A Sensory Neuroprosthesis Enhances Recovery from Treadmill-Induced Stumbles for Individuals with Lower Limb Loss

Dataset of treadmill-induced stumble recovery for three participants with lower limb loss who received a sensory neuroprosthesis that restores plantar somatosensory feedback corresponding to prosthesis foot-foor interactions. 

• Results.zip contains the metrics calculated for analysis of stumble recovery, which includes: trunk angular sway, peak trunk flexion angular velocity, and peak ground reaction force magntidues. 
• Raw.zip contains raw data collected during the experiments.
  • LLX - Data for a particular participant
    • SessionX - Data collected during a single session of the experiment, with each session being one day.
      • DFlow - Contains treadmill speed data. 
      • Visual3D - Contains biomechanical and force plate data. 
        • StaticX - Data collected during a static trial where the participant is standing in a T-pose. This data was used to calculate body weight during each session. Data on which blocks correspond to which static trials can be found in LLX_ProcessData.m This has the same directory structure as TrialX. 
        • TrialX - Data collected during one block of walking with multiple perturbations. Data on which blocks had the SNP active or inactive can be found in the FILE_DATA variable in LLX_Analysis.m files in Code.zip.
          • Analog - Contains force plate data collected from the instrumented treadmill, which was used to calculated ground reaction force metrics. 
          • ForcePlate - Contains force plate data calculated from the instrumented treadmill and interpolated to the Marker data.
          • LinkModel - Contains data from the biomechanical model, including trunk angle. 
          • Markers - Contains marker data from motion capture. 
• Code.zip contains the code used to process the raw data in Raw.zip and extract the metrics found in Results.zip. To run this, follow the steps below.
  
  1. Copy files from Projects/TreadmillPerturbations/Run/ to directory where processed data will be stored. 
  2. In this same folder from (1), create the following directories:
     • Processed/LL1/Session1
     • Processed/LL1/Session2
     • Processed/LL1/Session3
     • Processed/LL1/Session4
     • Processed/LL1/Session5
     • Processed/LL2/Session1
     • Processed/LL2/Session2
     • Processed/LL2/Session3
     • Processed/LL2/Session4
     • Processed/LL2/Session5
     • Processed/LL3/Session1
     • Processed/LL3/Session2
     • Processed/LL3/Session3
     • Processed/LL3/Session4
     • Processed/LL3/Session5
  3. In LLX_ProcessData.m, edit raw_data_all variable to path where raw data is stored (where data from Raw.zip has been extracted to).
  4. Run LLX_ProcessData.m. Processed data will appear as .mat files in directories created in (2).
  5. Run LLX_Analysis.m. This will reproduce the results found in Results.zip.