TIP Manual

Appearance

TIP V5.2 Latest ​

SuMo Optimizer — Pareto Front Quality

  • Improved SuMo optimizer Pareto front quality via three complementary enhancements to Step 6 (Post-Processing)
  • Replaced penalty-based handling of invalid electrode configurations with native Dakota nonlinear inequality constraints, producing a smoother surrogate fit and eliminating invalid solutions from the Pareto front
  • Added parallel multi-seed execution (default: 6 seeds) with Pareto front merging via non-dominated sorting, yielding a denser and more representative approximation of the true Pareto front without increasing wall-clock time
  • Introduced adaptive convergence stopping based on the hypervolume indicator (HV): the optimizer halts each seed once HV improvement falls below a threshold, avoiding wasted iterations

New User Controls

  • Added user-selectable convergence precision (Low / Medium / High), each backed by optimized evolutionary algorithm parameters, allowing users to trade off speed against Pareto front completeness

TIP V5.0 5.0.0 ​

Personalization
  • Introduced offline personalization workflow (run_personalizer.bat) for local MRI anonymization and segmentation using Sim4Life 9.4+
  • Raw MRI never leaves the user's machine; only the anonymized, segmented head model is uploaded to TIP
Exposure Analysis
  • Upgraded Exposure Analysis from a read-only analysis viewer to a full Sim4Life workbench (Modeling + Simulation + Analysis)
  • Added ready-to-use Sim4Life template projects for all precomputed models (MIDA isotropic/anisotropic, IXI Female, IXI Male, Mouse)
Performance
  • Reduced SuMo optimizer runtime from ~1 hour to 5–10 minutes via lead-field array caching and Numba-accelerated field superposition

TIP V4.0 4.0.0 ​

Personalization
  • Implemented automated personalization with neural network-based fiducial point detection
Optimizer
  • Introduced SuMo optimizer replacing the previous sweeper
  • Streamlined user workflow by eliminating electrode selector and sweep steps
Performance & Visualization
  • Enhanced parallel processing for up to 50% faster optimization runs
  • Improved visualization for Pareto-optimal solutions

TIP V3.0 3.0.0 ​

Personalization & Target Selection
  • Introduced personalized TI stimulation planning: users can upload subject-specific T1-weighted MRI (and optionally DTI) data for personalized exposure assessment and optimization
  • Added ICBM brain atlas registration for an expanded number of target regions available for stimulation optimization
Infrastructure
  • Migrated platform to Amazon Web Services (AWS) for increased reliability, stability, and scalability
  • Parallelized low-frequency EM simulations across individual electrodes using AWS resources, accelerating the overall workflow
New Features
  • Optimized for the TIBS-R multichannel device
  • Launched TIP.lite: a reduced-functionality tier giving broader audiences access to Classic TI mode and selected precomputed head models

TIP V2.2 2.2.0 ​

Standards & Imaging
  • Adopted the RAS (Right, Anterior, Superior) coordinate convention for image and exposure viewing, aligning with established scientific standards
  • Integrated co-registered brain atlases to expand the number of supported target regions for stimulation planning
  • Registered MRI data to field distributions for improved visualization and methodological rigor
Infrastructure
  • Expanded internal cluster capacity to increase availability and support multiple simultaneous users

TIP V2.0 2.0.0 ​

Head Models
  • Expanded head model library from one to four highly detailed computational anatomical models, enabling investigation of inter-subject variability
  • Added a mouse head model to support optimization of TI protocols for rodent research
New Stimulation Modes
  • Introduced Multichannel TI mode with up to eight independent channels, improving stimulation focality, selectivity, and multifocal targeting while distributing current over larger scalp areas
  • Introduced Phase Modulation TI mode, enabling advanced pulse shapes, burst stimulation, and complex modulation schemes for activity-synchronized and state-dependent stimulation
UX & Performance
  • Added visual guidance alongside the quantitative results table to help users interactively balance conflicting optimization goals (e.g., target exposure vs. collateral exposure)
  • Leveraged cloud computing and parallelization to enable interactive exploration of large electrode configuration spaces

TIP V1.0 1.0.0 ​

Initial Release
  • First release of TIP: an online-accessible, cloud-based TI stimulation planning tool built on o2S2PARC technologies
  • Optimization based on 71 precomputed 3D electric field distributions for the standard 10-10 International EEG Electrode System, computed using the MIDA head model with the Sim4Life low-frequency electro-quasistatic solver
  • Pareto-based multi-goal optimization using field superposition for efficient search over electrode pair permutations
  • Computed three key performance metrics per electrode configuration: target exposure strength, stimulation selectivity, and collateral exposure
Analysis & Reporting
  • Visually guided interactive refinement of the stimulation strategy with high-frequency and TI modulation field distributions overlaid on anatomical image data
  • 3D isosurface visualization of the TI exposure distribution within head anatomy to assess collateral stimulation
  • Automatic report generation covering electrode placement, current intensities, performance metrics, field distributions, and cumulative dose histograms
  • Field export in multiple formats (e.g., MATLAB) and integration with Sim4Life web for in-depth post-processing