Shop : Rezensionsexemplar
Shop
Rezensionsexemplar
Christoph Dinh
Brain Monitoring
Real-Time Localization of Neuronal Activity
With its millisecond temporal resolution, Magnetoencephalography (MEG) is well suited for real-time monitoring of brain activity. An MEG system with real-time analysis software would offer many new possibilities. For example, the timing and other characteristics of stimuli could be adjusted during the experiments on the basis of ongoing and stimulus-related brain activity. In addition, the analysis workflow, especially in clinical studies, could be made more efficient.
Two remarkable challenges exist in real-time analysis: the low signal-to-noise ratio (SNR) and the limited time available for computations. This thesis presents novel approaches to overcome these challenges. Low SNR reduces the number of reliably distinguishable sources. This thesis proposes an approach, which downsizes the source space based on a cortical atlas and allows to discern the sources in the presence of noise.
Using the clustering approach, two source estimation methods, one distributed and one sparse, were adapted for real-time source localization. The standard dynamic Statistical Parametric Mapping (dSPM) was tailored to provide distributed source estimates. Real-Time Clustered Multiple Signal Classification (RTC-MUSIC) was developed to obtain in addition to sparse localization results correlation information to infer connectivity metrics between cortical regions.
This work shows that real-time source estimations based on MEG are a feasible, useful addition to the standard on-line processing methods, and enables feedback based on neuronal activity during measurements.
In the scope of this thesis the new open source MNE-CPP real-time processing framework was also developed, which together with the novel algorithmic developments paves the way to non-invasive brain monitoring.
Two remarkable challenges exist in real-time analysis: the low signal-to-noise ratio (SNR) and the limited time available for computations. This thesis presents novel approaches to overcome these challenges. Low SNR reduces the number of reliably distinguishable sources. This thesis proposes an approach, which downsizes the source space based on a cortical atlas and allows to discern the sources in the presence of noise.
Using the clustering approach, two source estimation methods, one distributed and one sparse, were adapted for real-time source localization. The standard dynamic Statistical Parametric Mapping (dSPM) was tailored to provide distributed source estimates. Real-Time Clustered Multiple Signal Classification (RTC-MUSIC) was developed to obtain in addition to sparse localization results correlation information to infer connectivity metrics between cortical regions.
This work shows that real-time source estimations based on MEG are a feasible, useful addition to the standard on-line processing methods, and enables feedback based on neuronal activity during measurements.
In the scope of this thesis the new open source MNE-CPP real-time processing framework was also developed, which together with the novel algorithmic developments paves the way to non-invasive brain monitoring.
Schlagwörter: Computational Neuroscience; Inverse Problems; Neuroimaging
Shaker Verlag GmbHAm Langen Graben 15a52353 Düren
Mo. - Do. 8:00 Uhr bis 16:00 UhrFr. 8:00 Uhr bis 15:00 Uhr
Kontaktieren Sie uns. Wir helfen Ihnen gerne weiter.
