| dc.description.abstract | The brain is a highly dynamic system, performing complex operations at the millisecond scale. Cognitive processes studied in human participants manifest the same dynamicity through behavioral observations. Electrophysiology, through electroencephalography (EEG), can be used to understand the neural correlates of these cognitive and behavioral dynamics. Crucially, these methods function as more than just a mechanism for probing basic scientific questions. Electric signals are measured for medical diagnosis and research. In fact, many neurological disorders manifest changes in EEG. However, what EEG is doing for neurology and psychiatry is arguably not yet as powerful as, for instance, what the electrocardiogram does for cardiology. This is because we do not have a definite understanding of how electrical fluctuations generated in the brain comprise the EEG signal. As such, resolving the neurobiophysical generation of EEG represents a core objective for the field of neuroscience. We sought to resolve the neural generator of one EEG component implicated in cognitive processing, and more specifically, attentional selection. We employed a unique combination of neurophysiological recordings across the layers of visual cortical area V4 simultaneous with EEG. Macaque monkeys were trained to perform an attention-demanding task during recordings known to elicit the attention-associated component known as the N2pc. We found that the synaptic activity across the layers of V4 was sufficient to predict the N2pc. Moreover, biophysically plausible forward modeling of the empirical data demonstrated that the same synaptic currents were sufficient to recapitulate the canonical distribution of the N2pc. Further modeling demonstrated that other candidate brain areas for the generation of the N2pc were insufficient. Yet, cortical columns do not just serve to generate EEG, they also represent a canonical feature of neurobiology and seemingly play an important role as a fundamental unit of neural computation. In visual processing, the organization of columns across the cortical surface encodes information regarding the location and identity of objects. Accordingly, in our experiments this functional architecture predicted relative contributions of individual cortical columns to the overlying N2pc-EEG signal. This suggests that nuances of cortical organization influence the relative contributions of individual cortical columns to the global EEG signal. In sum, this dissertation accomplished two aims that propel our understanding of EEG forward. A combination of experimentation and computational modeling demonstrated that cortical columns in primates can generate EEG signals. Also, in evaluating the functional architecture of cortex we reveal new discoveries underpinning the mesoscopic organization comprising EEG production and attentional modulation, more generally. From these results, we gained insight into the origins and interpretative value of EEG signals, shedding new light on findings past, present, and future. | |
