Image: Bloorview Exterior RenderingImage: Boy and young woman having fun


Stories: Canada invests in engineer’s smart technology

Tom Chau awarded Canada Research Chair for devices that interpret intent behind eye blinks, hand gestures

Imagine your child has lost the ability to speak following surgery, and neither you nor the medical team can decipher why she’s agitated. Is she in pain? Is she cold? Does she fear you’ll leave?

Now picture your child with a serious swallowing problem. Every time you sit down to feed her, your heart pounds as you wait to see if she’ll aspirate.

Tom Chau is finding ways to circumvent these disabilities, bringing together intelligent computer chips and software to interpret the communicative intent behind a critically ill child’s eye blinks or hand gestures, or recognize unique throat vibrations that signal the onset of aspiration, prompting parents to intervene.

In fact, this biomedical engineer at Bloorview – just awarded a Canada Research Chair in pediatric rehabilitation engineering at the University of Toronto – is finding countless ways to use smart technology to enrich the lives of children affected by disability. His goal? “To place the onus for adaptation on technology, instead of the child, so that the technology fits the child and evolves as the child’s needs and function evolve.”

Tom’s research chair represents a federal government investment of $500,000 over five years, which will cover the bulk of his salary while he conducts research at Bloorview. It is a Tier 2 chair, recognizing emerging world-class researchers with the potential to lead their fields.

Tom’s rehab devices are unique because they adapt to a child’s abilities and environment “on the fly” – in the same way that artificial intelligence equips web sites to learn and respond to a customer’s preferences. “Their distinguishing property is that they can self-organize on their own,” Tom explains.

Part of Tom’s expertise comes from his work in data mining at the University of Waterloo, “where he looked at large data sets and teased out information that had relevance to the situation at hand,” says Dr. Mickey Milner, research consultant at Bloorview. “If you think about rehab and movement and intent to do things, it’s even before the child engages in the act that Tom needs to tease out factors that will tell us what the ultimate goal is.”

For example, in the area of prosthetics, Tom is training a computer chip to interpret the sounds that a child’s arm muscles make as they contract, so that those sounds can be used to control an artificial hand or wrist. “Each person has unique muscle sounds and we’re able to teach the computer chip to link certain actions with certain sounds.” The system – which relies on sound and vibration sensors in the prosthesis – must adapt to wide variations in a child’s environment, from a quiet bedroom to a noisy playground. “It needs to be able to weed out environmental noise it hasn’t detected before, so that it doesn’t interfere with the child’s ability to control the device reliably,” Tom says.

In partnership with The Hospital for Sick Children, Tom is building a communication system that will decode the most subtle of movements – an eye blink, head turn or thumbs up – giving a voice to children who have lost their speech in critical or palliative care.

Based on a questionnaire to parents and nurses, his research team devised a core vocabulary of about 60 words “that would be critical for the child to communicate their medical, emotional and family needs. If it’s known that a child will lose his speech pre-operatively, the system could be tailored to what that child’s core messages would need to be, and we can bank children’s voices for when they lose the ability to speak.”

The system – which will run on a laptop or palmtop device with a video camera – will adapt to the needs of the user, organizing the vocabulary menu based on the frequency and importance of messages. “For example, if a child consistently chooses the messages ‘Please don’t leave me’ and ‘Hold my hand’ at the same time, these messages will be presented together,” Tom says. Children will scan and select choices through eyelid, head or hand movements.

When a child is tired, the device will pick up this change and adapt. For example, if the child begins to respond more slowly, be inconsistent or change gestures, key messages may be made more easily accessible or the system may respond at a lower threshhold, accepting less precise or more limited movements.

This is important for children in palliative care who may lose function on a daily basis, Tom says. “One day they may be able to point, and the next they may have to use eye blinks, so it doesn’t make sense to change the ”
Tom supervises biomedical students at the University of Toronto and conducts most of his research at Bloorview, which he describes as “the premiere location if you want to work with children with disabilities.

We see the whole spectrum of disabilities here and that means I can develop systems that have as wide an impact as possible. I also have the support of colleagues from a unique blend of disciplines that you don’t find anywhere else: clinicians, therapists, physicians, dentists and engineers.”

To be connected with expert sources, contact:

Louise Kinross, Manager, Communications
Tel: 416-424-3866
Pager: 416-589-8826
E-mail: media at bloorview dot ca

Back to Stories

  • Family Corner
  • Youth Corner
  • Kids Corner