JMIR Publications

ABSTRACT

Background:

The ability to objectively measure the severity of depression and anxiety disorders in a passive manner could have profound impacts on the way in which these disorders are diagnosed, assessed, and treated. Existing studies have demonstrated links between both depression and anxiety and the linguistic properties of words which people use to communicate. Smartphones offer the ability to passively and continuously detect spoken words in a person’s environment, to continuously monitor and analyze linguistic properties of speech produced by the subject and other sources of ambient speech in their environment. The linguistic properties of automatically-detected and recognized speech may be used to build objective severity measures of depression and anxiety.

Objective:

The aim of this study was to determine if the linguistic properties of words passively detected from environmental audio recorded using a subject’s smartphone can be used to find correlates of symptom severity of social anxiety disorder, generalized anxiety disorder, depression, and general impairment.

Methods:

An Android app was designed, together with a centralized server system, to collect periodic audio recordings of subjects’ environments and to detect English words using automatic speech recognition. Subjects were recruited into a 2-week observational study where the app was run on their personal smartphone to record and analyze audio. Subjects also completed self-report severity measures of social anxiety disorder, generalized anxiety disorder, depression, and functional impairment. Detected words were categorized across 67 categories and correlations were measured between words counts in each word category and the four self-report measures to determine if any categories could serve as correlates of social anxiety disorder, generalized anxiety disorder, depression, or general impairment.

Results:

Participants were 112 Canadian adults from a non-clinical population, with 86 participants yielding sufficient data for analysis. Correlations between word counts across 67 word categories and each of the 4 self-report measure revealed a strong relationship between usage rates of death-related words and depressive symptoms (r = 0.41, P = .0001). Also presented are interesting correlations between rates of word usage in the categories of reward-related words (with depression, generalized anxiety) and vision-related words (with social anxiety).

Conclusions:

In this study group, English words automatically recognized from environmental audio were shown to contain a number of potential associations with severity of depression and anxiety. This work suggests that sparsely sampled audio could provide relevant insight into subjects’ mental health.