12th GIST BCI Workshop Date : 2014. 9. 26 (Friday), 15:30 ~ Place : SIC Bldg. B 206
Organizer : Donghyeon Kim
1
Program ► Oral presentation time (40 min.) = Presentation (20min.) + Discussion (20min.)
Time
Schedule
Contents
15:30~15:40
Introduction
Greetings
15:40~16:20
Sangtae Ahn
Inter-brain synchronization
16:20 ~ 16:30
Coffee break
16:30 ~ 17:10
Yongwoo Lee
Android application for wireless BCI system
17:10 ~
Discussion
Announcement & Schedule for next workshop
2
Notice ► International conference ■ 7th International IEEE EMBS conference on Neural Engineering Conference date, place : April 22-24, 2015 , Montpellier, France Submission deadline : October 15, 2014
■ 3rd International Winter Workshop on Brain-Computer Interfaces XX Conference date, place : January 12-14, 2015, High1 Resort, Korea Submission deadline : October 15, 2014
■ 40th International Conference on Audio Speech and Signal Processing (ICASSP) 2015 Conference date, place : April 19-24, 2015, Brisbane, Australia Submission deadline : October 5, 2014
■ HCI International 2015 (HCII) Conference date, place : August 2-7, 2015, Los Angeles, CA, USA Submission deadline : October 17, 2014 3
Notice ► Journal special issue ■ Multimodal Modeling and Analysis Informed by Brain Imaging IEEE Transactions on Autonomous Mental Development Manuscript Due: October 10, 2014
■ Advances in Brain-Computer Interface Publisher : Hindawi Publishing Corporation Manuscript Due : 7 Nov, 2014 Further information : http://www.hindawi.com/journals/tswj/si/387612/cfp/
4
Planning of next workshop ► Content changes ■ Research Evgenii Kim: MSPOC Thien: Correlation btw hemodynamic changes & neuro-activity
■ Discussions None Order
Expected date
Organizer
Presenter
9th
2014-06-xx
Hohyun Cho
10th
2014-07-xx
Younghak Shin
Kyungae Yoon, Seungchan Lee
11th
2014-08-xx
Sangtae Ahn
Hohyun Cho, Donghyeon Kim
12th
2014-09-26
Donghyeon Kim
Yongwoo Lee, Sangtae Ahn
13th
2014-10-31
Seungchan Lee
Evgenii Kim, Thien
14th
2014-11-xx
Hohyun Cho
15th
2014-12-xx
Younghak Shin
16th
2015-01-xx
Sangtae Ahn
Minkyu Ahn, Younghak Shin
5
BioCompuing Lab.
Inter-brain synchronization
Sangtae Ahn Ph. D. Candidate
[email protected] 14. 09. 26
BioCompuing Lab.
Contents
Hyper-scanning History • • • •
fMRI hyper-scanning EEG NIRS MEG -
Synchronization Basic methods Neuromarker during interaction : phi complex and TPJ Conclusion and future direction
BioCompuing Lab.
HYPER-SCANNING
BioCompuing Lab.
Hyper-scanning
What does hyper-scanning means? • Simultaneous recording of multiple subjects • Measurement of social interaction btw subjects http://labs.vtc.vt.edu/hnl/hyperScanFlash.html
MEG
BioCompuing Lab.
HISTORY
BioCompuing Lab.
History
First hyper-scanning study using two fMRIs (Montague et al., Neuroimage 2002) • •
Synchronization Sender & Receiver
BioCompuing Lab.
History
Hyper-scanning fMRI studies • • • • • • • •
(Montague et al., 2002) (King-Casas et al., 2005) (Tomlin et al., 2006) (Fliessbach et al., 2007) (Chiu et al., 2008) (Polosan et al., 2011) (Anders et al., 2011) (Spiegelhalder et al., 2014)
- Handy-dandy game - Trust game - Trust game - Trust game - Trust game - Stroop test - Facial expressions (joy, anger and etc) - Live verbal communication
Results? •
Which regions are precisely activated? high spatial resolution
BioCompuing Lab.
History
Hyper-scanning fMRI studies •
Activation during social interaction or coooperation
BioCompuing Lab.
History
First hyper-scanning study using EEG (Babiloni et al., IEEE EMBS 2006) • •
Flexer and Makeig., 2007 •
Playing card game 4 subjects
Two-person game playing
Since then, more than 30 studies were published
BioCompuing Lab.
History
Hyper-scanning EEG studies • • • • • • • • •
(Lindenberger et al., 2009) (Dumas et al., 2010) (De Vico Fallani et al., 2010) (Astolfi et al., 2010) (Babiloni et al., 2011) (Wang and Jung, 2011) (Sänger et al., 2012) (Gürkök et al., 2013) (Ménoret et al., 2014)
- Playing guitar - Hand movement - Prisoner’s Dilemma Game - Prisoner’s Dilemma Game - Playing saxophone together - Button press task - Face-to-face playing guitar - Mind the sheep game - Face-to-face interaction
Results? •
Which regions are related or correlated btw two or multiple subjects? High temporal resolution
BioCompuing Lab.
History
Hyper-scanning EEG studies
BioCompuing Lab.
History
First study using NIRS (Suda et al., 2010) Face-to-face conversation
Simultaneous NIRS at prefrontal (Funane et al., 2011) Button press task
BioCompuing Lab.
History
Hyper-scanning NIRS studies • • • • • • • •
(Suda et al., 2010) (Funane et al., 2011) (Cui et al., 2012) (Dommer et al., 2012) (Holper et al., 2012) (Jiang et al., 2012) (Duan et al., 2013) (Holper et al., 2013)
Results? •
Oxy-Hb changes and coherence
- Face-to-face conversation - Button press task - Button press task - Dual n-back test - Finger-tapping task - Face-to-face dialog - Tug-of-war game - Teacher-student
BioCompuing Lab.
History
First study using MEG (Baess et al., 2012) • • •
Two MEGs apart from 5Km Audio recording Synchronization
BioCompuing Lab.
History
Second study using MEG (Hirata et al., 2014)
• •
Two MEGs in a single room Mother-child facial expression
Only two MEG studies
Results? • •
No remarkable results Just experimental setup
BioCompuing Lab.
SYNCHRONIZATION
BioCompuing Lab.
Synchronization •
Synchronization •
Christiaan Huygens (1665)
•
Types of synchrony----------------------- A : in-phase ; reciprocal B : external induced – cooperation
C : driven – PDC
D : not real (moment)
BioCompuing Lab.
Synchronization
Inter-brain synchronization? • •
Correlation Coherence measures the degree of linear dependency
• Phase synchronization • Mutual information • Connectivity between two brains (or multiple)
BioCompuing Lab.
METHODS
BioCompuing Lab.
Methods
Brain-to-brain coupling • •
Partial directed coherence • •
Covariance in amplitude or power Event-related changes from FFT
Granger casuality Multivariate autoregressive
Phase synchrony •
Phase locking value (PLV)
BioCompuing Lab.
Methods
PLV
• • • •
N : the number of trials Φ(𝑡,𝑛) : phase on trials n at time t Φ, 𝜓 : different channel 0 ≤ PLV ≤ 1
1 : perfect phase locking 0 : no phase locking
BioCompuing Lab.
Phi complex and TPJ
NEUROMARKER
BioCompuing Lab.
Neuromarker
Tognoli et al., 2007 Proceedings of the National Academy of Sciences (PNAS) Dual-EEG, 60-ch
BioCompuing Lab.
Neuromarker
Self-paced rhythmic finger movement • • •
Before-vision (opaque) During-vision (transparent) After-vision (opaque)
Phi (9.2~11.5Hz) : right centro (temporo)-parietal region
BioCompuing Lab.
Neuromarker
Phi (9.2~11.5Hz) : right centro(temporo)-parietal
Parietal and superior temporal sulcus •
Human mirror-neuron system
BioCompuing Lab.
Temporo-Parietal Junction
Temporo-Parietal Junction (TPJ) •
The TPJ is an area of the brain where the temporal and parietal lobes meet, at the posterior end of the Sylvian fissure (lateral sulcus)
BioCompuing Lab.
right TPJ
rTPJ plays a pivotal role in analyzing signals from self-produced actions as well as with signals from the external environment
rTPJ plays a role in the way individuals observe and process information, thus impacting social interaction
Blue : rTPJ
Rebecca Saxe (2006)
PDC correlation , Dumas et al., (2010)
BioCompuing Lab.
left TPJ
the left TPJ contains both Wernicke's area and the angular gyrus
• •
Language Conversation
Kawasaki et al., Scientific Reports (2013)
BioCompuing Lab.
CONCLUSION & FUTURE DIRECTION
BioCompuing Lab.
Conclusion and future direction
Conclusion •
Inter-brain correlation and synchronization
Hot topic I just think and interesting issue right TPJ Verbal communication : left TPJ Coherence, correlation, PLV and etc
Yonsei Hospital
Future direction • • •
Need of novel paradigm Need of novel method Need of MEG study Technical problems → KRISS and Yonsei Hospital
KRISS
BioCompuing Lab.
Thank you Q&A
Android Application for Wireless BCI System Yongwoo Lee
Sep.26.2014 SIC Building B, 206 12th BCI Workshop
Contents I.
Demand for Mobile BCI
II.
Strong point of Android as Mobile BCI platform
III. INFONET BCI I.
Structure of INFONET BCI
II.
Strong point of this Application
III.
A demonstration
IV. A fast BCI keyboard for Android
Demand for Mobile BCI
1. BCI as a wearable Device •
Google glass with EEG
2. A great combination for wireless BCI
Google glass with EEG
• Google glass + MindWave Mobile NeuroSky • Taking picture by EEG Signal. • http://tvpot.daum.net/v/v8f6cMifWGWfPSP9wmPM LLw
A great combination for wireless BCI
• The wireless BCI should be used on wireless device. • Mobile devices are the most wireless, portable. • INFONET BCI is developed based on SeungChan`s wireless BCI device.
Strong point of Android as a Mobile BCI platform
1. Market share 2. Java – widely used language
Market share • Android is on Top of market share of Smartphone. • Also increasing.
Java – widely used language • Basically, android is based on Java. • Java is the most used language. • Lots of open libraries.
INFONET BCI
1.
Structure of INFONET BCI
2.
Strong point of this
Application 3.
A demonstration
Structure of INFONET BCI • Resembles bci2000. • Has main Operator. • Source Module, Signal processing Module, Application Module.
Structure of INFONET BCI(Cont.) • 3 important activities • Module selecting activity.
Structure of INFONET BCI(Cont.) • 3 important activities • Module selecting activity.
Structure of INFONET BCI(Cont.) • Application Activity
Structure of INFONET BCI(Cont.) • Application Activity – For Bluetooth.
Structure of INFONET BCI(Cont.) • Monitoring Activity
Structure of INFONET BCI(Cont.) • Monitoring Activity
Strong Point
1. High Compatibility 2. Easy to upgrade 3. Various Usage.
• It is designed for Seung-Chan`s device. • 100% Compatibility.
Strong Point
1. High Compatibility 2. Easy to upgrade 3. Various Usage.
• Modular programming • Develop additional module and adapt it. • If want, it is also possible develop a module for new wireless device.
Strong Point
1. High Compatibility 2. Easy to upgrade 3. Various Usage.
• Since it is easy to adapt new functions, It can have variable usage as updates.
Demonstration
A fast BCI keyboard for Android
A defect of Android for BCI key board
• Small display. • Not easy to use SSVP or P300 • Binary search is only way.
Binary selection is available.
• 6 selections/min -> to slow. • A new Algorithm for search is required.
Available choices.
• Weighted Tree of Characters. • Auto text. • Etc….
Thank You
[email protected] for further questions