• 💖 [Donate To Keep MyPTSD Online] 💖 Every contribution, no matter how small, fuels our mission and helps us continue to provide peer-to-peer services. Your generosity keeps us independent and available freely to the world. MyPTSD closes if we can't reach our annual goal.

Study PTSD can be diagnosed by computer with 89% accuracy through speech analysis

Not open for further replies.
It's an interesting premise - though it's also hard to imagine extrapolating this approach to a broader population. Speech patterns of the English language are subject to different tonal habits based purely on socialization as well as other factors (multi linguistic capability, situational stressors, gender identification...).

That being said, some of what they learned jives with existing observable phenomenon:
The selected RF used 18 speech features and the receiver operating characteristic curve had an area under the curve (AUC) of 0.954. At a probability of PTSD cut point of 0.423, Youden's index was 0.787, and overall correct classification rate was 89.1%. The probability of PTSD was higher for markers that indicated slower, more monotonous speech, less change in tonality, and less activation
(bolding for emphasis)

Slowed speech, lack of tonal variation, passive speech....these are all diagnostics that can apply to other mental health disorders (depression leaps to mind) - so it's not necessarily surprising that someone with PTSD would demonstrate them.

Speech samples were obtained from warzone-exposed veterans, 52 cases with PTSD and 77 controls, assessed with the Clinician-Administered PTSD Scale. Individuals with major depressive disorder (MDD) were excluded. Audio recordings of clinical interviews were used to obtain 40,526 speech features which were input to a random forest (RF) algorithm.

So - for a study with very specific criteria (war zone-exposed vets), and a total of 129 participants, I can see how the correct rate of 89% is feasible.

It's just a long way from that to developing a way of diagnosing any form of PTSD from an algorithm that sorts through recordings of the sufferer's voice. But I appreciate the research, because this:
The diagnosis of posttraumatic stress disorder (PTSD) is usually based on clinical interviews or self-report measures. Both approaches are subject to under- and over-reporting of symptoms. An objective test is lacking.
Is most certainly a frustration for many.
Some additional information concerning this study:

The Military Wants Better Tests for PTSD. Speech Analysis Could Be the Answer

"Post-traumatic stress disorder has long been one of the hardest mental health problems to diagnose because some patients try to hide symptoms while others exaggerate them."

"The most common method currently used for identifying PTSD relies largely on a subjective process in which patients answer questions about symptoms and clinicians make a judgment. That works well when patients report symptoms accurately, Marmar said, but — consciously or not — patients are notoriously unreliable."

"just one of a number of objective methods for diagnosing PTSD, including testing blood or saliva for stress hormones, that are on the cusp of being reliable enough to use in a clinical setting."
Makes sense. My t says she knows when i'm in pstd mode because I talk about horrific events like I'm ordering a coffee. It's a complete change in my speech pattern.
This! Because nothing awful in the past feels like it matters compared to the second my body hits fight or flight mode. If freeze and I’m mumbling and scrambling for focus for awhile. Cptsd long term emotional abuse triggers for me. after years of therapy many triggers have receded but dang are there a ton.
Not open for further replies.