Neuron-level dynamics of oscillatory network structure and markerless tracking of kinematics during grasping

by Swathi Sheshadri
Doctoral thesis
Date of Examination:2020-12-01
Date of issue:2021-11-18
Advisor:Prof. Dr. Hansjörg Scherberger
Referee:Prof. Dr. Alexander Gail
Referee:Prof. Dr. Michael Wibral
Persistent Address: http://dx.doi.org/10.53846/goediss-8960

 

 

Files in this item

Name:GGNB2020_NetworkDynamicsAndKinematics_Sheshadri.pdf
Size:23.9Mb
Format:PDF
Description:Neuron-level dynamics of oscillatory network structure and markerless tracking of kinematics during grasping

The following license files are associated with this item:

Abstract

English

Oscillatory synchrony is proposed to play an important role in flexible sensory-motor transformations. Thereby, it is assumed that changes in the oscillatory network structure at the level of single neurons lead to flexible information processing. Yet, how the oscillatory network structure at the neuron-level changes with different behavior remains elusive. To address this gap, we examined changes in the fronto-parietal oscillatory network structure at the neuron-level, while monkeys performed a flexible sensory-motor grasping task. We found that neurons formed separate subnetworks in the low frequency and beta bands. The beta subnetwork was active during steady states and the low frequency network during active states of the task, suggesting that both frequencies are mutually exclusive at the neuron-level. Furthermore, both frequency subnetworks reconfigured at the neuron-level for different grip and context conditions, which was mostly lost at any scale larger than neurons in the network. Our results, therefore, suggest that the oscillatory network structure at the neuron-level meets the necessary requirements for the coordination of flexible sensory-motor transformations. Supplementarily, tracking hand kinematics is a crucial experimental requirement to analyze neuronal control of grasp movements. To this end, a 3D markerless, gloveless hand tracking system was developed using computer vision and deep learning techniques. 
Keywords: Network Neuroscience; Markerless grasp tracking; Computer Vision and Deep Learning for behavior tracking
 
Statistik