Knowledge Vault 3/36 - G.TEC BCI & Neurotechnology Spring School 2024 - Day 3
Hyperscanning EEG recordings from multiple subjects with cognitive load
Francisco Fernandes, g.tec medical engineering GmbH (AT)
<Resume Image >

Concept Graph & Resume using Claude 3 Opus | Chat GPT4 | Llama 3:

graph LR classDef hyperscanning fill:#f9d4d4, font-weight:bold, font-size:14px; classDef applications fill:#d4f9d4, font-weight:bold, font-size:14px; classDef experiments fill:#d4d4f9, font-weight:bold, font-size:14px; classDef tools fill:#f9f9d4, font-weight:bold, font-size:14px; classDef future fill:#f9d4f9, font-weight:bold, font-size:14px; A[Francisco Fernandes] --> B[Hyperscanning: simultaneously
recording brain activity. 2] A --> C[Francisco: hyperscanning experiments,
G-Tech equipment. 1] B --> D[Applications: education, entertainment,
healthcare, marketing. 3] B --> E[1965: EEG synchronization,
identical twins. 4] B --> F[Hyperscanning improves brain-computer
interface accuracy. 7] B --> G[Hyperscanning studies empathy,
mother-child synchrony. 25] B --> H[Hyperscanning: social interactions,
teamwork, dyads. 29] C --> I[G-Tech setup: EEG caps,
projection, PC. 5] C --> J[2013: 8 subjects, P300 speller,
100% accuracy. 8] C --> K[2022: subjects, rooms,
D2 test. 9] C --> L[Demo: 4 g.HIamp,
2 g.Nautilus subjects. 10] L --> M[Tasks: fixation cross,
visual attention. 11] L --> N[Real-time engagement displayed
per subject. 12] L --> O[Paradigm: alternating rest,
attention tasks. 13] L --> P[Francisco: sets up
8-channel caps. 14] L --> Q[Simulink: 32 g.HIamp,
16 g.Nautilus channels. 15] Q --> R[Markers: Unity to
Simulink via LSL. 16] Q --> S[Simulink: filtering, feature
extraction, LDA classifier. 17] Q --> T[Script trains LDA
post unclassified run. 18] Q --> U[Real-time: classified state
aligns with tasks. 19] Q --> V[Stopping task decreases
classified attention. 20] Q --> W[Multi-device block:
synchronizes EEG streams. 21] Q --> X[LSL: markers: Simulink:
high-bandwidth EEG. 22] C --> Y[Demo showcases complex
setup's ease. 27] C --> Z[G-Tech tools enable
hyperscanning experiments. 30] B --> AA[2016: pilots' EEG
synchrony, tasks. 6] B --> AB[Multiple subjects improve
BCI accuracy. 28] B --> AC[Engagement: alpha, theta changes
between states. 26] A --> AD[Astronauts, ISS, moon
hyperscanning: interesting. 23] A --> AE[October: zero-gravity flight
experiment, G-Tech. 24] class A,B,E,F,G,H,AA,AB,AC hyperscanning; class D applications; class C,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z experiments; class AD,AE future;


1.- Francisco presents hyperscanning experiments using G-Tech equipment with multiple subjects performing tasks simultaneously.

2.- Hyperscanning is the simultaneous recording of brain activity from two or more subjects who are interacting or performing the same task.

3.- Hyperscanning has applications in education (studying student performance), entertainment (audience engagement), healthcare, marketing, and social interaction studies.

4.- A 1965 study found synchronization of EEG activity between identical twins, demonstrating early interest in hyperscanning.

5.- G-Tech's setup includes EEG caps on subjects, a projection wall displaying tasks, and a PC processing the hyperscanning data.

6.- A 2016 study examined EEG synchrony between airplane pilots during takeoff, cruise, and landing while addressing an electrical fault.

7.- Hyperscanning can enhance accuracy in brain-computer interface applications by combining data from multiple users.

8.- In a 2013 G-Tech experiment, 8 subjects collaborated on a P300 speller task, achieving 100% accuracy after combining their data.

9.- At the 2022 spring school, a hyperscanning study was conducted with subjects in different rooms performing tasks like the D2 test.

10.- In today's live demo, 4 subjects with 8-channel caps will be recorded with a g.HIamp, and 2 with g.Nautilus.

11.- Subjects will perform a fixation cross resting task and a demanding visual attention task selecting specific targets within time constraints.

12.- Real-time engagement levels will be displayed for each subject based on their EEG during the resting and attention tasks.

13.- The experimental paradigm alternates between resting and attention tasks for 2.5 and 1 minute respectively, totaling around 15 minutes.

14.- Francisco sets up the 8-channel electrode caps on the 6 volunteer subjects, ensuring proper positioning and electrode contact.

15.- The Simulink model acquires 32 EEG channels (8 each from 4 subjects) from the g.HIamp and 16 channels from the g.Nautilus.

16.- Markers indicating the task state are sent via lab streaming layer (LSL) from the Unity paradigm to the Simulink model.

17.- The Simulink model applies filtering, separates EEG bands, extracts features, and trains an LDA classifier to detect attentional states.

18.- After the initial unclassified run, a script processes the data to train the LDA classifier for real-time classification.

19.- In the real-time feedback run, fluctuations in the classified attentional state align with the task periods.

20.- Stopping the attention task causes the classified attentional state to decrease, demonstrating the validity of the real-time EEG classification.

21.- The multi-device block in Simulink enables synchronization of multiple EEG streams without additional hardware triggering.

22.- LSL is used for lower bandwidth event markers, while high bandwidth EEG data is shared directly via memory in Simulink.

23.- Frankie is unsure if hyperscanning has been used with astronauts on the ISS, but suggests it would be interesting, even on the moon.

24.- G-Tech equipment will be used in a zero-gravity flight experiment in October.

25.- Hyperscanning has been used to study empathy, such as synchrony between mother and child or in response to emotional stimuli.

26.- Engagement is quantified by relative changes in EEG alpha and theta band power between attentional and resting states.

27.- The demo showcases the ease of setting up a complex 6-subject hyperscanning experiment in a short amount of time.

28.- Combining data from multiple subjects can improve BCI accuracy, even with a few poor performers, by averaging.

29.- Hyperscanning enables studying brain synchronization during social interactions like teamwork, between couples, twins, or parent-child dyads.

30.- The presentation demonstrates G-Tech's hardware and software tools for easily implementing hyperscanning experiments with real-time EEG processing and feedback.

Knowledge Vault built byDavid Vivancos 2024