García-Magariño et al. use smartphones in their research as tools to support the diagnosis of Parkinson’s disease. Smartphones are used in several studies with clinical assessments based on the Unified Parkinson’s Disease Rating Scale due to their capabilities to collect and analyze data on a daily basis. This research is based on the presumption that tremor is the most common motor disorder of Parkinson’s disease. Early detection of Parkinson’s disease may help in selecting appropriate treatments, and it plays a crucial role in managing and treating the disease. Some studies have used wearable sensors as accelerometers attached to specific body segments in order to monitor patients in daily activities and detect tremors in an unconstrained environment. However, from the patient’s perspective, smartphones are easier to use than specialized wearable devices that are attached to the patient’s body. Thus, the authors found a way to detect tremors through laboratory measurement methods and were able to exclude inconvenient specialized devices. With the authors’ approach, patients utilize their smartphones in daily activities and, by using the hand-trembling detector application, involuntary hand tremors, distinguished from activities of daily life, are detected.
Related works on the methods based on detecting tremors using electromyography without sensors as well as accelerometer-based detectors for tremors are presented in order to instruct readers in basic methods for the assessment of tremor activity in Parkinson’s disease. Further, the brief insight in detection of tremors with smartphones that include accelerometers is also presented as the basis for general discussion of the application development and its implementation as a smartphone application for detecting tremors. The authors develop the algorithm that detects hand tremors based on parameters that determine which kinds of shakes will be detected including number of shakes, still acceleration, maximum inclination of the device, minimum and maximum intensity of a shake, and minimum and maximum interval from one shake to another. The current algorithm has a function that listens to the accelerometer changes, and it is clearly presented by the included dataflow diagram. Special attention is given to the detection of real tremors and filtering out tremor-like activities of daily life.
At the end of the study, readers will find a detailed description of the application implementation that could be used on any Android-based device with an accelerometer sensor. The deployment activities are accompanied by a performance evaluation and cross-validation that includes the appropriate calibration of detector parameters. This process requires training the system with a representative amount of data, taking such parameters as sensitivity, specificity, and accuracy into consideration.
With the novelty of the approach and the measurement of tremor occurrence with a smartphone application, this study could help other researchers gain new findings in the early diagnosis of Parkinson’s disease.