• functional imaging
• source localization
• computational neuroscience
• bioelectronic medicine
• vagus nerve stimulation

Background and Objective: Bioelectronic medicine is an emerging field aiming to develop therapies for the treatment of various chronic diseases through targeted organ neuromodulation. One of the most interesting possibilities for administering bioelectronic medicine treatments is through selective stimulation of the autonomic nervous system (ANS), since it plays a crucial role in the control of the whole-body homeostasis. In particular, the vagus nerve stimulation (VNS) is a promising therapy for treatment of various conditions that are resistant to standard medication. However, the fascicular organization of the ANS is surprisingly not yet well understood, with consequent intersubjective variability of the efficacy of the treatments, sometimes with the generation of side effects, due to an unselective stimulation that recruits non-target fibers. Here, an advanced method for the functional characterization of the left vagus nerve was developed, i.e., the identification of the nerve regions involved in cardiovascular or respiratory functions, using a computational model.
Methods/Approach: A simulation framework to identify the regions involved in a certain function were created by combining Finite Element Method (FEM) analysis of the macrostructure of the nerve (simulating the distribution of the electrical potential at a given stimulation) and the behaviour of neurons, simulated through the NEURON environment (https://neuron.yale.edu/neuron/). The method used to define the functional topography was to spatially locate the sources associated with a certain function, through the extraction of the discriminative indexes from the recordings. A set of realistic detailed models, designed based on the morphological parameters of the human cervical vagus nerve, were used to simulate realistic neural recordings. Through a model that exploits prior knowledge, i.e., only the geometry of the electrode and the macro structure of the nerve (size and shape), without any information about fascicles distribution, several tests were done to assess the performances of three employed source localization algorithms. In detail, we compare the results of two algorithms present in the literature (Beamforming and Discriminative Field Potential) with a new method presented in this study that we called Discriminative Beamforming (DBF).
Main results: The recordings were simulated both with intraneural electrodes (TIME) and with extraneural electrodes (cuff), working with different levels of current sources’ complexity and with different noise level, characterizing the general performance of the algorithms for different electrode types and in different conditions. The DBF results the best performing algorithm and has therefore been adopted to create functional maps for estimating the nerve regions associated with a function.
Conclusions/Significance: Achieving an accurate localization in a real case scenario will lead to a functional characterization of the vagus nerve: this in a first place will aims to improve the accuracy of the decoding and subsequently will improve the effectiveness of preclinical and clinical treatments, through more targeted stimulation.