David Brent and PIttsburgh and Carnegie Mellon colleagues used machine learning to identify suicidal thoughts in subjects based on fMRI scans.
In a recent study, 18 suicidal participants and 18 members of a control group were presented with three lists of 10 words related to suicide, positive, or negative effects. Previously mapped neural signatures showing brain patterns of emotions like “shame” and “anger” were incorporated.
5 brain locations, and 6 words, were distinguished suicidal patients from the controls. Using those locations and words, a machine-learning classifier was trained and able to identify 15 of the 17 suicidal patients and 16 of 17 control subjects.
Suicidal patients were divided into those that had attempted suicide (9) and those that had not (8). Another classifier was then able to identify 16 of the 17 patients.
If healthy patients and those with suicidal thoughts are proven to have such different reactions to words, this work could impact therapy and, it is our hope, prevent lost lives.
Join ApplySci at Wearable Tech + Digital Health + Neurotech Silicon Valley on February 26-27, 2018 at Stanford University. Speakers include: Vinod Khosla – Justin Sanchez – Brian Otis – Bryan Johnson – Zhenan Bao – Nathan Intrator – Carla Pugh – Jamshid Ghajar – Mark Kendall – Robert Greenberg – Darin Okuda – Jason Heikenfeld – Bob Knight – Phillip Alvelda – Paul Nuyujukian – Peter Fischer