Neuronal mechanisms of perception and cognition
- Autores: Parth Chholak
- Directores de la Tesis: Alexander Pisarchik (dir. tes.)
- Lectura: En la Universidad Politécnica de Madrid ( España ) en 2021
- Idioma: español
- Materias:
- Enlaces
- Tesis en acceso abierto en: Archivo Digital UPM
- Resumen
This work examines neuronal substrates functioning as complex networks in various prevailing neuroscientific paradigms, including motor imagery (MI), visual flickering, picture naming, semantic anomaly, auditory oddball, and motion-onset visual evoked potentials.
We present the corresponding experimental designs and implementation details of recording electrophysiological data such as magnetoencephalography (MEG) and electroencephalography (EEG) alongside the performance of experimental tasks by the subjects. We describe popular data processing tools as well as our own novel methods to transform the recorded electrophysiological signals into useful measures and features.
Models of underlying mechanisms are discussed, proposed, and validated against state-of-the-art and our own experiments. Key models/mechanisms include neuronal communication during MI, generalised perception model, coherence resonance assisted visual stimulus detection, and neuronal network in object recognition and naming.
Novel brain-computer interface (BCI) applications and refinements to current systems are put forth. Key propositions include appropriate filtration for artificial neural network classification in MI-based BCI, a model-free brain noise estimator, efficient percept-tracking in BCIs using wavelet transforms, measuring voluntary attention performance, live monitoring of pedestrian traffic based on wireless EEG data, and a cryptosystem only accessible through cognitive activity.
