Begin by running the commands to install and configure an uncomplicated firewall in the PC terminal. For the server VPN installation, execute the installation of the open-source WireGuard VPN protocol on the computer. Navigate actively to the installation directory and run the unmask 007 command in the PC terminal to update the directory access rules appropriately.
Then run the command to generate a public or private key pair for the VPN server in the PC terminal, which can be shared with any client PC that connects to the VPN. Next, run the commands to create a server configuration file and paste the server rules into this file. Enter the command to activate the VPN.
Also execute the command to program the VPN protocol to launch each time the PC reboots automatically. Take the public key from the configuration file, generated during installation and paste it into the peer section of the server's configuration file. Set up a new client tunnel, copy the client public key and paste it into the server configuration file.
Now copy the software link from GitHub and install the software on the computer to schedule and control video recordings. Choose a monitor that will show the video recording and process and allows users to terminate recordings seamlessly. Select USB-compatible webcams that offer a high resolution to accurately calculate poses within the designated space.
Identify a space for webcam installation that offers minimal disruptions to the home environment. Then mount the webcams in this designated area, using the selected mounting hardware. After discussing a suitable recording schedule with the patient, set this schedule in the video-recording software.
Confirm that the PC is equipped with the latest version of the chosen video-recording software, using the GitHub machine user account installed earlier. Next, log into the PC, using the installed remote desktop software and activate the video-recording software. Acquire a checkerboard pattern to support the 3D calibration of the pose-estimation software.
Capture a video from each webcam, ensuring the checkerboard is angled appropriately within the frame for all cameras. To initiate data collection, request the patient to inspect the battery and power supply of the INS device to ensure it remains on for continuous stimulation. Request the patient to activate the tablet PC and verify that the CTM for the left and right INS devices are switched on and fully charged.
Prior to any data collection, the patient has two INS devices implanted in both sides of their chest. Instruct the patient to position the CTMs on both sides of their chest. Upon the tablet's activation, request the patient to launch the Deep-Brain Stimulation or DBS application and select the connect option to initiate a Bluetooth connection with the CTMs and ultimately the INS devices.
Ask the patient to turn on their smartwatches and smartphones by pressing the power button. Further, direct them to open the smartwatch application to initiate data recording and Parkinson's disease symptom tracking. To ensure the reliability of timestamp alignment and multi-device clock-based synchronization, instruct the patient to perform a specific gesture at the start of each new recording session.
Switch the stimulation group to the patient's preferred clinically assigned team. Submit a symptom report on the patient-facing graphical user interface of the DBS application. Then close the DBS application to disconnect the clinical trial monitors and stop the INS streaming.
In the manual time, align the graphical user interface and overlay data streams for visualization. Zoom in on the time axis and move the viewing window to the chest-tapping portion of the recording. Adjust the alignment for peak chest taps overlap between INS and smartwatch signals.
Ensure the alignment consistency throughout the time series. Then hit the switch-aligning key and initiate alignments on any remaining data streams. In 1.5 years, 293 hours of INS data, 224 hours of smartwatch data and 2037 hours of video data across three webcams were recorded, which allows longitudinal changes in neural data to be observed.
Kinematic trajectories of movement, including isolated finger movements that cannot be captured by a smartwatch can be obtained through pose estimates. These estimates can be analyzed together with other recorded modalities.
