The relationship between language patterns and cognitive impairment have been the subject of many scientific studies. In a study coordinated by researchers at the Massachusetts Institute of Technology as well as the Boston University School of Medicine and Public Health, an automated system that evaluates features of speech and language was used to examine 92 subjects’ speech samples. From the experiment, the scientists constructed a binomial logistic regression model of the speech and language features to classify the presence of cognitive impairment. This study evaluated language features such as patterns of word usage, speech waveform, intervals of time between words, and voice quality measures like pitch. They found that speech characteristics such as decreasing pitch, presence of jitter, shorter responses and segments of speech, and responses phrased like questions are positively associated with cognitive impairment. Many other studies offer similar conclusions and expand further by demonstrating the role of language characteristics as a biomarker for future Alzheimer’s disease.
While the first study discussed in this article establishes a relationship between linguistic variables and cognitive impairment, this study uses this relationship to develop a simple language test for Alzheimer’s disease. In a study conducted by researchers at IBM Thomas J. Watson Research Center and Pfizer Worldwide Research and Development, written language samples were analyzed to predict Alzheimer’s disease more than seven years before diagnosis. 270 participants were divided into two sets: 80 participants in a test set and 190 participants in a training set. Half of the participants in the test set had developed symptoms of Alzheimer’s by the age of 85 and the other half were cognitively normal. All of the participants performed a written picture analysis task. The writing samples from the training set were collected by the researchers and analyzed via automated linguistic analysis to identify 87 linguistic markers. The researchers developed models to predict the future development of Alzheimer’s disease by assessing language performance. From the writing samples, they found that writing patterns such as writing short and simple phrases, repeating and misspelling words, and skipping punctuation were associated with future onset of Alzheimer’s. The study led to the development of a simple language test that can predict the development of Alzheimer’s disease seven years before diagnosis with about 70% accuracy. According to the researchers, this rate of accuracy is much higher than older predictive methods like neuropsychological testing. The exploration of relationships between linguistic variables and verbal language patterns in cognitive impairment may further advancements in non-invasive and inexpensive tests to assess risk of developing Alzheimer’s. Despite the absence of an effective cure or preventative treatment for the disease, early detection of risk factors may help clinicians delay its progression.
Eyigoz E, Mathur S, Santamaria M, Cecchi G, Naylor M. Linguistic markers predict onset of
Alzheimer's disease. EClinicalMedicine. 2020 Oct 22.
Tuka Alhanai, Rhoda Au, James Glass, “Spoken Language Biomarkers for Detecting Cognitive
Impairment.” Massachusetts Institute of Technology. 2017.