( Katie Drummond) The Pentagon hasn’t made much progress in solving the PTSD crisis plaguing this generation of soldiers. Now it’s adding new staff members to the therapy teams tasked with spotting the signs of emotional pain and providing therapy to the beleaguered. Only this isn’t a typical hiring boost. The new therapists, Danger Room has learned, will be computer-generated “virtual humans,” used to offer diagnostics, and programmed to appear empathetic.
It’s the latest in a long series of efforts to assuage soaring rates of depression, anxiety and PTSD that afflict today’s troops. Military brass have become increasingly willing to try just about anything, from yoga and reiki to memory-adjustment pills, that holds an iota of promise. They’ve even funded computerized therapy before: In 2010, for example, the military launched an effort to create an online health portal that’d include video chats with therapists.
For now, the system, called SIM Sensei, is being designed for use at military medical clinics. A soldier could walk into the clinic, enter a private kiosk, and log on to a computer where his or her personal simulated therapist — yes, you can pick from an array of different animated docs — would be waiting. Using Kinect-like hardware for motion sensing, a microphone and a webcam, the computer’s software would take note of how a patient moved and how they spoke.
The video [here] offers a demonstration of what a SIM Sensei would look like, and how they’d interact with a patient. In fact, the video is a demonstration of another Pentagon-funded program, called SIM Coach, upon which SIM Sensei will be based. SIM Coach is meant to be used by soldiers inside their own homes, and doesn’t incorporate analysis tech the way SIM Sensei will.
SIM Sensei won’t replace human clinicians. Instead, it’ll supplement them, and help military clinics prioritize which patients need care most acutely, and which can wait to see a flesh-and-blood doctor. If a soldier talking to the SIM exhibits minor symptoms, the Sensei might help him or her schedule an appointment to see a human therapist in two weeks’ time. But if the Sensei detects “red flags” in an individual’s behavior — vocal patterns that signal depression, for example — the SIM could schedule that patient to see a doctor immediately.
“Let’s say you have a more serious case, where it becomes evident to the Sensei that a patient is exhibiting major depression or might be a suicide risk,” Dr. Rizzo tells Danger Room. “The computer could immediately call for a human doctor to come take over.”
The initiative is a collaborative effort between the University of Southern California’s Institute for Creative Technologies (ICT) and Cogito Health, a spin-off company developed by MIT researchers.
ICT has deep experience with virtual therapy. Under Rizzo’s leadership, the institute was the first to develop immersive programs that allowed patients afflicted with PTSD to revisit combat scenarios. The programs have been widely lauded, and are now used by more than 60 military medical clinics across the country.
Cogito’s role, on the other hand, raises something of a red flag. The company was developed out of the lab of MIT scientist Alex Pentland. He’s the number-cruncher whose “reality mining” spurred Darpa to throw millions into a dubious program to mine social data and then yield conclusions about U.S. progress in Afghanistan, known as Nexus 7. The initiative, as Danger Room reported exclusively last year, has been something of a disaster.
Cogito is also grounded in data mining. But the company’s aim is to evaluate a single person’s well-being, rather than an entire community’s. The company will incorporate its bespoke software suite, called “Honest Signals,” into the new Darpa program. It “assesses cues in an individual’s natural speech and social behavior” to spot potential mental health problems, according to a statement that Cogito e-mailed to Danger Room. The company declined to offer studies on the efficacy of “Honest Signals,” but did point to a book — co-written by Pentland — on that very subject.
Rizzo acknowledges that pulling accurate data out of an individual’s face, voice and other such metrics remains a challenge. “We’ve got some heavy lifting ahead of us,” he says. But he’s also extremely confident that Pentland and Cogito are well equipped with data that can turn SIM Sensei into a success. “These guys are bright as hell,” he says. “They’re pioneers in the field, and they’ve got an amazing capacity to detect the smallest problems that pop up in someone’s behavior.”
That said, the SIM Sensei idea is also bogged down by another downside. Computer-based therapy, in comparison to face-to-face treatment, is inevitably impersonal.
Studies on the efficacy of telemedicine (therapy via video chat with a human therapist), where PTSD or depression are concerned, have been mixed. But in aninterview with PBS published last year, Stars and Stripes reporter Megan McCloskey summed up the shortcomings of such therapy for mental health conditions. “Many of those who need more intensive counseling … don’t like the impersonal nature of talking to a TV screen,” she says. “For some, telemedicine doesn’t meet their needs and adds to their sense of isolation.”
Cyber therapy would be even more vicarious. Soldiers will talk to a videogame character, rather than a real person, through their computer screen.
But a robust virtual option would give soldiers, many of whom still shy away from face-to-face mental health treatment, the option to seek solace in a more anonymous alternative. Eventually, Rizzo and his colleagues hope to see SIM Sensei available for soldiers within the comforts of their own home, rather than a military clinic.
“A lot of people still don’t want to stop by the clinic and meet with a real person,” he says. “Technology is ripe for us to leverage. I’m extremely confident that we can use it, leverage it, to help people who otherwise wouldn’t get better.”