In this MEDICA-tradefair.com interview, Dr. Lisa Klingelhöfer talks about the European project, which aims to harness big data and smart devices to facilitate the early detection of Parkinson's disease.
Dr. Klingelhöfer, what is the iPrognosis project all about?
Dr. Lisa Klingelhöfer: At the heart of the project is a smartphone app that is being developed by our consortium. It aims to detect and measure the motor and non-motor symptoms of Parkinson's disease. This does not require any extra active involvement of the smartphone user and takes place while he goes about his/her everyday life while using the smartphone. In doing so, we hope to identify individuals who experience early symptoms of Parkinson's disease, prompting them to consult a physician at a much earlier stage. We collect this data during the GData phase, which stands for "general usage data" and has been in place since May 2017.
What data do you aim to collect with the iPrognosis app?
Klingelhöfer: A main symptom of Parkinson's disease, which has many effects, is bradykinesia and hypokinesia, slowed movement and reduction in the size of movements. A specially developed keyboard within the iPrognosis app measures the speed, regularity and pressure the user exerts to switch between keystrokes when typing messages for example. A gradual slowing down of movement might also cause a reduction in distances an affected user covers each day. That’s why we also conduct a distance analysis. There is a decrease in facial expressions as well. Especially with the younger generation, we can analyze this via selfies, which might make this an important function as it pertains to generational change. Another main symptom is tremor, which we measure via the smartphone's acceleration sensor every time users hold the smartphone when they make a call or type messages for example.
Parkinson's disease typically also causes speech to be more monotonous and softer. That’s why we conduct a voice frequency analysis during phone calls, without being able to understand the context of the conversation.
Lastly, we also try to monitor the mood of the user, because scientific studies have shown that the onset of Parkinson’s disease is often preceded by depressive mood disorder and depression. To analyze messages, we have developed a kind of dictionary to encode words with numeric values. Neutral words like "tree", "car" or "bottle" are valued at "zero", while emotional words are either given a positive or a negative value. This adds up to a numerical code from which we hope to identify a change in emotions over time.