Computer scientists from the University of York have teamed up with a world-renowned surgeon to develop better ways of treating children with abnormally-shaped heads. STEPHEN LEWIS reports

WE all like to think there's a streak of the rebel in us, or at least something that makes us that bit different from the crowd.

But the truth is that when it really comes down to it, we just want to fit in.

Most of us are much better these days at covering up our reaction to people who look different, says craniofacial surgeon Christian Duncan.

But an unusual appearance is one of the things that can still trigger very deep reactions - reactions that we may try to mask, but that we struggle to fully control.

Imagine, then, what it must be like to grow up with what would be widely regarded as a physical abnormality. Friends and loved ones wouldn't even notice. But each time you went out in public, you'd be aware of the gaze of strangers - not critical or judgemental, necessarily, but just surprised, and seeing you as different.

As head of surgery at Alder Hey children's hospital in Liverpool, Mr Duncan is one of the country's leading experts on a condition known as craniosynostosis.

This is a condition in which the plates of a young baby's skull fuse together too quickly.

When you're born, the separate bony plates which make up your skull are not fixed firmly together. This is so that your skull can expand as you grow, allowing your brain to develop normally.

If some of the plates of your skull fuse together too early, however, the skull is unable to grow properly. It may remain unusually small; or it may be too long and narrow, or bulge at one side to compensate for restricted growth on the other side.

Craniosynostosis is comparatively rare. But a child growing up with such a condition, if left untreated, can suffer headaches caused by pressure on the brain, and breathing problems. In severe cases the development of the brain itself may be affected. A child may also be teased or bullied, and suffer feelings of low self-esteem.

Mr Duncan and his team at Alder Hey operate on something like 100 craniosynostosis patients every year, essentially rebuilding the patients' skulls so that they are able to grow properly. The surgery usually involves making a cut across the top of a child's head, removing and reshaping the affected parts of the skull, and then fixing them back in place.

His aim, Mr Duncan says, is to 'give children back their anonymity, to make them unrecognisable in a crowd.'

Giving a child the gift of looking normal can make a huge difference to the quality of their life.

The problem surgeons have is how to recognise what is normal. What head shape should they be aiming for when they operate?

Until recently, there was no objective test that could tell a surgeon whether they had created a 'normal' head shape. "The only way was to try to measure the head with calipers," Mr Duncan says. "That gave you an idea of the dimension, but not the shape." Experienced surgeons would resort to their own judgement - or ask the patient or the patient's family. But given the difficult and sensitive nature of the surgery, that was always highly emotive and very subjective.

A few years ago, Mr Duncan realised that what was needed was a large database of images of 'ordinary' human heads to serve as a reference point.

His solution: a unique, interactive installation put together by three graduates of Liverpool School of Art and Design at Fact, the Foundation for Art and Creative Technology in Liverpool.

Over the space of a couple of years, visitors to the gallery were invited to step into a specially-built booth where their heads were photographed from five different angles with the help of a 3-D camera.

Ultimately, more than 1,200 people - a mix of men and women, young and old - had their heads scanned during the Headspace project, producing an extraordinary database of head and face images.

But how to turn these into an interactive, 3-D computer model of a 'normal' human head and face that could be used to help surgeons?

Step forward Dr Nick Pears, an expert in computer science at the University of York.

Four years ago, Dr Pears was at home in York when his wife Maxine called him through to see an item on the news. It was about Christian Duncan and his Headspace project.

Dr Pears, an expert in robotics who had been working on artificial intelligence with his colleague Dr Will Smith and PhD student Hang Dai, could see the possibilities immediately.

"I called Christian Duncan the same day," he says. He and his team have been working with Alder Hey ever since.

The model the teams in York and at Alder Hey have developed between them is known as the Liverpool-York Head Model.

It allows surgeons to call up a range of 'normal' human head shapes on a computer screen, then rotate them and study them from every conceivable angle.

The models can be adjusted to take account of age and sex and even, to a limited degree, ethnicity.

They can show a range of head shapes that we would regard as normal, from broader and shorter to longer and narrower. And they are, Mr Duncan says, invaluable.

At the moment, the model is mainly used in planning and preparing for surgery, and also in assessing the success of an operation.

That last is crucial, Mr Duncan says.

Before, surgeons had to rely on their own sense of judgment - or the judgement of their patients - in deciding whether they had succeeded in creating a head shape that was 'normal'.

But now they have an objective standard against which to measure their work. "It's unique," Mr Duncan says. "It is the only tool we have that let's us say definitively 'yes, that is in the normal range'."

That is important because it means success or failure can be objectively measured - vital for prospective patients and their families who want to choose the best possible surgeon.

The potential of the model goes much further than that, however.

It could ultimately be used to 'guide' surgeons during the complex process of surgery.

And it could also be used to reconstruct the heads or faces of patients other than those with craniosynostosis who are severely injured or deformed.

"There seems to be a relationship between one part of a skull or face and the rest of the skull and face," Mr Duncan says. So even if only a part of the skull or face is intact, the model could be used to reconstruct what the rest of the face or skull should look like. "It could absolutely be used for that."

So unique is the model that it has attracted the attention of Google.

The California-based tech giant gave Dr Pears and his team at York £40,000 to develop the model in 2017/18.

Earlier this year, Dr Pears travelled to Google's ultra-modern headquarters at Mountain View, California, to report back on the progress of the project.

He then gave his presentation to Google staff while his sponsor from Google's site in Cambridge, Massachusetts, watched overt a video link.

His presentation, which covered how the head model had been developed, and how it was being used at Alder Hey to assess the results of surgery, seemed to have gone down well, he says.

And so it should have, says Christian Duncan. "It is very exciting. It is the only model of its kind that currently exists. There's a huge amount of interest in it."