We have constructed a monitoring system to measure the degree of stress and depression from the voice in a telephone conversation so far. The number of users of our research application in Japan has already exceeded 7,000, and some companies have adopted it in the field of industrial health. In addition, it is pre-installed on some domestic smartphones, and the number reaches 500,000.
Recently, vocal biomarkers have attracted attention in diseases other than stress and depression. Voice is non-invasive, easy and inexpensive, can be measured remotely, and thus is an ideal biomarker for screening.
Attempts to use voice as a biomarker are not so new. In particular, it has been practiced for a long time in the psychiatric and neurologic diseases area where subjective judgment is mainly performed, and in the otolaryngology area in which voice is directly involved. The reason for the rapid development in recent years is largely due to the development of computers. The capabilities of current smartphones are comparable to supercomputers 30 years ago. In addition, the fact that voice is becoming the mainstream of human-machine interface also supports the development.
However, in order to use voice as a biomarker, not only comparison with healthy persons, but also discrimination for each disease is required. In this talk, I will introduce recent findings on this point.