A computational outlook on neurostimulation

Capogrosso and Lempka Bioelectronic Medicine https://doi.org/10.1186/s42234-020-00047-3 (2020) 6:10 Bioelectronic Medicine PERSPECTIVE Open Access A computational outlook on neurostimulation Marco Capogrosso1,2*† and Scott F. Lempka3,4,5† Abstract Efficient identification of effective neurostimulation strategies is critical due to the growing number of clinical applications and the increasing complexity of the corresponding technology. In consequence, investigators are encouraged to accelerate translational research of neurostimulation technologies and move quickly to clinical applications. However, this process is hampered by rigorous, but necessary, regulations and lack of a mechanistic understanding of the interactions between electric fields and neural circuits. Here we discuss how computational models have influenced the field of neurostimulation for pain and movement recovery, deep brain stimulation, and even device regulations. Finally, we propose our vision on how computational models will be key to accelerate clinical developments through mechanistic understanding. Keywords: Neurostimulation, Neuromodulation, Computational modelling, Finite element modelling, Spinal cord stimulation, Chronic pain, Spinal cord injury Background In this perspective article, we sought to provide our personal experience and thoughts on the impact of computational models in the field of neurostimulation. We describe the general framework of technology development in the neurostimulation industry and provide examples of past, present, and potential future utility of computational models in accelerating technology development. We believe that our interpretation of the recent advancements in the field could help motivate other investigators to invest in the use of computational models, hopefully leading to a more precise interpretation of pre-clinical and clinical results. Main text In the age of fast information transfer and social media, we are getting used to direct access to information and * Correspondence: † Marco Capogrosso and Scott F. Lempka contributed equally to this work. 1 Department of Neurological Surgery, University of Pittsburgh, Pittsburgh, PA, USA 2 Rehabilitation Neural Engineering Laboratories, University of Pittsburgh, Pittsburgh, PA, USA Full list of author information is available at the end of the article technology, on demand. We are convinced that this new interconnected environment allows inspiring ideas to quickly spin-off into high-profit, fast-success, high-tech solutions to environmental, social, and healthcare challenges of modern societies. However, this is not quite the pace of scientific discoveries. Scientific advances occur through a slow, laborious, and rigorous process requiring multiple experimental verifications and crossvalidation procedures. This process is particularly true for biomedical applications in which significant costs and strict, but necessary, regulatory constraints bind the technology development to an even slower pace. Despite this fact, the scientific community, and in particular the neuroscience community, is too quickly focusing on “translational applications” (i.e. the translation of scientific discoveries in neuroscience to clinical settings). Fostered by the urge to solve the impelling needs of an aging society, funding bodies provide everincreasing support to this type of research. Given the stakes, as members of the scientific community and information-era human beings, we should question the very concept of translation, and approach this task with the most rigorous scientific attitude (Arber and Arber 2016). © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Capogrosso and Lempka Bioelectronic Medicine (2020) 6:10 Neuromodulation, or neurostimulation, technologies offer a clear example of this frantic race to clinical implementation. Both the scientific and industry communities seek new tools to interact with the nervous system and its computational architecture, without having a clear understanding of its particular features. Therefore, when developing neuromodulation technologies, engineers are asked to design devices that interact with largely undetermined systems, sometimes without having identified the actual neural targets. Verification of such systems is then sought in the preliminary outcomes of exploratory pilot clinical studies. However, given the titanic efforts and costs of clinical research, this purely experimental evidence-based approach is sub-optimal. Therapy optimization would be more efficient if at least part of the system efficacy was verified prior to finalization of the design. This initial verification would help focus design efforts on specific features, thus reducing the number and risks of experimental trials needed to refine therapies. Computational models are natural candidates to perform this initial testing. The synthesis of state-of-the-art neuroscientific concepts into in-silico models of the nervous system simultaneously serves two purposes. First, it highlights how much we know of a specific system and where we should direct experimental research to acquire new knowledge (Markram et al. 2015). Second, it provides a virtual testing platform to study the interactions between neuromodulation technologies and the computational structure of the nervous system (McIntyre and Foutz 2013). After all, the efficacy of neuromodulation is determined by our ability to modify the outcome of the mathematical operations performed by complex networks of neurons. We can simulate these operations by implementing artificial representations of networks and their interactions with neuromodulation technologies. We and others in the field have applied this strategy to characterize the interactions between spinal cord stimulation (SCS) and the dynamics of spinal circuits for the design of neuromodulation protocols to reduce chronic pain and to improve motor control in people with spinal cord injury. We believe that our personal experience in the use of computational models might provide a helpful example of the role that models could have in addressing important clinical and scientific que (...truncated)