SciTechBits – Interesting Bits of Science & Technoglogy Served For You!

Researchers have designed a virtual reality based piano, which can be used to train stroke survivor’s finger motion. The piano actively engages the patient and slowly makes them produce more degrees of motion.

Stroke is the largest cause of long term disability in America and affects about 5.7 million people. Only 5% of them are able to regain their hand & finger controls. While some stroke survivors experience hemiplegia – which is the total impairment of one side of their body, some survivors experience hemiparesis – which is weakness in one side of the body. A system designed to produce hand & finger motion, in a continuous and interactive manner and thus making the target skill level progressively higher is proved to avail better and quicker rehabilitation of the hands & fingers.

With these in mind, a group of researchers from New Jersey have designed a “Virtual Piano Trainer” to provider finger motion training for people with hemiparesis. The system comprises of a screen – which shows a first person view of an avatar playing a 64 keys keyboard, a glove fitted over both hands – which can read every fine movement produced in the hands & fingers, an arm tracking device over both the arms and a haptic feedback & exoskeletal device over the dorsum of the hand (on top of the gloves) – which is capable of producing a force on each finger to either cause resistance or motion.

The VR system provides cues for the song/scales being played, by highlighting a key on the virtual keyboard and the corresponding finger to be used. The patient then has to move only that finger down. The number of degrees the active finger flexes (bends) as compared to the most flexed inactive finger, gives a quantitative measure (fractionation angle) of how much the patient was able to press the key. If the flexion of the active finger does not go beyond the most flexed active finger, the score would be a zero or –ve. The algorithm keeps adjusting the targeted flexion, so that the user feels challenged but not frustrated. The VR software constantly interacts with the glove and arm band to keep track of the fingers, hand and arms position and suitably adjusts on the screen. The exoskeletal system provided haptic feedback (a small resistance upon pressing a key) to imitate the actual feedback got from a piano. This haptic feedback, combined with audio and video feedback, provides adequate engagement/immersiveness.

However, most patients with hemiparesis would tend to flex all or multiple fingers as compared to a single finger. In such situations, the VR software provides signals to the exoskeletal device, which then offers a force (resistance) to the other fingers of the hand and thus helping the participant to move only the actively needed finger. Over trials, the assistive force offered is slowly reduced as the patient picks up the movements of single fingers.

The researchers measured two different performance parameters. Accuracy – a measure of the number of correct key presses in the first attempt for each cue, over a single trial and duration – the average time taken for a key press in each trial. A group of 4 patients showed a collective improvement of 14% in duration and one particular participant showed an impressive 116% improvement in duration. The biggest improvement in accuracy was about 50%. Consistently, patients performed better when training with one hand as opposed to two hands. Patients also showed improvement in the fractionation (ability to move fingers individually) score, ranging from 19% to 61%. There were some variations in the scores based on the various training algorithms used.

This technology is very promising and engages patients more than just sitting with a therapist and stretching the hands looking at a wall. More thorough testing with a larger group of patients would help us know even better about the actual potential this system has.

Journal Reference & Image Credit: doi:10.1186/1743-0003-6-28