Validierung der automatischen Analyse periodischer Beinbewegungen in der Polysomnographie

Introduction Periodic leg movements during sleep, as well as their pathology in the form of PLMD, characterize a previously underdiagnosed disorder of the sleeper, which, by occurring simultaneously with a sleep disorder, can severely reduce the quality of life of those affected. Most of them are unconscious movements during sleep, the controlled comorbidities and the differential diagnoses often become subjective attitudes. The occurrence in the specialists in pediatrics as well as neurology, highlights the far-reaching, demanding people of all ages. The social development under the demographic change, the justification of the subjects of the view in the future. Question Automatic methods that are currently used in the analysis of periodic leg movements sometimes have inaccuracies, the results of which can lead to incorrect interpretation of the results. This work was created with the question whether this problem is actually due to the software. Research and testing should be carried out to determine the reasons for any errors. The research work is intended to optimize the validity of the automatic software. Method For this purpose, 20 patients with periodic leg movements were examined retrospectively. The biosignals were evaluated in two ways and then compared. The data were first analyzed using automatic algorithms from the software program. The periodic leg movements were then recorded again by trained personnel in accordance with the AASM criteria. After a detailed comparison of the two analysis methods with a critical look at the real recognition of the movements, the strengths and weaknesses of the automatic analysis could be highlighted. Results The exact comparison showed a clear tendency towards false positive results of the automatic analysis. In total, the automatic software recognized 16271 events, whereas only 11091 events after critical re-analysis met the AASM criteria. As a result, 6461 movement events of the 20 patients can be described as false positive. The statistical evaluation reflects the false positive results in the form of low precision with considerable dispersion (47-72%). The F1 score, as a representative of the accuracy of the measuring method, also indicates the deficits mentioned at 71%. If the events were recorded well in general, a high sensitivity with low scatter (82-92%) was noted. The analysis of the methods in detail revealed systematic errors, the rectification of which can promise optimization and at the same time improved validity of the automatic analysis. Discussion and Outlook With this background and the intention to use a PSG as a search test for PLMD, its validated assessment with suspicion of PLMD can make it easier to assign the movement disorder to individual symptoms. Taking into account any concomitant diseases, an appropriate start of therapy is possible. Misguided and therefore inappropriate therapies can be avoided by validated diagnostics. In the evaluation of periodic leg movements during sleep, as well as sleep disorders in general, a wide range of diagnostics takes place, which takes subjective questionnaires, validated measurement methods and a precise anamnestic history into account. Only when the diagnostic aspects are brought together pathologies of the sleeping person can be explored in this strange state.