…Continued from “Parkinson’s Diagnosis”…

The program works in much the same way as our IVR speech recognition software here at Plum.

The computer listens to the speaker and converts the sound into hard data. In our IVR call centers, the computer uses this data to synthesize meaning (understanding the caller) and from that arrives at the appropriate action or response.

Unlike a traditional IVR system, Little’s algorithm takes the vocal data and looks for speech patterns characteristic of Parkinson’s disease, like shortness of breath and vocal tremors. From the results and frequency of matches, the system offers a diagnosis.

In blind testing, Little reports an 86% accuracy rate when using his vocal recognition software and a 99% accuracy rate among those in the mid to late stages of the disease.

The beauty of Little’s system is that placing a phone call takes far less time and energy than an in-clinic assessment, is non-invasive and can be completed from the comfort of home.

What’s more, patients can call in regularly to the system, giving doctors more data to make individualized treatment decisions, and the program can quickly and easily reach out to mass-recruit patients for new treatment trials in the search for a cure.

This means fewer trips to the doctor’s office, more accurate care and wider testing.

In 2011, 75% of the entire world’s population was connected to the mobile cellular network, which means there is the potential to quickly and easily screen about five billion people for Parkinson’s with just one quick, local call each.

By creating an automated IVR system with this vocal recognition software, researchers will be able to locate, diagnose and hopefully treat individuals with Parkinson’s while fast-tracking progress towards a cure.

To experience it for yourself, you can go to their webpage and make a three-minute phone call to input your voice into the system. With the addition of more example voices (of callers both with and without Parkinson’s disease), the software will soon hopefully push its accuracy rate up from 86% to 99%.