heonslee.github.io

Heonsoo Lee

Works - Education - Publications - Talks

Computational Neuroscientist (github - linkedin - google scholar)

Photo

I am a computational neuroscientist, analyzing and simulating neuronal firing activity and EEG data. My research focuses on dynamical patterns in unconscious states (anesthesia, sleep, and disorders of consciousness) and their impact on machine learning tasks of spiking neural networks. Recurrent spiking neural network is constructed and forced to generate dynamic patterns of conscious and unconscious state while performing various machine learning tasks. This will suggest how and why slow oscillations and ordered neural network dynamics are highly associated with unconsciousness or cognitive deficits.

My past research focused on mechanisms of loss and recovery of consciousness. The investigations covered from firing activity of neural networks at the mesoscopic level to multi-channel EEGs at the macroscopic level. I postulated that the conscious brain operates near critical state (i.e., being at the boundary between order and disorder) and anesthetized or damaged brains would be far away from critical state. My approach included measuring statistical interdependency and information flow between brain signals (and between neurons), graph theoretical analysis, nonlinear dynamical methods, and computational modeling of brain signals. I have also worked on development of medical device for monitoring consciousness. A novel analytic method for measuring connectivity complexity between brain regions was proposed, and further developed as a brain monitoring algorithm. Its clinical usefulness under anesthesia has been demonstrated.

works

education

publications

talks