Neuronal Mechanisms of Perception and Cognition : an anthology of abstract models and experimental deductions in prevailing neuroscience paradigms

Resumen

The thesis explores neural substrates functioning as complex networks in several predominant neuroscience paradigms, including motor imagery, flickering images, picture naming, semantic anomalies, auditory oddball, and motion-onset visual evoked potentials. The corresponding experimental designs and algorithms for recording electrophysiological data, such as magnetoencephalography (MEG) and electroencephalography (EEG), obtained during experimental tasks by the subjects are presented. We use standard tools and develop new physical and mathematical methods for data analysis. Models of underlying mechanisms of perception and cognition are discussed, proposed, tested, and compared with modern approaches and our own experiments. Key models / mechanisms include neural communication during motor imagery, generalised perceptual models, coherence resonance in visual perception, and the use of a neural network for object recognition. New brain-computer interface (BCI) applications are introduced, and existing systems are improved. Key suggestions include proper signal filtering for using artificial neural networks in BCI to classify imaginary movement, model-free estimation of brain noise, efficient BCI percepttracking using wavelet transforms, measuring voluntary attention performance, and real-time monitoring of pedestrian traffic based on wireless EEG data, as well as designing a cryptosystem, whose access is only possible through cognitive activity.