Recent technical advances have enabled the simultaneous recording of brain signals from over 100 recording sites during behavioral tasks in monkeys. From these signals, action potentials — also known as spikes — of numerous neurons can be detected. This expanded access to whole populations of neurons has driven a paradigm shift from studying behavior-related activity in single neurons to examining changes in collective neural activity patterns, also known as neural population dynamics. This shift, however, also required new analytical tools named "dimensionality reduction methods," to reveal behavioral-dependent patterns in collective neural activity.
Using simultaneous recordings in conjunction with these methods has provided a variety of new insights, such as a better understanding of movement initialization, movement control, and decision-making processes. We use and develop large scale simultaneous recordings together with dimensionality reduction techniques to investigate the neural population dynamics underlying hand-grasping movements, the control of brain-computer interfaces (BCIs), and decision processes across various brain regions.
Recording large neuronal populations also enables the exploration of network interactions of theses neurons. A variety of different network interaction analyses allows to estimate co-activations of neurons as well as inferring anatomical connections between neurons. Using such methods, we study the structure of neural networks within and between brain regions and their relevance for behavior-related information processing. Furthermore, we investigate how rhythmic activity within these networks influence network communication and processing.