Thursday 29 May 2014

Further ideas for leapmotion

So, following the excitement of the NHS hackathon I was at recently, it has been difficult to resume the normal medical pattern of studying 7am - 9 pm. Fortunately Monday was a bank holiday, so nothing was "officially" expected of me. It is also fortunate that we are currently in the neuro modules, which is something I semi-understand a bit of already.

A number of people have expressed excitement at the project - both clinicians and researchers as well as students. Most recently today at a placement with the INS I spoke to one of the Occupational Therapists who was interested in the potential. So far my ideas had focused on the potential for detecting severity and types of tremor in a patient, with maybe the development of games for physiotherapy.

On the way back from the INS placement I was discussing it with a fellow student, and between his feedback and what we had seen during the day, an idea crept in regarding people with Parkinsons.

The disease presents with a severe tremor, such that eating and drinking or dressing and going to the toilet are made very difficult. They also find it difficult to write, a sign that medication is working is regaining the ability to sign your name. Yet it is often still difficult to hold a pen dexterously and typing is similarly difficult.

Now, the app. If it were possible to map out a Parkinsonian tremor and tell the computer what to expect, then someone could trace out the letters they want in the air and, even if their tremors made the writing illegible, by allowing for this the computer could discern where the hand "intended" to go and react accordingly.

Possibly a simpler method (no complicated algorithms) - a virtual keyboard that moved itself relative to the finger position. Since the Leapmotion software isolates each finger, it knows which one has moved to where and can move the virtual keyboard accordingly. Even if ten-fingered typing were still impossible and it only worked with one finger, it could still improve communication.

Of course, this requires a library of different tremor types first which as far as I'm aware does not yet exist - but that is my summer plan. It may also require a more in depth knowledge of how tremors present, but I'll be at the weekly neurology grand round at the hospital tomorrow so I will pick the brains of the learnèd faculty.

Monday 26 May 2014

NHS Hackathon

I've just had an awesome weekend of awesome. Honourable mention in the NHS hackday, won a t-shirt and surgical book, and was a far cooler bank holiday weekend than...for instance...playing in the sun. I met lots of awesome people, and if anyone wants to see the presentation it is here over at google docs

And a video of it (skip to 11:45)


We were an honourable mention (the first honourable mention, which I count as runner up ;) and I have a team photo (as well as other photos of the day) here.
 
(Edit - I'm putting in here a better discription of what we did - apologies, I've spent too long staring at medical textbooks and not long enough writing creatively)

Hackathons are a day where all manner of people (programmers, business managers, mathematicians etc) get together to discuss problems and find solutions. In this case, I got attached to a group of programmers who were looking for possible uses of a Leap Motion device  (https://www.leapmotion.com/). The device uses infra red to detect the position of the palm of the hand and all five fingers and records the position.

Working with the team over the weekend, we discovered how it could be used to measure tremors in the hand. Amongst the possibilities this opens, it can be used to objectively record a patient's progression over time, whether it is monitoring a disease such as Parkinsons, assessing the effect of medication or rehabilitative physiotherapy. So far as I am aware, there is no clinical objective measure on the severity of tremor, there only seems to be approximations based on quality of life measures.

The possibilities of what we created seem colossal, and I was excitedly talking to other clinicians about how neurological diseases could be monitored, and how GPs, on seeing a new patient can exactly track the extent of a disease, or when transferring patients around a hospital, a doctor can look at a graph and see instantly the extent of a patient's tremor.

It opens possibilities for further subdividing tremors which currently consist of "does it interfere with a patient's life or not", and (excitingly) it may uncover new types of tremor to subdivide existing disease classifications, discover new syndromes, and find ways to assess the early onset of conditions.