September 22, 2005

Games For Health 1

by Michael Mateas · , 9:55 am

Thought I’d do some live blogging from Games for Health in Baltimore.

Steven Downs
Robert Wood Johnson Foundation

Talked about why the foundation is interested in funding games for Health. They fund it under their “emerging health care applications” umbrella – he noted that the irony is that games are not “emerging”, they’re already here.

Bruce Jarrell
University of Maryland School of Medicine

Ben Sawyer introduced him by saying that when he met with Bruce to convince him to host a Serious Games conference on his campus, he didn’t need to do any convincing – Bruce immediately started talking with him about the patient simulations he works on. Medical and patient education needs simulations that exercise the head, not the hands. Current simulations exercise the hands, e.g. surgery simulations. We need “customized, complex, multi-level language and engaging interface.” I’m sitting here thinking “Façade, Façade, Façade…” Simulation won’t have a deep impact on med school education until it’s engaging high-level problem solving, not low-level mechanics.

Ben Sawyer

Two sides of game’s for health: Personal Treatment (game-like applications for patients, e.g. exergaming, braintraining games, VR psychotherapy, pain distraction, Gluco Boy for diabetics), professional practice (modeling, simulation, training, game interfaces for medical applications like pre-op interviews, health messaging). The personal treatment side seems more fleshed out – there are more honest-to-god games that have been developed on that side. Sketches a future vision: you’ll have a personal health record (PHR) that is fed by doctor/medical visits, data recording from various devices. The PHR itself then feeds the selection/generation of customized health games – traits of your online avatar may also be influenced by your PHR. For me, immediately brings up privacy concerns; the PHR must be owned and maintained by the patient, not by some central authority.

Mentions WoW story from last week: “The corrupted blood disease is, in short, out of control and rapidly taking on epidemic status.”

Show’s game introduction curves: a brain training and personal health game show consistent sales for weeks, while traditional games have a big blip, then quickly fall. Brain training is a significant driver for Nintendo DS sales in Japan (above Nintendogs – braintraining is driving 60% of sales (!!)).

Showed video of new one-handed Nintendo DS controller that uses motion sensing. Showed people playing virtual drums, conducting symphonies, fishing, performing dental surgery, by gesturing with the one-handed controller. It’s a marketing video, but show’s some interesting thinking breaking out of the “gamer” box.

Eric Johnston
Lucas Arts

Eric talks about the development of Ben’s Game, a game funded by the Make-A-Wish foundation. Ben Duskin is a Leukemia patient who’s wish was to make a video game that would be helpful for kids who have cancer, that would allow them to fight back, and help relieve pain and stress. Eric read about the wish, talked to his employers at Lucas Arts, and got permission to work on the game. Lucas Arts game him after-hours use of the office and computer, sound studio, legal and PR help for Make-A-Wish, and gave Ben a tour of Skywalker Ranch and the archives. Eric’s first game with Lucas Arts, Pipe Dream (search for Pipe Dream on this page), had one programmer per platform – Eric was the programmer for the Mac version. He enjoyed working on Ben’s Game, as it was a return to the days of small scale development teams like Pipe Dream. Project budget: $87.00 (1 USB Flash Drive, 16 liters of Limeade).

Basic play field is based on game of life (cellular automata); cancer cells mutate and spread across the board. But, for Ben, he doesn’t directly experience the cancer cells, he experiences the nasty side effects. So they added monsters associated with the side effects: Fire monster (fever), Q-Ball (hair loss), Ro-Barf (vomit), Vampire (bleeding), Tornado (rash), Snow Monster (cold), Evil Chicken (chicken pox). By fighting the monsters, you gain tools for fighting the cancer.

The port is a device inserted in the sternum of young cancer patients so they don’t have to keep sticking them with needles. Ben showed Eric the scare from his port, and said “this is how the player enters the game”. This was the genesis of the notion of player as medicine.

The basic gameplay is regulated by having to maintain health, ammo and attitude.

TATRC – Telemedicine & Advanced Technology Research Center
Federation of American Scientists building a research roadmap for virtual humans in health training and education. Includes work on emotion, conversational capability, as well as physiology. The did a workshop on patient scenario generation – the holy grail is to be able to quickly, and at low cost, build an interactive patient scenario. Key areas that come out: AI tutors (knowledge modeling and scaffolding), case generation (bringing in clinical records, etc.), relevance of learning (engagement, meaningful scenarios – hope to leverage design ideas from gaming). Health administrators have a big interest in bioterrorism preparedness.

Two distinct training tracks: physiology simulations for medical interventions vs. conversational training for patient interaction. One guy who builds commercial conversational trainers does video lookup – pre-film all possible conversational snippets and sequence them. Some discussion about the distinction between realism vs. believability. Chatted with Patrick Kenney from ICT about this. He feels that ICT is going the realism route (build people simulations without artistic “tricks”) vs. my work which is about balancing the tension between authorship and autonomy. Referred to the ICT work as “hard AI”. I demurred with this characterization, since what I do is hard; expression as a focus opens up a different, but no less hard, technical research agenda.

Noah Fahlstein is at the conference and is participating in the TATRC sessions. It was nice to finally meet him. He mentioned that Façade has been getting a lot of good buzz among his game design/storytelling friends, though he himself has not had time to play it. Over lunch a group of us had a nice conversation about “smoke and mirrors tricks” in game design (I call this authorship), and how smoke and mirrors still beats the best, and may always beat the best, purely autonomous approaches. I agree, with the caveat that you can build architectures that are specifically designed to support authoring (aka “smoke and mirrors”).

Open-source rendering engines that came up in the conversation:
Delta 3D

During the proposal brainstorming session (suggestions for future TATRC RFPs), Ben brought up conversation engines, that there are no engines currently available, open-source or commercial, that help you author interactive conversations. Going around the room, several people mentioned NLU and conversation as issues. We were asked “what stumbling blocks do you have in your work”. For me, in the short term it is authoring (Andrew and I have talked about this a lot); it is too hard to train new people how to use ABL to write conversational characters. Over the last few years I’ve had to train a number of students on how to use ABL. It’s true that, in a commercial contex, training would be less of an issue because developers would be focused full-time on learning the development framework, but to the extent that students represent “casual” authors, ABL is difficult for causual authors to pick up and use. In the long term, there are a zillion hard AI and design problems (e.g. personality-rich natural language generation, rich declarative representations of conversational state, behavior generation, generative drama management, etc.).

Some discussion about a consortium that is meeting with Valve and Epic about inexpensive source licenses for serious games researchers and developers. Reminds me of the similar game AI consortium that tried this a year or two ago, unfortunately to no avail.

Brian Morrison, Believe in Tomorrow Foundation
Lyn Dahlquist, University of Maryland
FreeDive is a scuba diving simulation designed as a pain distraction tool for children undergoing painful medical procedures. The game was developed by Breakaway Games, a major developer of serious games and funded by Believe in Tomorrow, a foundation whose mission is to “bring comfort, joy and hope to critically ill children and their families.” Games as a pain distracter are based on the fact that pain processing requires cognitive attention; the idea is to distract cognitive attention so that pain is not experience. Testing the hypothesis the interaction is more distracting the passive stimuli (like watching TV). Lyn did a laboratory test, using the number of seconds a child is able to keep their hand submersed in ice water, as the pain stimuli. No distraction: 29 seconds, passive distraction (watching someone else play a video game), 25 seconds, active distraction (playing the game themselves), 65 seconds. So interaction dramatically improved pain distraction. Currently testing FreeDive on its effectiveness to distract from clinical pain, specifically the injection pain experienced by child cancer patients undergoing chemotherapy.

Games for Treating Post Traumatic Stress Disorder (PTSD)
Russell Shilling, Office of Naval Research

Russell was the sound designer for America’s Army. Did research on emotion, and how to evoke emotion in VR, at the Navy Postgraduate School; this lead him to become interested in treating phobias in VR. Currently a program officer running programs on VR mental health applications; with the war in Iraq, the Navy has become more interested in mental health issues. Historical list of PTSD VR treatment systems: Virtual Vietnam, World Trade Center, Terrorist Bus Bombing, Motor Vehicle Accidents, Virtual Angola, Virtual Iraq (ICT project – their current research in this area). In all of these, you slowly expose the PTSD patient to a more and more detailed simulation of the traumatic event. For Virtual Iraq, they are using assets from Full Spectrum Warrior. They’re creating a Wizard of Oz interface for the therapist; the therapist controls the simulation as a function of the patient’s response. Not only interested in audio and visual cues, but also smell; smell powerfully evokes memories. If you pass garbage, for instance, you smell garbage.