Joint neuromechanics arise from the complex interaction of dynamic, nonlinear elements including muscles, tendons,proprioceptors, and neural circuits. The system has limited non-invasive observability and exhibits time-varying behaviour during functional tasks. This necessitates the use of advanced data analysis and modeling techniques - reductionist and holistic (or system-level), physics-based and data-driven, etc. Therefore, the movement, and neuromechanics communities have developed and validated a variety of software tools to analyze and model their data. However, these tools are fragmented, lack standardization, and are rarely intended for use by the larger community. These limitations slow scientific progress as researchers often must spend time writing the code needed to replicate the analysis of a scientific paper which is always prone to errors, or to piece together fragmented pieces of code written in different languages.

In this project, we will address these problems by developing the NEUromechanics and MOvement ANalysis (NEUMAN-Py) package. We will achieve this through a combination of developing new code, integrating Python packages that contain the required methods, porting code from other languages, customizing code to address specific needs of neuromechanics analysis, developing easy to use graphical user interface (GUI), and developing a neuromechanics data repository that will serve as a stepping stone to standardize maintenance and sharing of human movement and neuromechanics data for testing, validation and exploration purposes.

NEUMAN-Py will comprise the following key modules:

-Time-Invariant (TI) Dynamic Joint Stiffness : Parallel-Cascade, SubSpace, Structural Decomposition SubSpace (SDSS), Short-Segment SDSS methods.
-TI Intrinsic Stiffness : non-parametric Impulse Response Function (IRF), and parametric mass-spring-damper models.
-Reflex Stiffness Parameterization: parameterization of the static nonlinearity and linear dynamics element of the reflex stiffness.
-Endpoint Stiffness: 2D/3D endpoint elasticity ellipsoids estimation and related tools.
TI Intrinsic Compliance : non-parametric (IRF), and parametric (transfer function, state-space, and mass-spring-damper) models.
-Time-Varying Joint Stiffness Identification: ensemble-based and temporal expansion methods,
Parameter Varying Stiffness: Linear Parameter Varying (LPV) of IRF models, LPV Subspace, LPV Laguerre, NPN Hammerstein, NPV Parallel-Cascade models.
-EMG Analysis: EMG filtering, activation calculation, time- and frequency domain EMG analysis, reflex EMG modeling, voluntary EMG modeling, EMG-EMG and EMG-force synchronization
-Muscle Synergies: non-negative matrix factorization (NMF).-
Motion Capture Data Analysis: Parsing various data files from different marker-based motion capture systems; inverse kinematics analysis; inverse dynamics analysis.
-IMU and Gyro Data Analysis: Kalman filtering to estimate joint kinematics (position, velocity, and attitude/orientation/posture) from accelerometer data recorded by IMUs and magnetic field recorded by gyros. Overall, this is the module that requires most research as many of the algorithms are still only available in scientific papers.
-Actigraphy: Analysis of actigraphy data including accelerometer data, body temperature, skin resistance or conductance, and EEG; as well as environmental data such as ambient light and sound levels.
-Data Visualization: Visualization of human movement and joint neuromechanics data as well as the results of modeling and analysis. This includes developing context dependent and physiologically relevant visualization capabilities for NEUMAN.
-Graphical User Interface (GUI): A GUI to facilitate and promote the use of NEUMAN among research communities who use data analysis but are not used to (or mandated to use) coding/scripting and prefer to analyze their experimental data via a graphical interface.