Project Description

Physiological human brain function can be studied almost exclusively in a non-invasive manner, as invasive recordings from the human brain require stringent clinical indications and also because the available recordings typically stem from pathological brain tissue. However, the essence of how the brain works lies in the way that neurons self-organize in local and more extended networks. The human Magneto encephalogram (MEG) is a non-invasive measurement of the electrical activity in the brain and recent advances have allowed the localization of such activity to specific brain regions and the reconstruction of high Signal to Noise Ratio signals. Some functionally important phenomena such as high frequency oscillations playing a major role in visual perception have been first described in invasive animal recordings such as the Local Field Potential (LFP)- a measure of the activity in local networks of 102-4 neurons- and spiking data-single neuron activity- and subsequently analogues have also been observed in human MEG reconstructions-activity in brain regions summing other 105-6 neurons. In this project we will utilize already existing single unit, LFP data in the macaque and human MEG data that were acquired in the same visual stimulation paradigm. We will then use neuronal network modelling (mathematical neuronal models) to study the observed connections between unit activity(microscopic level), local networks (mesoscopic level) and entire brain regions (macroscopic level).Most importantly we will derive the models of the observed macroscopic phenomena (high frequency oscillations in brain regions such as the visual cortex) that will be informed by the invasive data, which include information on underlying layer and columnar cortical circuitry, which cannot be accessed in the human.