Scientists have developed a new smartphone app which can detect early signs of Parkinson's disease much before people experience tremors.

The app picks up subtle changes in a person's voice, walk, and manual dexterity to diagnose the condition.

"The condition is hard to diagnose, with specialists having to take a detailed history of people's symptoms and analysing them for physical signs of the disease. Using smartphone data may help to make this process much easier," said Dr Max Little, a mathematician from Aston University who is leading the project.

Reportedly, the app is being tested in a group of 2,500 people with diagnosed Parkinson's symptoms, or genes known to be linked to the disease, and healthy individuals. An earlier analysis of a voice recording revealed that the app can detect Parkinson's with 99 percent accuracy from effects such as increased breathiness, drifting pitch and altered vowel sounds.