Parkinson’s disease is so devastating that early detection is a boon to recovery when a brake is put on its progression. In this field, Khalifa University at Abu Dhabi, UAE, takes the lead by developing a new tool which uses sensors designed in smartphones.
One percent among those who are above 60 years, suffer from this disease which is termed as the second neurodegenerative disorder that prevails commonly across the planet. The beginning is humble with prominent hand tremor but gradually the complication takes control over the movement of limbs, muscle and also hampers one’s balance. Among those diagnosed with early stage of Parkinson’s disease, Fine Motor Impairment (FMI) is the usual criteria of diagnosis but still such clinical techniques are not considered prime and efficacious enough.
Basking under towering tech diligence which is widely acknowledged, a flock of research luminaries from various nationalities, i.e. Germany, Greece and UK came together to develop a Parkinson’s Disease detection tool. The group also included Dr. Leontios Hadjileontiadis, professor of Biomedical Engineering as the central jewel and their efforts culminated in the development of a tool, which is supreme enough to screen for initial symptoms related to Parkinson’s Disease and can indicate to patients in same way on their smartphones.
The thing is, a framework is rolled out which is deeply infected by the premier technology Deep Learning and which sorts data that smartphones capture during normal use and presents the outcome in Scientific reporting.
Dr. Handjileontiads explains, “Remote unsupervised screening via mobile devices can raise awareness for medical care, with daily assisting diagnosis. User interaction with smartphones can unveil dense and multi-modal data to reveal patterns that can be connected with both motor and cognitive function. In particular, Hold Time, the time interval between the press and release of a key, offers insights to the probability of a subject suffering from Parkinson’s”.
Clearly, we use fingers during smartphone operations and the micro-speed at which our finger pushes an icon and releases the pressure from it, gets captured and quantified and represents our brain’s control over our muscles. Now, in the event body starts shaking our brain’s motor cortex informs the spinal neurons for muscle activation. Now, here Dopamine also plays a pivotal role which is one of the neurotransmitters that triggers series of events leading to feeling, some action or any such movement of our limbs. Now, among such patients, cells that produce Dopamine get inactive and dopamine loss gives way to movement issues. Gradually, patients start showing apparent symptoms alongside tremors, while walking and other such movement issues get to prominence too.
Dr. Handjileontiads claims, “Detecting these smaller tremors at the start of the disease can lead to earlier diagnosis and allow us to implement management strategies earlier. The standard medical practice in diagnosing Parkinson’s Disease requires years of expertise. Using a smartphone provides an unobtrusive way of capturing data as we link keystroke typing with an enriched feature vector to describe the keystroke variables”.
Besides, in smartphones, there is stored a sensor called Inertial Measurement Unit (IMU) which produces acceleration values, in a bid to carefully oversee hand tremors. Further, such also captures data when users operate their smartphones normally, such as to call or to text.
Now, when deep Learning technology churns out this data, a promising system gets into the front seat which records the slight fine motor impairment that simply indicate initial stage of Parkinson’s Disease.
Undoubtedly, the Deep Learning technology has shown its functional precision and effectiveness when crucial patterns and representations are determined complex from data comprising complex dimensions, such as images and research team displayed the promise of Deep Learning in quantifying the typing done on touchscreen which deeply correlates the FMI clinical recording.
Further in the loop, Deep Learning algos can also uncover the Parkinson’s from MRI scans, from tremors that accelerators capture and even from the tempo and tone of our voice and by testing for any degradation therein. Notwithstanding, smartphone based typing can also check the keystroke orientation during our daily and usual activities.
Dr. Handjileontiads came in further, “We tried to detect Parkinson’s Disease using a multi-symptom approach that merges passively-captured data from two different smartphone sensors via a novel deep learning framework. Our method is inspired by the typical workflow of a neurologist, in the sense that it outputs a score for tremor and FMI, two of the most common motor symptoms, as well as a score for Parkinson’s disease.
This is, however, not something new and unique when Parkinson’s is detected in automated way and in parts, in fact, there have been a fistful of sensors to grab certain aspects of distinct symptoms. For instance, microphones for speech impairment, to detect change in gait, we have IMU sensors, keyboards for typing speed and device for writing to record fine motor impairment.
Most commonly Parkinson’s disease is diagnosed through a single symptom, which is again dubious as this problem manifests among patients in many ways and thus any system meant for disease inference needs to cover a range of symptoms before reaching at any conclusion. As such, the achievement of Dr. Handjileontiads indeed deserves great appreciation as this is multi-modal and captures data unobtrusively.
Finally, as far as sensitivity and specificity of disease detection is concerned, it stands at 92.8 % and 86.20% respectively. This highlights that the system not only puts up stellar performance but then, such is scalable to too, to incorporate other data in similar framework, such as speech information.
Not surprisingly, Dr. Handjileontiads is filled with pride, “Performance-wise, our approach produced good classification results and this is the first work to address the problem of detecting Parkinson’s from multi-modal data. This is a solid first step towards a high-performing remote Parkinson’s disease detection system that can be used to discreetly monitor subjects and urge them to visit a doctor signs of disease are detected”.