Speech data collected from older adults could help determine the risk of driving accidents, according to a new study out of Japan.
Published in the Journal of Medical Internet Research, the study found that drivers age 60 and up who experienced accidents or near accidents had changes in their speech features indicating cognitive impairments. Researchers also found that the algorithms used to predict accident risk using speech data could make those predictions with up to 88.3% accuracy.
“Speech in daily life can be used [as] potential data sources for determining cognitive impairments related to driving safety. Speech involves multiple interacting cognitive abilities including attention, memory, and executive functions,” researchers wrote.
The study was conducted in Ibaraki, Japan, with approval from the University of Tsukuba Hospital Ethics Committee and with grants from the Japan Society for the Promotion of Science KAKENHI.
Of the 60 individuals who participated in the study, 23 reported having near-accidents, two reported accidents and one reported both.
Researchers did not find significant differences in cognitive assessment variables between the individuals who reported accidents or near-accidents and those who did not.
However, the study found significant differences between the groups' speech features. Specifically, researchers found 10 speech features with differences.
The participants who reported accidents or near-accidents had decreased speech rate, and jitter, as well as increased response time and long pauses.
“These speech features were reported in previous studies as significant indicators of changes in cognitive function, and the trends in their changes were consistent with those observed in individuals with cognitive impairments, and patients with Alzheimer disease and mild cognitive impairment (for speech rate, for jitter, for response time, for proportion of long pauses),” researchers wrote.
HOW IT WAS DONE
Data for the study was collected over a year-and-a-half. To start, 71 healthy adults age 60 and up were recruited to participate in cognitive assessments and speech data collection.
Participants took cognitive assessments typically used to diagnose dementia. Two participants did not meet the criteria and were removed from the study.
Researchers collected speech data by simulating conversations with a voice assistant on modern smart speakers and smartphones. Participants were tasked with asking for tomorrow’s weather, booking a movie ticket online and creating a calendar event. The follow-up questions for each task became more complex as participants were asked for more detailed information.
Though participants believed they were talking with a computer system, an experimenter mediated the interaction by prompting the system to move onto the next question as participants responded.
About a year and a half after collecting the speech data, researchers distributed a questionnaire to participants asking about their driving experiences within the past year. Of the 71 original participants, 60 answered the questionnaire, which contained free-form questions about accidents and near accidents.
Researchers defined near accidents as infractions and any other driving incidents that participants felt were dangerous.
As the world’s 60-plus population continues to grow, researches see an increasing need to address driving safety in older adults.
A 2018 study published in Taylor and Francis Online found that adults ages 65 and older share common concerns regarding their own driving and the driving of others on the road.
While much has been done to study the impact of age and cognitive ability on driving safety, the connection between speech data and driving risk appears to be a fairly new area of study.
“Although further studies with speech data collected in everyday life and objective data for near-accidents are needed, our study provides the first empirical results suggesting that speech data during interactions with voice assistants in smart speakers and smartphones could help predict future accident risks of older drivers by capturing subtle impairments in cognitive function,” the researchers wrote.
Future studies need to utilize more objective measures, the researchers wrote. They noted that their study had several limitations, including the sample size, its lab setting, and reliance on self-reported accidents and near accidents.
Though further study is needed, the researchers maintained that their “results can be used in future efforts toward preventing driving accidents of older adults through continuous passive unobtrusive monitoring.”