Signatures of monitoring in speech and their precursors : a multi-facet outlook.

Speech is a multifaceted process that rapidly and precisely combines cognitive processing with motor actions. Despite this complexity, speech errors occur infrequently. This remarkable precision can be attributed to the presence of self-monitoring, allowing to detect and promptly rectify various types of overt or impending errors. The primary focus of this thesis is to comprehensively investigate the phenomenon of speech monitoring, seeking to address how it operates within the complex framework of speech production and beyond. The research aims to explore distinct monitoring mechanisms, identify key moments in the implementation of self-monitoring in speech, examine how different levels of representation influence the monitoring process, and contextualize self-monitoring within the broader spectrum of human capacities along the motor-cognitive continuum. In the first chapter, an event-related fMRI study assessed the roles of different brain structures associated with theoretical frameworks for self-monitoring operations. The findings indicated a role of temporal cortex in detection of overt errors through comprehension-based monitoring; a role of medial frontal cortex in feedback-based control and a role of the cerebellum in the internal modeling of upcoming speech. A subsequent chapter using EEG delved into the temporal aspects of the monitoring process, investigating different sources of error probability (lexical and articulatory-phonetic)as well as overt speech errors. The results indicated that monitoring processes come into play at specific stages of speech production, concentrating on various types of information relevant at each stage. In the next chapter, an fMRI study explored action general monitoring by examining a simple motor task that was complexified by the introduction of a cognitive monitoring rule. The focus was on the shift in brain structures associated with the increase of cognitive demands of the task. The results were consistent with an adaptive and hierarchical fronto-parieto-cerebellar control network. Finally, the fourth study investigated the relationship between a morphological specificity of the brain and functional activations in the anterior cingulate cortex,(ACC), a key area of interest in monitoring research. This study revealed that activation patterns elicited by speech errors depend on sulcal variations in the ACC. In summary, this multifaceted exploration of speech monitoring contributes to our knowledge about its multiple mechanisms, temporal dynamics, adaptability to levels of representation and its place along the continuum of motor-cognitive functions. By employing various neuroimaging techniques and diverse tasks, it sheds light on the complex organization of brain structures involved in monitoring and its flexible nature.