Mental illness is a disease that packs two punches: first symptoms, then stigma. The stigma surrounding mental health — especially men’s mental health — only contributes to the struggles of those already suffering, and at its extreme culminates in 700,000 individuals lost to suicide each year.
Within mental health diagnoses, depression and anxiety are among the most common. Depression is a worldwide epidemic regularly overlooked due to the ‘invisible’ nature of its diagnosis and compacted by societal pressure to sweep it under the rug. The World Health Organization estimates 3.8% of the total population struggles with the chronic illness, totaling 280 million individuals worldwide. For reference, in the first two weeks of 2022, around 3.37% of New Jersey residents were considered active cases of COVID-19, which was profound enough to delay Stevens students’ return to campus for the Spring 2022 semester.
Raising awareness towards the significance of the number of individuals struggling, development of earlier detection methods, and advancements in the ease of access to treatment all progress efforts to curb the ongoing mental health crisis.
In December 2022, BMC Psychiatry published a groundbreaking study in early detection methodology. Alexandra König and colleagues were able to predict if an individual had clinically significant levels of depression — with 93% accuracy — from a two minute speech analysis.
Until this publication, prior art for speech analysis and depression primarily focused on the comparison of patients already diagnosed with a major depressive disorder versus that of their healthy counterparts. The work put forth by the French National Institute for Research in Computer Science and Automation’s König is unique in that it brings knowledge from prior models and applies it to a non-clinical setting. König and colleagues successfully developed an analytical program capable of detecting early and sensitive markers of clinical depression in individuals with no previous clinical diagnoses.
It’s been widely understood that patients with depression may present with more monotonous speech, less expressive rhythm and intonation, and utilize more words between pauses. Clinicians and practicing psychologists have been analyzing the speech of their clients for years. It is unsurprising, however; that this subjective form of evaluation, whether conscious or unconscious, is often unreliable and hard for a clinician to quantify. It takes years of professional experience to train the ear to these subtle differences. König’s team’s 93% successful identification rate defined the length of a pause to be a mere 10 milliseconds.
Computational analysis of speech in depression-prone patients presents a significantly more objective pathway to diagnosis and treatment. The ones and zeros dialect of computers translates beautifully into temporal measurements and is very conducive to reading between the lines with less of an emphasis on the very human desire to understand and empathize with semantics. The study’s evaluation was rooted primarily in the number of words participants utilized between pauses in speech. Additional acoustic and segmental features such as pitch, speed of speech, and length of speaking segments also informed the prediction.
While a computer may excel at quick detection, its ability to develop individualized treatment plans crafted to the patient’s unique struggles falls a bit flat, and is best left to a practicing psychologist. A program such as König’s is perhaps most effectively used as a tool to screen individuals, and minimize the time clinicians spend screening. Non-clinical detection methods shorten the agonizing wait for counseling by maximizing the time clinicians can spend practicing. Quick, emotionally non-invasive, early detection methods such as this study’s analytical model provide the promise of a future where individuals struggling with depression spend less time facing their health struggles alone.
Until early screening methods such as König’s become more mainstream, I urge you to work on becoming part of the change to destigmatize mental health.
Be First to Comment