DIRECT ELECTRICAL STIMULATION OF PRIMARY MOTOR AND FRONTAL PREMOTOR REGIONS: MAPPING AND PRESERVING NETWORKS FOR HAND MOTOR CONTROL DURING BRAIN TUMOUR RESECTION

This PhD project was funded by EDEN2020 (Enhanced Delivery Ecosystem for Neurosurgery in 2020). Brain disease represent a cost about 800 billion euros per year in Europe and outcome of treatment is demonstrated to critically depend on the knowledge of functional anatomy and its preservation. By combining pre-operative MRI and diffusion-MRI imaging, intra-operative ultrasounds, robotic assisted catheter steering, brain diffusion modelling and a robotics assisted neurosurgical robot (the Neuromate), EDEN2020 aims at realizing an integrated technology platform for minimally invasive neurosurgery which will provide a significative step change in treatment of brain disease. To date, neurosurgical instruments for diagnostic and therapy (drugs infusion) are inserted via rigid cannulas. This represents a primary technological limitation of treatment with direct consequences in patient’s post-operative outcome, since the insertion of rigid cannulas cannot be planned along procedure-optimised trajectories which take into account tissue microstructures and respect the bundles’ topographical anatomo-functional organisation. To bridge this gap, main aim of EDEN2020 is to engineer a steerable catheter for chronic neuro-oncological disease than can be robotically guided and kept in situ for extended period, which insertion can be tailored on clinical conditions and individual anatomy. The correct trajectory and final positioning of a catheter must be planned and guided through the brain structures by the knowledge of the anatomo-functional organization of the neural circuits subserving the essential motor and cognitive functions to avoid lesions resulting in permanent deficits impacting on the quality of life of patients. In addition, since the diffusivity is enhanced when it follows the white matter pathways belongings to the network in which an individual tumour has grown, as showed by data generated by EDEN consortium, these circuits became the target of the drug. In EDEN animal trials (ovine model), the circuit targeted for delivery was the corticospinal tract, due to the anatomical restriction imposed by the sheep brain. In humans, this descending system, which is essential for everyday life activities allowing the skilled use of the hand (i.e. the ability to manipulate objects and tools), has a much higher level of complexity and its functional organisation has not yet been described in detail as in other animal models. The complexity of the neural organization underlying motor control of hand gestures in humans results in a dramatic degree of freedom, but at the same time in a poor ability to recover after lesions. When this connectivity is infiltrated by a tumour and thus became the possible target of drug delivery devices (EDEN), its complexity must be taken into consideration to avoid the onset of deficits. The great majority of brain tumours occurs in the frontal lobe and, particularly Low Grade Gliomas (LGGs), develop close or within the cortical but mostly subcortical structures involved in motor control. Therefore, to track safely the entrance and the trajectory of catheters, a reference atlas of the neural circuitry controlling hand movement is mandatory to identify which cortical and subcortical areas must not be lesioned to avoid permanent inability. Based on this premises, this PhD project investigated, with a multidisciplinary approach, the frontal networks subserving hand function to provide a frame for understanding the connectivity involved in hand skilled movements, which became a possible target for drug delivery in tumours developing in primary motor and/or pre-motor regions. The skilled use of the hand is allowed by the high level of human sensorimotor control implemented by the corticospinal system, particularly developed in primates, connecting distant and functionally different areas via subcortical bundles and finally acting on the spinal cord with a huge bundle of descending fibres. This complex network computes the sensory information related to the goal of the action to shape the appropriate motor command for motoneurons, the final common path to muscles. Non-human primate studies have demonstrated that the main motor output of the corticospinal tract is the primary motor cortex (M1), which act on the spinal motoneurons, in producing voluntary hand and finger movements. The monkey M1 has recently been demonstrated to represent an anatomo-functional non-unitary sector, subdivided in a caudal region dense with cortico-motoneuronal cells and a rostral sector with few monosynaptic connections to alpha-motoneurons and/or with slower projections to the spinal cord. To control and execute skilled hand movements, M1 is highly interconnected with other frontal and parietal cortical pre-motor regions, with subcortical structures such as the basal ganglia and with the cerebellum. A precise description of the human circuitry allowing for realization of dextrous hand movement is still missing in the human, as the electrophysiological and anatomical experimental approaches developed in animal models cannot be performed (i.e. intracortical microstimulation, neuronal tracing, lesion studies etc.). The unique setting of brain tumour resection with the brain mapping technique gives a great opportunity to use clinical data to evaluate neural networks in humans. In this setting, during surgical resection, Direct Electrical Stimulation (DES) is applied onto the exposed cortical and subcortical areas in order to identify the eloquent sites, i.e. where DES elicits motor responses, thus individuating the structures directly acting on the motor descending pathways, or induces transient impairment of the execution of a task, due to its interference with the physiological activity of the stimulated area. This approach allows for the extension of the resection of the tumour beyond its boundaries, increasing the patients’ survival while preserving their functional integrity. As has emerged by recent publications of our group, among the different stimulation paradigms available for intraoperative monitoring, the high frequency stimulation (‘the pulse technique’), which elicits motor evoked potentials (MEPs), is the most reliable paradigm for mapping the descending fibres originating form primary and non-primary motor areas, also in lesions infiltrating M1, while long and short-range fronto-parietal premotor pathways are well identified when low frequency stimulation (‘the Penfield technique’) is applied while the patient is performing a dedicated object manipulation task, clearly interfering with its performance. With a multidisciplinary approach, by combining electrophysiological data with virtual anatomical dissections by means of high angular resolution diffusion imaging (HARDI) tractography we correlated the functional properties of the stimulated sites with specific anatomical structures. In this PhD project, we focused on: the anatomo-functional properties of the human hand representation in M1 (study 1); the oncological and functional efficiency of high-frequency mapping in tumours harbouring within M1 (study 2); the frontal premotor pathways involved in controlling fine hand movements (study 3). Study 1, conducted on 17 patients who underwent an awake procedure, reported a possible subdivision, based on anatomo-functional analysis, of the human hand-knob in two sectors (a posterior one, close to the central sulcus, and an anterior one, close to the precentral sulcus) with different cortical excitability, different hand-muscle electromyographic (EMG) pattern when stimulations were delivered during the object manipulation task and, finally, with different local cortico-cortical connectivity. Overall data suggests that the two sectors may exert different roles in motor control. Study 2 consisted of a retrospective analysis of 102 patients who underwent an asleep procedure for the removal of tumours harbouring with M1 and its descending fibres. The neurophysiological protocols adopted for the intraoperative brain mapping were correlated with the clinical condition, the tumour imaging features, the extent of the resection and the post-operative functional outcome. First, results indicated that M1 tumour removal is feasible and safe and the high frequency stimulation was revealed as the most efficient and versatile paradigm in guiding resection of M1, affording 85.3% complete resection and only 2% permanent morbidity. The study confirmed the possible subdivision of M1 in a rostral less excitable region and a caudal more excitable region reported in Study1 with its clinical impact: the rostral sector can be indeed considered a safe point of entry for surgery and thus for catheters. Study 3 aimed at characterizing the effect of DES on the electrical activity (EMG) of hand movers during a dedicated object-manipulation task during subcortical stimulation of the frontal white matter anterior to M1 (precentral gyrus) and the anatomical evaluation of the stimulated sites by means of diffusion tractography, in 36 patients who underwent an awake surgery. Results indicated that stimulations of dorsal premotor connections with the spinal cord, dorsal striatum, local U-shaped connections and the superior longitudinal fasciculus I and II resulted in abrupt arrest of the hand, while more ventral stimulation, mainly targeting the third branch of the superior longitudinal fasciculus (SLF III) resulted in clumsy hand movements. Resection cavities analysis showed that transient post-operative upper-limb motor deficit occurred only disconnecting the supplementary motor area corticofugal fibres and the frontal U-shaped connections. Overall data suggests that DES on dorsal premotor white matter could interfere with areas involved in the very final stages of the motor program, while DES on ventral premotor white matter could halt the sensorimotor transformations necessary for correct hand shaping.