Combined accelerometer and genetic analysis to differentiate essential tremor from Parkinson’s disease

Abstract

Essential tremor (ET) and Parkinson’s disease (PD) are among the most common adult-onset tremor disorders. Clinical and pathological studies suggest that misdiagnosis of PD for ET, and vice versa, occur in anywhere from 15% to 35% of cases. Complex diagnostic procedures, such as dopamine transporter imaging, can be powerful diagnostic aids but are lengthy and expensive procedures that are not widely available. Preliminary studies suggest that monitoring of tremor characteristics with consumer grade accelerometer devices could be a more accessible approach to the discrimination of PD from ET, but these studies have been performed in well-controlled clinical settings requiring multiple maneuvers and oversight from clinical or research staff, and thus may not be representative of at-home monitoring in the community setting. Therefore, we set out to determine whether discrimination of PD vs. ET diagnosis could be achieved by monitoring research subject movements at home using consumer grade devices, and whether discrimination could be improved with the addition of genetic profiling of the type that is readily available through direct-to-consumer genetic testing services. Forty subjects with PD and 27 patients with ET were genetically profiled and had their movements characterized three-times a day for two weeks through a simple procedure meant to induce rest tremors. We found that tremor characteristics could be used to predict diagnosis status (sensitivity = 76%, specificity = 65%, area under the curve (AUC) = 0.75), but that the addition of genetic risk information, via a PD polygenic risk score, did not improve discriminatory power (sensitivity = 80%, specificity = 65%, AUC = 0.73).

Essential tremor: diagnosis and management

“…The development of consumer grade accelerometry, which can be helpful in distinguishing essential tremor from Parkinson’s disease tremor, may emerge as a viable clinical diagnostic tool as the technology becomes more accessible. Spiral drawings can be performed directly on digital tablets capable of analyzing the characteristics and severity of tremor 909192. Accessibility and cost represent a barrier to the use of imaging and other technology.…”

Future Opportunities for IoT to Support People with Parkinson's

“…One strand of this work has focused on the detection of Parkinson's symptoms such as freezing [26,75], and tremor and dyskinesias [18], through wearable devices-typically utilizing inertial sensors such as accelerometers and gyroscopes to determine movement of the trunk and limbs. Similarly, several researchers have used smartwatches to generate time-based estimates of bradykinesia (slowness of movement) during daily activities [54], to support the differential diagnosis of Parkinson's tremor against other types of tremor during daily activities [37], and to improve medication regimes [20].…”

“…The potential for technology to implement automated and continuous assessment, using environmental and body worn sensor systems, alongside machine learning approaches, has insurmountable potential for making this assessment process more objective. Indeed, previous work has already looked at improving this through the detection and classification of Parkinson's symptoms using wearables [27,75,18,54,37] and mobile based assessment [79,83]. This not only has impacts for our Western society, but also has the potential to improve this assessment process within lower income countries, where specialist services for conditions like Parkinson's are only beginning to emerge [77].…”

Multi-Dimensional, Short-Timescale Quantification of Parkinson's Disease and Essential Tremor Motor Dysfunction

“…Although the textbook clinical pictures of these two movement disorders are distinct, misdiagnosis persists in both directions (45). Some quantitative diagnostic tools designed to distinguish between PD and ET are accelerometer-based and rely on differences in tremor characteristics alone, such as frequency (19,46,47). Others require significant training for proper administration, such as those that use electromyography or transcranial sonography (48,49).…”