ACES profile: Dybuster develops software to treat dyslexia

Markus Gross, co-founder of the ACES award winner Dybuster, didn’t set out to create a software company. He just wanted to help his son Adrian, who has dyslexia.

Gross and his wife had spent a lot of time and money on therapies purporting to improve Adrian’s reading and writing, but nothing had worked. “I saw my son suffering,” says Gross, professor of computer science at the Swiss Federal Institute of Technology in Zurich (ETH Zurich).

Flying back from a conference in the US in 2002, Gross decided to tackle the problem himself. “I thought there must be a way to develop software to treat dyslexia based on our experience as computer scientists,” he explains.

He ultimately found it. Dybuster, the company he co-founded with Christian Vögeli, launched its first patented software product in 2007, quickly selling more than 4,500 German-language copies to parents and therapists. In Switzerland alone, the software generated revenues of almost CHF 1 million (€ 773,555). A school license released last spring, typically covering one site, brings the total number of users to 20,000. Now established as a public company, Dybuster is close to breaking even.

Early on Gross recognised that dyslexics often have remarkable skills. His son has musical talent, for instance, and is a whiz at reading maps. So, Gross wondered, could the areas of the brain that are obviously working well be enlisted to compensate for the regions that are functioning poorly? More specifically, could visual and musical information help dyslexics with one of their biggest problems – matching the sound of a word to the individual letters that spell that word?

Gross sat down at the computer to design software to do just that, using colour, sound and visual patterns to train students to spell spoken words. His idea was to assign a colour and musical tone to each letter, and then divide words into syllables and individual letters in what looks like a family tree (see picture) on the screen. These sights and sounds would help students pick the right letters for words spoken by the software. And the program would be smart enough to use students’ mistakes to adjust the training and focus in on their weakest areas.

Gross developed the basic concept but had no time to develop it. He handed the project to Vögeli, then an energetic Master’s student, “I saw immediately that this could have a big impact,” Vögeli says.

Vögeli was responsible for the complex programming that turned the idea into software. In the main training exercise, the program speaks a word, then the user breaks it into syllables and tries to spell it, constructing a colourful, musical tree on the screen. But would that actually help? To find out, Gross and Vögeli organised a trial in 2006 with 80 children in the Zurich area, 43 of whom had dyslexia. All were given spelling tests. Then, over a period of three months, about half the children used the software for 15 to 20 minutes, four times a week; the other half followed their normal schooling. At the end, they were given another test.

Study shows improvements in spelling scores

The day the data came in was tense. “We knew if there were no improvements, that would have been the end of the project,” Vögeli says. He was hoping the dyslexic children might improve by 10 to 20 per cent. That would have been encouraging enough for him to continue working on the software as a PhD thesis. But the results exceeded his and Gross’ most optimistic expectations.

Without the training, both dyslexic and non-dyslexic students’ performance barely budged. With the software, the dyslexic children improved their spelling scores by an impressive 30 per cent and began to close the gap between themselves and non-dyslexic classmates. Follow-up tests showed the improvement was maintained, raising the tantalising hope that the software had enabled the students to capitalise on the plasticity of the brain, and were now using other areas to compensate for those regions that weren’t fully functional. “We were delighted,” says Vögeli.

So were the parents of children who took part in the trial. “They asked us if the software would be available, and encouraged us to commercialise it,” says Gross. What had been a research project became a joint-stock company, Dybuster, with Vögeli in charge. After raising CHF1.6 million in investment and CHF 800,000 in research grants, the company now has seven full-time employees.

Having completed an English language version of the software at the end of 2010, Dybuster is working on a French one. It’s not an easy business, however. As head of sales, Vögeli is trying to expand into the English-speaking world, either by partnering with an educational tools company, or by finding parents and teachers to spread the word. But he faces a wary public hardened by too many disappointing experiences, and by stiff competition.

English version of the software needs evaluation

“There are about 500 companies out there that claim to have the cure for dyslexia,” according to Nadine Gaab, professor of pediatrics at Harvard Medical School. She has tested a number of them and says few offer any real improvements.

Can Dybuster do better? “It’s not crazy – it could work,” says Gaab, who has looked over Dybuster’s original study. “What it needs is a really good evaluation, beyond the German language,” she says. German is a relatively straightforward language to write, since every letter has one sound. In contrast, notes Gaab, an English letter may have a number of different sounds, helping to explain why the incidence of dyslexia is up to 17 per cent of students in English-speaking countries, compared to 7 per cent in Germany.

Gaab also points out that dyslexia in the English world is diagnosed with a reading test, not writing evaluations. So improvements in writing might not be the right measure of effectiveness.

Another study is about to be published that should bolster Dybuster’s case. In addition, Gross says that feedback from therapists, parents and teachers, indicates Dybuster has improved reading comprehension, “by encouraging students and taking away their fear of words and letters.”

According to one such testimonial from Klaus Ellrich, of Bingen, Germany, “Since working with Dybuster, my son has improved his unsatisfactory grade point average by a whole grade.” If this encouraging result is borne out by further studies, Dybuster may be a winner.

So has Dybuster helped Gross’ son Adrian, now aged 20? Not yet, says Gross, “He got so annoyed being the test case that I never got him to train consistently.” But Adrian may be the reason that some other dyslexics are better at reading.