It’s something that more than 6 million Americans face every day: Alzheimer’s disease. And their numbers are expected to rise to around 13 million by the year 2050. Today, more than 1 in 10 Americans over the age of 65 have the disease, but tracking the progression of Alzheimer’s can be a challenge as it is marked by subtle changes in a person’s cognitive ability. For decades, physicians have relied on a simple, subjective cognitive test to track Alzheimer’s. But recent findings from the Toronto-based Winterlight Labs suggest that AI-based analysis of speech patterns may be a better way to gauge the cognitive decline associated with Alzheimer’s.
So how does it work? Winterlight’s clinical research shows that applying artificial intelligence to natural language processing is an effective way to detect changes in psychiatric and neurological symptoms. Their platform extracts and measures more than 550 features of language and speech, including a person’s word choices, sentence structures, and a range of other linguistic components of speech. Then algorithms group these variables into composite metrics, such as lexical richness, syntactic complexity, and word-finding difficulty. Using these data sets, the speech-based digital biomarker platform can track changes over time and assess responses to treatment.
To assess the real-world effectiveness of its platform, Winterlight engaged 101 patients with early stage Alzheimer’s disease who were part of an ongoing trial with the biotech company Genentech. The trial used the traditional Clinical Dementia Rating test, in which the patient is interviewed with a standard questionnaire designed to assess cognitive abilities, including orientation, judgment, and problem solving. These interviews were recorded, allowing Winterlight’s platform to analyze over 500 linguistic and acoustic markers, and determine that a select number of these markers indicated neurological decline.
While the technology is primarily used to assess Alzheimer’s, Winterlight says their platform can also detect the disease with a high degree of accuracy by analyzing vocal biomarkers and linguistic cues. Winterlight researcher Jessica Robin says, “Given the crucial role that language changes play in Alzheimer’s, automated language processing represents a new tool to characterize speech and language patterns and provide additional insight into a patient’s condition.” And a tool that could bring Alzheimer’s care to people living in remote areas and in underserved communities. Robin adds, “We’ve worked to streamline the process of implementing automated tools into research, allowing for low-burden assessments suitable for remote testing.”