User Testing 2

On Thursday October 12, 2023 at Alberta Children’s Hospital, BCI4Kids Lion’s den.

User(s):

• Beverly Collison
• Marie Dylke
• Georgie Brewin

Student(s)

• Beverly’s 2 pieces – the Real users

Dev(s):

• Anup
• Emily
• Dan

Set-up

Used DSI-Flex connected to “Morpheus” desktop computer via USB.

What versions were tested

• CLIP main version dfc0559.
• CLIP-Unity was confirmed to be working from both the source files run out of Unity editor and from the exe compiled by Elena that morning.
• CLIP-Python was confirmed to be working from the source files run from a Python terminal but not working from the exe.

How it went

• The set-up for the actual user testing was using the Unity exe and Python source files with the DSI-Flex connected by USB.
• Emily helped set-up the DSI-Flex on the Students. Contact quality wasn’t 100% on all electrodes – an average of 4/7 electrodes were “green” and the other 3/7 were “yellow” as displayed by the DSI software.
• Anup launched the Unity game, created a profile, entered the main scene. Launched Python files after the game was in the main scene.
• The Users set up the shelf above the monitor. Anup set-up the digital shelf to recreate the visual scene in-game.
• Greyscale and Outline features were demonstrated, but not used for the purpose of the demo.
• Anup initiated training. Anup had to instruct the students on which stimulus they needed to look at during training, as the “indicator” used to direct attention to the given training stimulus (the quick grow-shrink animation) was very fleeting.
• After training, we ran 4 stimulus selection opportunities to each Student. The Students were directed by the Users to just look at an instrument of their choosing. The correct selection was made (in reference to the student’s declared preference) 3/4 times in the older student (~8 years old) and 2/4 times in the younger student (~5 years old)

Bugs noted

1. Major – LSL output stream from Unity only happening on primary shelf scene, but not on menu scenes. The Unity software doesn’t send a LSL marker stream when not in the primary shelf scene (e.g. in the main menu or options menus) – This cause the Python software to bug out and/or crash when it stops receiving Unity markers via LSL, such as from the beginning of the software, or when the User leaves the primary shelf scene.
2. Major – The Python exe doesn’t work.

What needs to be done.

1. Fix the LSL output stream bug

2. Fix the Python exe bug.

3. Finish hooking up the “Display settings”

4. Hook up the BCI stimulus settings

   1. flash - pretty much done
   2. texture - not too much harder than flash, similar
   3. zoom bounce - this is the bigger challenge <—- defer this

5. Verify the different BCI stimulus settings actually work

   1. test flash
   2. test texture
   3. test combinations - there is uncertainty here, probably can’t do it <— defer

6. Is the method used for directing the training appropriate? The “stimulus indicator” during training is too fast.

   1. this is a defect (Eli can fix it, probably, if not defer)

7. Prevent stimulus training and stimulus selection from being initiated while the inventory menu is open

8. Instrument transparency – should it remain transparent

Deadline:

• Nov 27 the interventions start
• Internal deadline of Nov 20-22 is needed.