The investigation of the level of processing responsible for the transformation of an abstract linguistic input into readable motor commands — also known as motor speech encoding — is a rapidly growing domain from the last 25 years. In an effort to understand, characterize and contribute to fundamental and clinical research, the present thesis aims at providing empirical evidence from Electroencephalography (EEG) recordings during several experiments using the delayed production task. Especially, various experimental constraints have been used including speech modes of different kinds, the articulatory complexity of the syllables and the type of speakers. First of all, the investigation in Chapter 2 of whispered speech and loud speech in two distinct experiments yielded very thrilling results. On one hand, behavioral analysis on utterances produced louder than usual demonstrated that participants needed more time to produce those utterances. This additional initialization time was associated to changes in EEG activity in a large time window containing the same brain networks as those supporting the encoding of normally produced speech. On the other hand, it has been observed that although producing whispered utterances entailed increased latency of initialization, this behavioral encoding cost was not observable in electrophysiological analyses with almost identical results across conditions. Furthermore, results from Chapter 3 demonstrated that faking an english accent and producing loud utterances induced an EEG encoding signature in the last 200 ms preceding the vocal onset. However, despite covering the same time period, the spatial and temporal modulations of the neuronal activity induced by each speech mode were unique indicating no singular speech mode encoding mechanism. Eventually, analyses from Chapter 4 highlighted that the same encoding signature for loud speech — i.e., increase in brain activity from a brain network represented by a prototypical topography — was observable with the same properties in each chapter. This result was even more intriguing when put into perspective with single trials analyses on patients suffering from specific subtypes of motor speech disorders. Gathered together, evidence across chapters will provide important insights to the following questions arising from the literature:

How does the brain encode speech modes?

How can speech modes be integrated in theoretical models?

What is the time course of motor speech encoding?

How many processes are underlying the motor speech encoding level of processing?