3) Complex brain activity and cognitive variability
The brain displays complex, spontaneous ongoing background activity and generates stimulus-induced responses, which are likely to interact nontrivially with each other, since neural network in the brain is a system of interacting nonlinear elements by biophysical and physiological nature. Largely, complex brain activity is analyzed with two groups of methods in order to search for correlates to brain functions. Classical cognitive theory assumes that these two activities are independent of each other and regards the background activity as noise which is eliminated by averaging over many trials of electroencephalographic (EEG) records to obtain the event-related potential (ERP) for a given stimulus condition. Alternative hypotheses assume that the oscillation components, such as alpha and gamma waves, are correlated with cognitive functions; time-frequency and phase are analyzed to search for changes in spectra and phase coherence during cognitive processing. These two branches of approaches and methods pay attention to different aspects of complex brain activity, but both explore only part of the rich information. An evident limitation of ERPs is that they cannot address strong cognitive variability across trials and individual subjects.
In close collaboration with cognitive neuroscientist Prof. Werner Sommer from Berlin Humboldt University, we are interested in analyzing the interaction between ERP and ongoing brain activity, since it is crucial for understanding how the ongoing brain activity impacts on cognitive processing. We take the response variability in cognitive experiments as the window for studying this important problem. There is a long-lasting problem in cognitive neuroscience that cognitive sub-processes, such as stimulus perception, decision-making and motor execution, have strong variability in latency and the conventional averaging ERP method will mix the components associated to the cognitive sub-processes. Recently, we have developed a new method called Residual Iteration Decomposition (RIDE) which can decompose single trial brain activity into several components associated to sub-processes and identify their latencies respectively. This method will have great many applications in cognitive experiments.
Our next steps are to study the nonlinear interaction between ERP and ongoing activity by analyzing the dependence of component latencies on amplitude, phase and synchronization of ongoing activity. This will allow us to obtain much more information from the complex brain activity and its relationship with the functional performance.
We will use modeling to obtain understanding of the mechanism underlying the response variability, which can help us develop new methods to better understand the experimental data.
|G. Ouyang, G. Hermann, C.S. Zhou and W. Sommer, "Residue Iteration Decomposition (RIDE): a new Method to Separate ERP Components on the Basis of Latency Variability in Single Trials" , Psychophysiology, 48, 1631-1647 (2011).|
Back to Research Webpage