Differentielle Auswirkungen der subthalamischen Stimulation bei Patienten mit idiopathischen Parkinson Syndrom

Idiopathic Parkinson’s Syndrome is one oft he most important indications for deep brain stimulation. This invasive intervention makes it increasingly possible to achieve better control of motor and non-motor symptoms. Due tot he progressive neurodegenerative character oft he disease, it is necessary to continuosly adapt the therapy with a restructuring oft he electrical stimulations. An objective and efficient methodology ist o be introduced by integrating technical innovations and machine learning processes into these settings, which are still very detailed and demanding for examiners and patients. Against the background that more and more devices such as smartwatches, smartphones, bracelets and others have found use in people’s everyday lives, the registration of symptoms and movement data oft he patients can also take place independently oft he respective institute. This can result in a long-term benefit in Parkinson’s treatment. The monitoring and the objective data would make it possible to treat the individual characteristics of the clinical picture for each individual case. Taking economic aspects into account, this could also bring a significant advantage. In the long term, the use of self-adjusting automated systems in clinical practice can mean another major step for all involved. In the present experimental work, 32 patients with idiopathic Parkinson’s syndrome who are treated with an implantation of a deep brain stimulation system on the subthalamic nucleus were examined. A clinical evaluation of specific movements such as diadochokinesis, finger tapping, elevation of the forearm, resting the arm on a surface, and rigor was performed. In addition, these were recorded in parallel with an examination designed for this purpose using peripheral sensors on the upper extremities. These movements were repeated for the 16 contacts of a directional electrode system while increasing the stimulation amplitude in 0,5 mA steps. The data collected in this way could be succesfully further differentiated, whereby defined characteristics could be filtered out. In a further step, these were used to contruct a random forest model fort he creation of an algorithm. Using these, it was possible to more precisely depict the relationship between the clinical rating values, the objectively measured IMU data and the setting parameters oft he electrodes. The extent to which these values created using the model correspond to the real values could be determined more closely by means of statistical correlation studies. There was significant agreement here. Furthermore, this could be confirmed by analyzing the Dice coefficients in a consistent imaging representation oft he clinically determined stimulation settings and those constructed by the model. In summary, it can be said that a model fort he creation of paradigms fort he present DBS systems could be successfully generated, which correlates significantly with the values from clinical studies. From this it can be concluded that it is possible to acquire objective data using commercially available sensors with an inertial measuring unit. The conclusion ist that by collecting IMU data via sensors on the forearm, it is possible to reliably collect objective data. An algorithm from this was able to generate significantly correlating results tot he clinically collected and evaluated data in the present work. It is optimistic that this can be contributed to further research on an automated DBS system and possibly serve as a supporting contribution for such.