Navegar por los elementos (1 total)
- Resumen es exacto "In recent decades, cognitive neuroscience has shown that cognitive processes rely on dynamic couplings between brain areas over short and long distances. Cognitive events are thus no longer taken to depend solely on a specific region, but rather on the largescale integration of many regions. In human research, most relevant evidence stems from functional magnetic resonance or electromagnetic techniques, marked by limitations in their temporal and spatial resolution, respectively. In this context, invasive intracranial recordings represent a unique opportunity for studying cognition. This technique has a spatial resolution in the order of millimetres and a temporal resolution in the order of milliseconds, alongside one of the best signal-to-noise ratios. Therefore, it is very well suited to evaluating neuroanatomical models and complex temporal dynamics of elemental cognitive domains (e.g., emotion, moral judgement, and language) at the single-subject level. Building on such assets, this thesis relies on intracranial recordings combined with classical evoked activity methods, measures of large-scale brain connectivity, and machine learning approaches to study the coupling of neural networks in key cognitive processes of everyday life. The dissertation comprises three studies. The first one examines neural correlates of the recognition of harmful intentions, a decisive element of moral cognition. Intentional harm induced early amygdalar activity (< 200 ms), with modulations that discriminate among intentional harm, unintentional harm, and neutral actions. Furthermore, connectivity analyses showed a fast coupling with prefrontal regions. These results highlight the key role of the amygdala in the rapid encoding of intentional harm, a critical ability for survival and to moral cognition. The second study assessed emotion recognition in faces and words, two pillars of social interaction. Machine learning algorithms were applied to temporal series data and connectivity coefficients to examine individual signatures of neural activity in subjects with contrastive behavioural performance. The participant’s differential performance profiles were captured by connectivity coefficients. This work represents a contribution to the individual differences approach, a recent neuroscientific trend that focuses on individual neural variability. Lastly, through a combination of intracranial and scalp-level recordings, the third study focused on semantic processing to examine the temporal organization of multimodal systems (areas related to abstract concepts that integrate different sensory pathways) and/or embodied mechanisms (specific sensorimotor networks for each perceptual modality) in the construal of meaning. We studied the dynamics of the modulations of linguistic stimuli denoting to facial body parts (nouns such as ‘nose’ and ‘mouth’) through classical oscillatory analysis together with machine learning, and the interaction between multimodal and embodied systems using connectivity methods. We showed, for the first time, that these stimuli rapidly modulate early markers in facial processing areas (embodied regions), with multimodal contributions, suggesting that linguistic meaning partly relies on fast reactivations of relevant sensory experiences. Together, the three studies provide direct contributions for the construction of models in cognitive neuroscience through invasive recordings in humans. Finally, the methods implemented for these and other works were included in a data analysis toolbox. This tool serves as a repository to ease the organization, reuse, and sharing of scripts, thus supporting student and researcher training. Already equipped with the functions used for present analyses, the toolbox is easily extensible and each laboratory or work group can populate it with their own methods. In sum, building on the uniqueness of intracranial human recordings, the present thesis affords methodological and theoretical contributions in hot topics of cognitive neuroscience, combining novel connectivity and machine learning methods and providing a tool that assists future studies and enhances other potential investigations."
Ordenar por: