detect cognitive impairment from audio recordings of neuropsychological tests without the need for an in-person appointment. The results of the study were published in the journal Alzheimer's & Dementia.

He says faster and earlier detection of Alzheimer's disease could lead to larger clinical trials targeting people in the early stages of the disease and potentially enable clinical interventions to slow cognitive decline.

The research team trained their model using audio recordings of neuropsychological interviews with more than 1,000 people from the Framingham Heart Study, a Boston University-led long-term project that studies cardiovascular disease and other physiological conditions.

Using automated online speech recognition tools and a machine learning technique called natural language processing, which helps computers understand text, they had their program transcribe the interviews and then encode them into numbers. The final model was trained to assess the likelihood and severity of an individual's cognitive impairment using demographic data, text encodings, and real world diagnoses made by neurologists and neuropsychologists.

The model was able not only to accurately distinguish healthy people from people with dementia, but also to detect differences between people with mild cognitive impairment and dementia.

And as it turned out, the quality of the recordings and the way people spoke—whether their speech flowed quickly or constantly stuttered—was less important than the content of what they said.