Automated balance assessment for blind and non-blind individuals using mini-BESTest and AI

The dataset consists of the following files:

  1. `Test_3`: IMU data for the mBESTest test 3. Inside the folder you will find 58 CSV files with the accelerometer data for each leg of participant that performed the test 3. The data is sampled at 50 Hz and it was acquired using the OpenSense RT IMU sensors (see how to use it here). File naming convention: `ParticipantID_raw_acc_Leg.csv`, where `ParticipantID` is the participant ID, and `Leg` is the leg used for the test. Each CSV file contains the following columns:

     - `ID`: The ID of the participant.
     - `timestamp`: The timestamp of the data starting from 0.
     - `pelvis_ax`, `pelvis_ay`, `pelvis_az`: The accelerometer data for the pelvis sensor.
     - `torso_ax`, `torso_ay`, `torso_az`: The accelerometer data for the torso sensor.
     - `l_shank_ax`, `l_shank_ay`, `l_shank_az`: The accelerometer data for the left shank sensor.
     - `l_foot_ax`, `l_foot_ay`, `l_foot_az`: The accelerometer data for the left foot sensor.
     - `r_shank_ax`, `r_shank_ay`, `r_shank_az`: The accelerometer data for the right shank sensor.
     - `r_foot_ax`, `r_foot_ay`, `r_foot_az`: The accelerometer data for the right foot sensor.
     - `l_forearm_ax`, `l_forearm_ay`, `l_forearm_az`: The accelerometer data for the left forearm sensor.
     - `l_hand_ax`, `l_hand_ay`, `l_hand_az`: The accelerometer data for the left hand sensor.
     - `r_forearm_ax`, `r_forearm_ay`, `r_forearm_az`: The accelerometer data for the right forearm sensor.
     - `r_hand_ax`, `r_hand_ay`, `r_hand_az`: The accelerometer data for the right hand sensor.
     - `neck_ax`, `neck_ay`, `neck_az`: The accelerometer data for the neck sensor.
     - `head_ax`, `head_ay`, `head_az`: The accelerometer data for the head sensor.
     - `label`: The evaluation of the mBESTest test.
  
  2. `participants.csv`: The participants' information. The file contains the following columns:

     - `ID`: The ID of the participant.
     - `Blindness`: The blindness status of the participant (blind or non-blind).
     - `Age`: The age of the participant.
     - `Height (m)`: The height of the participant in meters.
     - `Weight (kg)`: The weight of the participant in kilograms.
     - `Gender`: The sex of the participant.

  3. `mbestest_punctuation.csv`: The mBESTest evaluations for the participants. The evaluations were performed by a physiotherapist and can take values from 0 to 2 (0: bad, 1: mild, 2: good performance). The file contains the following columns:

      - `ID`: The ID of the participant.
      - `Blindness`: The blindness status of the participant (blind or non-blind).
      - The evaluation of each mBESTest test from 1 to 14. For Test 3 and Test 6, the evaluation is performed for each leg. The mBESTest score is the worst of the both legs.
      - The aggregation by category of the mBESTest tests: `Anticipatory`, `Reactive Postural Control`, `Sensorial Orientation`, `Dynamic Gait`.
      - `mBESTest Score`: The total mBESTest score.

  4. `manual_stratification.csv`: A shortcut to the stratification of the participants. A manual statification to ensure that the stratification is correct in the cross-validation of the ML classification. The file contains the following columns:

      - `fold_x`: The ID of the participants in the fold x.