THE NEURAL NETWORK UNDERLYING SPEECH IN HUMANS: INTRAOPERATIVE INVESTIGATION OF MOTOR CONTROL OF SPEECH IN BROCA, VENTRAL PRE-MOTOR AND PRIMARY MOTOR CORTICES

The language, as form of verbal communication –and of thought-, is a unique human ability. It is complex form of communication, consisting of different levels of representation (phonological, syntactic and semantic), translated into words by a sensory-motor system controlling the phono-articulatory apparatus. The loss of language ability, as a result of injury, is an intolerable disability with, at present, very little chance of functional recovery. In order to design efficient rehabilitation actions, the knowledge of the functional organization of neural circuits that underlie language is mandatory. The aim of PhD project was to investigate the functional properties of the cortical areas involved in sensory-motor control of speech by acting on the phono-articulatory apparatus. The phono-articulatory apparatus is innervated by the motor nuclei of the cranial nerves receiving, with few exceptions, bilateral input from the Primary Motor cortex (M1). At present, however, it remains unclear which other cortical areas in the frontal lobe are actually involved in shaping the motor program to be executed by M1 to allow verbal production and which is their precise functional role in sensory-motor control of the phono-articulatory gestures. Main focus of the project were the three frontal area classically considered involved in motor control of speech: Broca's area, vPM and M1. A putative direct role of Broca’s area in motor control of speech must be exerted either in shaping the activity of M1 or through its independent control of bulbar motoneurons. In both cases, Broca’s area is expected to significantly affect, either directly or indirectly via M1, the motoneuronal excitability and, in turn, the activity of phono-articulatory muscles. However, the observation that injuries of sole Broca’s area do not result in a motor deficit of speech, but rather in improving mutism (see Introduction), raised doubts on its significant role in control of speech motor output. Interestingly, the “apraxia of speech”, a clear phono-articulatory dysfunction (see Introduction), follows to lesions of the ventral Pre-Motor cortex (vPM), suggesting to be involved in motor programming of speech. However, all these issues remained unresolved for many years, due to the lack of appropriate experimental tools to study these areas in humans, in ecological conditions. In the last two decades, the introduction of the intraoperative brain mapping technique allowed a direct investigation of the functional properties of M1, vPM and Broca’s area, adding new elements to the debate regarding the role of these areas in speech. However, the absence of a careful analysis of phono-articulatory activity during the brain mapping, highlights confounding evidence regarding the actual role of three areas in motor control of speech. This study was performed by analysing data collected intraoperatively in 70 patients during surgical removal of gliomas performed with the aid of the brain mapping technique by a high skilled neurosurgical team. The instrumental setup and the methodology used to perform the brain mapping technique gives the unique opportunity to investigate human circuits underlying language with a direct approach (see Introduction). During resection of gliomas located within or in proximity to the cortical areas and tracts involved in language neural network (at level of frontal, temporal or insula lobes), Broca’s area, vPM and M1 are exposed and electrically stimulated as essential part of the clinical procedure. During the intraoperative phase, the patient is awakened and asked to perform different types of language tests assisted and evaluated online by neuropsychologists. During the language tests (object picture naming and counting tests), the Direct Electrical Stimulation (DES) is applied on the exposed areas (M1, vPM and Broca’s area), in order to identify the eloquent cortical sites, i.e. the sites where DES actually “interferes” with the language function inducing a deficit in performance. To perform a reliable mapping procedure, at the beginning of surgery the precise site onto the three areas expected to control the phono-articulatory apparatus and therefore to be likely involved in speech must be identified. It is indeed known in literature that both M1 and vPM host the representation of both oro-facial and hand-arm muscles (M1 controls also foot-leg muscles). Thus, during procedure first it is mandatory to identify, during each area the hand-arm and the oro-facial, the latter to be then exposed to stimulation during speech tasks. Once the oro-facial representation in the areas was disclosed, DES was applied on the areas during language tasks and the performance was compared with the performance of the same task without DES (natural performance). During all the surgical procedure, the electrical activity (EMG) of some of the muscles involved in the phono-articulation has been recorded. The EMG signal was analysed offline with a quantitative approach allowing to investigate the pattern of motor units voluntarily recruited during the language tasks performed by the patient in absence of stimulation (natural performance) to be then compared to the pattern of recruitment recorded when DES was applied on the three different cortical areas in the same conditions. This analysis was designed to disclose the specific pattern of specific alteration in motor unit recruitment in phono-articulatory muscles due to the DES-induced “transient inactivation” of the three areas. These data, interpreted in light of the animal studies and human studies, were used to infer the putative specific role of M1, vPM and Broca’s area in motor control of phono-articulatory muscles in the attempt to shed light on the cortical network underlying the executive branch of language network.