Data from "Sensing the Full Dynamics of the Human Hand with a Neural Interface and Deep Learning"

Format
This dataset contains the ground truths and predictions of all deep learning models from Sîmpetru et al (_{https://doi.org/10.1101/2022.07.29.502064}).
Fig. 2-4
The data is formatted as follows:

The data used for <em>Fig. 2</em> is stored under <strong>3D_points_{random, testset}</strong> depending on whether the AI was tested with the random or the test set.
The data used for <em>Fig. 3</em> is stored under <strong>angle_{random, test set}</strong> depending on whether the AI was tested with the random or the test set.
The data used for <em>Fig. 4</em> is stored under <strong>force</strong>.
Each of these folders contains the predicted and expected data for each of the 13 subjects stored as <strong>Subject{1, ..., 13}_{predicted, expected}.pkl</strong>.

Fig. 5
The data for Fig. 5 is stored under <strong>comparison</strong>.  
<br>
To distinguish between the EMG forms used, we use:

5 Hz low-pass filtered as <strong>3D_points_5</strong>
5 Hz low-pass filtered and rectified as <strong>3D_points_5_rectified</strong>
20 Hz low-pass filtered as <strong>3D_points_20</strong>
20 Hz low-pass filtered and rectified as <strong>3D_points_20_rectified</strong> 
raw as <strong>3D_points_raw</strong>
raw and rectified as <strong>3D_points_raw_rectified</strong>
raw and 20 Hz low-pass filtered combined as <strong>3D_points_raw20</strong>

<br>
Each of these folders contains the predicted and expected data for each of the 4 subjects, stored as <strong>Subject{1, ..., 4}_{predicted, expected}.pkl</strong>.<br>

General file format
The saved <strong>.pkl</strong> files are tables where the columns represent the joint labels (<em>Fig. 1B</em>) and the rows represent the bin number.
<br>
Each cell of the table contains a series of 3 values representing the x, y and z coordinates of the respective joint label at a given time.<br>
<br>

Loading
The stored files are <em>pandas</em> (high-level python library for tables manipulation) dataframes stored using <em>pickle</em> (python object serialisation library).<br>
<br>
To read such a file, we give the following minimal example:<br>
<br>
import pandas as pd<br>
<br>
dataframe = pd.read_pickle("PATH_TO_FILE.pkl")<br>
<br>
# display head (top 5 rows) of the dataframe<br>
print(dataframe.head())<br>