AI studies find that patients with Parkinson’s disease speak differently

Using artificial intelligence (AI) to process natural language, a research group evaluated speech characteristics among patients with Parkinson’s disease (PD). AI analysis of their data determined that these patients spoke using more verbs and fewer nouns and fillers. This study was led by Professor Masahisa Katsuno and Dr. Katsunori Yokoi, Nagoya University School of Medicine, in collaboration with Aichi Prefectural University and Toyohashi University of Technology. They publish their results in a journal Parkinsonism & Related Disorders.

Natural language processing technology (NLP) is a branch of AI focused on enabling computers to understand and interpret large amounts of human language data using statistical models to identify patterns. Given that patients with PD experience a variety of speech-related problems, including impaired speech production and language use, the group used NLP to analyze differences in patients’ speech patterns based on 37 characteristics using text generated from free speech.

Analysis revealed that patients with PD used fewer common nouns, proper nouns, and fillers per sentence. On the other hand, they speak using a higher percentage of verbs and variations for the case particle (an important feature of the Japanese language) per sentence.

According to Yokoi, “When I ask them to talk about their day in the morning, a PD patient might say something like the following, for example: ‘I wake up at 4:50 am. I thought it was a little early, but I woke up. It took about half an hour to go to the toilet, so I showered and got dressed around 5.30am. My husband cooks breakfast. I have breakfast after 6 am. Then I brush my teeth and get ready to go out.’”

Yokoi continues: “Whereas someone in a healthy control group might say something like this: ‘Well, in the morning, I get up at six, and get dressed, and, yes, wash my face. Then, I feed my cat and dog. My daughter prepared food, but I told her I couldn’t eat, and I, umm, drank some water.’”

“While these are examples of conversations we created that reflect characteristics of people with PD and healthy people, what you should see is that they are the same length,” explains Yokoi. “However, PD patients spoke shorter sentences than people in the control group, leading to more verbs in the machine learning analysis. Healthy controls also use more fillers, such as ‘good’ or, ‘umm’, to connect sentences.

The most promising aspect of this study is that the team performed the experiment on patients who had not yet exhibited the characteristics of cognitive decline seen in PD. Therefore, their findings offer a potential means of early detection to differentiate PD patients.

“Our results show that even in the absence of cognitive decline, the speech of patients with PD is different from that of healthy subjects”, Professor Katsuno, head of the study, concluded. “When we attempted to identify PD patients or healthy controls based on these conversational changes, we were able to identify PD patients with greater than 80% precision. These results demonstrate the possibility of language analysis using natural language processing to diagnose PD.”