This new release incorporates a number of ‘world firsts’.
A particular highlight of V4.4 is the integration of the new Virtual Population (ViP) phantom Yoon-sun V4.0 that features neuro-functionalized nerve trajectories compatible with T-NEURO –, modeled as splines and assigned nerve electrophysiology and neuroanatomy parameters. In combination with the coupled electromagnetic (EM) neuro multi-parameter optimization within Sim4Life’sOptimizer framework and unparalleled high-performance computing/ graphics processing unit (HPC/GPU) performance, V4.4 supports an impressive range of applications from device design to personalized treatment planning and safety assessment, e.g., in the context of electroceuticals.
Sim4Life V4.4 also features the first robust "Maximum Exposure" algorithm for determining the largest power density of antenna arrays of any size on any arbitrary surface via a semi-analytical solver and five iterative optimization methods running in parallel.
In response to specific user requests, we have included enhancements of the IMAnalytics, HPC/GPU and LF solver extensions, together with a whole set of productivity boosts for data analysis and post-processing.
The new release improves user-friendliness and productivity, and includes general improvements and bug fixes, to make your daily work more effective. Feedback, as always, is greatly welcomed – we listen!
YOON-SUN V4.0: FIRST NEURO-FUNCTIONALIZED ANATOMICAL MODEL
The spline trajectories and tissue anatomy of the new ViP model Yoon-sun V4.0 (IT’IS Foundation) are now fully compatible with the Sim4LifeT-NEURO tissue simulation module. The tissue list has been substantially extended and includes separately named and delineated nerves, muscles, and blood vessels.
ViP phantom Yoon-sun V4.0: movie showing the neuro-functionalized nerve trajectories modeled as splines and assigned default nerve physiology parameters based on literature data.
The neural trajectories in Yoon-sun V4.0 have been tagged to facilitate simulation setup. Dragging-and-dropping the model on a NEURO-simulation automatically assigns suitable neuronal dynamics models (sensory and motor fiber dynamics) and the appropriate electrophysiological parameters, as well as the discretization into nodes, internodes, and paranodal compartments. Currently, the parameters are optimized for conservative exposure safety assessment.
EM simulations newly support voxeling of dielectric trajectories. This feature is used to ensure that neural fibers are surrounded by a sufficient amount of nerve tissue to realistically dampen the impact the contrast in dielectric parameters at tissue interfaces. The increased homogeneity of the fiber environment corresponds to anatomical reality and typically increases the effective stimulation threshold.
COUPLED EM-NEURO MULTI-PARAMETER OPTIMIZATION
The Sim4LifeOptimizer can now be applied to coupled physics computations, such as the simulation of EM field-induced heating or neurostimulation, making it possible not only to parameterize the model geometry, simulation parameters, and analysis settings, but also to identify whether the parameterization impacts the underlying first simulation or that repetition of only the dependent simulation is sufficient.
Example applications include optimization of the placement of spinal-cord stimulators and stimulation settings, such as electrode voltages or pulse shape, based on the computed neural responses.
Beyond optimization, support of parameterized modeling of unidirectionally coupled simulations is of value, allowing, e.g., the use of Sim4Life’sSweeper engine to determine recruitment curves or assess the impact of electrophysiological parameters, such as fiber diameter.
Coupled EM-neuro optimization in Sim4Life V4.4: Model of spinal cord stimulation where specific nerves need to be selectively stimulated or inhibited to promote a natural gait. The electrode excitation patterns and power levels are optimized to generate a subject-specific stimulation model.
Neuronal Dynamics Simulation
NEW MODELS AND PROCESSING OF EM-NEURON INTERACTION
New electrophysiological neuron fiber models, targeting in particular electroceutical and neuroprosthetic applications, have been added.
The new C-fiber model, the first unmyelinated fiber model available in Sim4Life, is based on the publication from Sundt and has been validated against results from that publication.
The previously implemented McIntyre, Richardson, and Grill (MRG) model – one of the most frequently employed models of myelinated nerve dynamics – has been extended to distinguish between sensory and motor fibers, in accordance with recent publications.
Already supported is one of the most recent research results – one that attracted a lot of attention – on deep brain stimulation by temporal interference. The formula derived in the above publication to determine the distribution of the envelope modulation amplitude along, which has been speculated to be related to the stimulation magnitude, is now available as a special post-processing/analysis filter.
In Silico Implant Safety Assessment
The Python API of IMAnalytics, easily accessible via the Notebooks provided, has been extended to include more options for normalization, allowing users to perform Tier 3 evaluations using limits derived from the B1 field or its B1+ and B1- components.
The B1 normalization can now be performed based on either the value of B1 at the isocenter of the birdcage or on its mean value across a slice.
Convenient 2D graphing and plotting features in the notebooks have been introduced for easy visualization of exposure distributions across a large range of scenarios.
Unique 5G Simulation Toolbox
NOVEL AND FAST 5G MAXIMUM EXPOSURE ALGORITHM
Novel and exclusive algorithms have been added to the 5G simulation toolbox to make it easier than ever to include quantitative data on power density in regulatory submissions.
The first robust and comprehensive methodology for determination of the maximum power density (norm or total pointing vector) on any surface for large arrays has been developed and implemented. The methodology, comprising a semi-analytical solver and five iterative optimization methods that run in parallel, ensures robust identification of worst-case configurations without overestimation.
Surface-averaged power density on flat surfaces (fSAPD) can now be computed even faster with the fSAPD algorithm. A user-friendly interface allows choosing the exposure plane, the averaging area and the quantity of interest (norm or normal component of Poynting vector).
The new Maximum Exposure Evaluator has already been tested by a user. Watch the movie here.
Solvers and High-Performance Computing
B-FIELD AS EXCITATION FOR INDUCED CURRENT SOLVER MQS
The requirement in previous releases, i.e., knowledge of the vector potential field (A-field) for the magneto quasi-static simulation (MQS) to compute the induced E-field, has been relaxed; it is now sufficient to provide only the magnetic B-field as the excitation field.
The B-field provided by the user is used to reconstruct a vector potential field suitable for driving the magneto quasi-static simulation. The user can then easily check the accuracy of the reconstructed A-field in the post-processor.
The new feature enables straightforward coupling of any external EM solvers to our magneto-quasi-static solver and is fully compatible with all ViP and standard phantom models, thus intensifying the focus on induced currents in human body environment setups.
ENHANCED GPU COMPUTING
Sim4Life V4.4 now supports the NVIDIA Volta generation GPU architecture (Titan V, Quadro GV100, Tesla V100, etc.) which leads to a substantial boost in performance over previous architectures (e.g., doubled speed compared to Kepler).
Also new is that the AXE library now allows for computation jobs to run concurrently, i.e., shared memory machines with multiple GPUs/CPUs installed can run their simulations simultaneously on a subset of cards/processors.
NEW ANALYSIS METHODS AND FEATURE ENHANCEMENTS
A very-long-awaited feature is the integration of the Gamma Comparison Method, which permits visual comparison of distributions and provides a metric of their agreement. This very powerful approach, which has been used by our research partners to compare simulations and measurements for quantitative validation purposes, provides a means for handling variations in distribution patterns.
A new histogram/cumulative histogram filter allows not only visualization of the distribution of values, but also computation of the exact isopercentiles and isovolumes. Potential applications are the determination of (i) certain standardized dosimetric quantities, where the most exposed 1% of the volume can be discarded, (ii) dose quantities used to predict the efficacy of thermal oncology treatments, e.g., the temperature achieved in 90% of the tumor, and (iii) brain exposure quantification, where the volume or percentage of brain volume exposed to field strengths higher than the threshold needs to be quantified, and more.
Selections of multiple post-processing algorithms can now be grouped into a single algorithm, allowing complex workflows to be reduced to just a few customized post-processing blocks with no compromise on the power and versatility of Sim4Life’s pipelining analysis tools.
Further, a selection of post-processing algorithms can be easily saved into analysis pipelines for re-use later in other simulations, in different projects, or by different teams.
A comprehensive list of all new features, improvements, and bug fixes is included in the Release Notes.
To experience the power and elegance of Sim4Life V4.4 or for further information, please email us at firstname.lastname@example.org or call +41 44 245 9765.
The relevant Sim4Life installers for Win7/Win8/8.1/Win10 64-bit platforms can be downloaded here. Existing customers with an up-to-date annual maintenance and support plan, will receive individually all pertinent information regarding the Software Installer Download and updated licenses.
At ZMT we are committed to supporting our customers with the most innovative software solutions, testing equipment, and service.
The Sim4Life Team
EM-neuro stimulation modeling in Sim4Life V4.4: Model of a cuff vagus nerve stimulator (VNS) surrounding the left cervical tract of the vagus nerve in the ViP Yoon-sun V4.0 phantom (including fascicles, perineurium, and different electrophysiological models of myelinated/unmyelinated C-fibers). The close proximity of the carotid artery (red) and vein (blue) affects the performance of the device and the outcome of the treatment. Yellow lines: nerve trajectories in Yoon-sun V4.0.
Risk of potential peripheral nerve stimulation by magnetic resonance imaging (MRI) gradient switching in Sim4Life V4.4: Model of Yoon-sun V4.0 with crossed hands to investigate the effect of different body postures on stimulation thresholds to improve MRI safety regulations and standards (coupled with the NEURON solver to predict the impact of arbitrary gradient waveforms, units, body position, etc. on the neuroelectric response of nerves).
IMAnalytics V2 in Sim4Life V4.4: Computation of the distribution of power density deposited at the tip of a pacemaker across a huge range of MRI exposure conditions taken from the MRIxViP library [link to website]. Results can easily be sliced, diced, and projected onto any dimension thanks to the flexible visualization tools.
Yoon-sun V4.0 with detailed blood vessels (red/blue) and nerve spline trajectories (yellow). The model with all its details has been posed with the Sim4LifePoser tool in Sim4Life V4.4.
Magnetic (H-) field generated in Sim4Life V4.4 by a resonant wireless power transfer (WPT) system for charging an electric vehicle at 85 kHz. The H-field distribution was obtained in silico after analytical reconstruction of the vector potential A from sparse H-field samples.
User-defined analysis pipelines in Sim4Life V4.4: Customized post-processing tasks such as mass-averaged specific absorption rate (SAR) evaluations or other dosimetry analysis are now easy to save and reuse.