Showing papers by "Changsong Zhou published in 2018"

Mathematical Modeling for Description of Oscillation Suppression Induced by Deep Brain Stimulation

Abstract: A mathematical modeling for description of oscillation suppression by deep brain stimulation (DBS) is explored in this paper. High-frequency DBS introduced to the basal ganglia network can suppress pathological neural oscillations that occur in the Parkinsonian state. However, selecting appropriate stimulation parameters remains a challenging issue due to the limited understanding of the underlying mechanisms of the Parkinsonian state and its control. In this paper, we use a describing function analysis to provide an intuitive way to select the optimal stimulation parameters based on a biologically plausible computational model of the Parkinsonian neural network. By the stability analysis using the describing function method, effective DBS parameter regions for inhibiting the pathological oscillations can be predicted. Additionally, it is also found that a novel sinusoidal-shaped DBS may become an alternative stimulation pattern and expends less energy, but with a different mechanism. This paper provides new insight into the possible mechanisms underlying DBS and a prediction of optimal DBS parameter settings, and even suggests how to select novel DBS wave patterns for the treatment of movement disorders, such as Parkinson’s disease.

Repetition priming effects for famous faces through dynamic causal modelling of latency-corrected event-related brain potentials.

Abstract: Repetition priming, that is, the repeated processing of a stimulus, facilitates performance. However, the neural underpinnings of repetition priming for famous faces in terms of effective connectivity are not known. Here we investigated this problem using dynamic causal modelling of latency-corrected event-related brain potentials (RERPs). Source waveforms of RERP-derived sources in the Occipital Lobe, Fusiform Gyrus, Mediotemporal Lobe, Prefrontal Cortex and Anterotemporal Lobe of each hemisphere entered into models with only forward (F) or also with backward (FB) connections. Based on the framework of predictive coding formulated for repetition suppression, modulations of F and FB connections were expected for sources that displayed priming effects in their source waveforms. Hence, neural sources in each hemisphere were fitted with either F or FB connections. Inter-hemispheric connections were considered between homologous areas and were allowed to be modulated in an incremental manner resulting in a model space that comprised of 24 models. Bayesian model averaging across models revealed effective bidirectional connectivity between the Fusiform Gyrus (face perception) and Prefrontal Cortex (decision-making) in both hemispheres to be modulated by priming. In the left hemisphere, there is also a substantial involvement from the Mediotemporal Lobe, indicating the facilitation of automatic retrieval of the famous person's name. Furthermore, there is evidence that the priming is supported by connections from the right to the left Fusiform Gyri possibly in the service of inter-hemispheric cooperation. Altogether, the study indicates that along with top-down modulations, efficient processing within and across the two hemispheres is crucial for famous face priming.