BZA BCI Projects

1
Selective attention SSVEP
Herrmann et al, Exp. Brain Research 2001
2
SSVEP
6 Hz
15 Hz
Steady state visual evoked potential (SSVEP) BCI
(Kelly et al., 2005)
3
SSVEP
Data from an SSVEP BCI using 6 and 15 Hz
checkerboxes (Allison et al., 2008)
4
SSVEP IAT Bremen
SSVEP spelling system
13 14 15 16 17
5
SSVEP IAT Bremen
6
CeBIT 2008
BCI performance by age and gender.
7
CeBIT 2008
Replies to post test questionnaire items by age.
8
Conventional BCI View

• Why use a BCI if youre healthy?
• BCIs
• Only provide communication.
• Provide the same information available via
  conventional interfaces.
• Are exclusive interfaces.
• Thus are of no practical value to people who can
  otherwise communicate.
• Will not attain wider adoption without dramatic
  improvements in information transfer rate (ITR).

9
Conventional BCI View

• Replacing conventional interfaces for
  conventional users in conventional settings.

The future of BCIs is not Keanu Reeves.
10
Emerging User Goals

• Replacing conventional interfaces for disabled
  users in conventional settings.
• BOTH for communication and rehab.
• Replacing conventional interfaces for
  conventional users in specific settings.
• Supplementing conventional interfaces.

11
Emerging User Groups

• Rehabilitation stroke, autism, attention
• Less disabled users
• Healthy users
• Gamers
• Surgeons, drivers, soldiers, mechanics
• Lazy people

12
BCI Stroke Rehabilitation
13
BCI Autism Rehabilitation
UCSDnews.ucsd.edu
14
Consumer games
15
New User Groups Who and When?
Moderately disabled
Eager healthy
Mainstream healthy
Schalk (2008)
16
Induced Disability

• Bayliss et al (2000, 2003) for drivers,
  household control
• Middendorf et al (2000) for pilots
• Pineda et al (2003), Scherer (2007) virtual
  navigation
• Trejo et al (2006), Millan (2006) for
  astronauts
• Also cellphone users, mechanics, surgeons,
  soldiers
• And (sadly) disability by laziness remote
  control or cellphone users

Trejo et al (2006)
Scherer et al (2007)
17
BCI Limitations

• BCIs have poor information throughput (ITR)!
• BCIs may require bulky equipment and/or surgery.
• BCIs may work poorly in realworld environments.
• BCIs are harder to use than other interfaces.
• BCIs may not work with everyone.
• BCIs may require training.
• BCIs are expensive.
• BCIs are unsupported by common software.
• BCIs are unfashionable (to tasteless people).
• BCIs are unknown or scary to the general public.
• BCIs are exclusive interfaces.

18
BCI Limitations

• BCIs have poor information throughput (ITR)!
• Some BCIs require bulky equipment and/or surgery.
• Some BCIs work poorly in realworld environments.
• Some BCIs are harder to use than other
  interfaces.
• Some BCIs do not work with everyone.
• Some BCIs require training.
• Some BCIs are expensive.
• Most BCIs are unsupported by common software.
• BCIs are becoming more wearable and fashionable.
• BCIs are becoming more popular and respected.
• Are BCIs exclusive interfaces?

19
BCI Advantages

• Better ITR than an unavailable interface
• That requires impractical hardware
• That cannot be easily used
• That could never provide the same information
• BCIs may be easier to use than other interfaces
• More portable, accessible, or convenient in real
  world settings
• Induced disability is major, even from laziness
  (TV remote control)
• More natural and intuitive
• Less training
• BCIs are the only interface capable of total
  privacy
• BCIs may be faster than other interfaces
• BCIs may seem more novel or fun
• BCIs might supplement other interfaces (hybrid
  BCI)

20
Key BCI factors
21
BCI Technologies
Sensors
Cognemes
Side Effects
Related Technologies
Electronics
Signal Processing
HCI
Cognitive Neurosci.
Communications
Medicine
Psychology
Manufacturing
Wearable Computing
ExG Sensors
22
BCI Technologies
Sensors
Cognemes
Side Effects
Related Technologies
Electronics
Signal Processing
HCI
Cognitive Neurosci.
Communications
Medicine
Psychology
Manufacturing
Wearable Computing
ExG Sensors
23
BCI Technologies
Sensors
Cognemes
Side Effects
Related Technologies
Electronics
Signal Processing
HCI
Cognitive Neurosci.
Communications
Medicine
Psychology
Manufacturing
Wearable Computing
ExG Sensors
24
BCI Technologies
Sensors
Cognemes
Side Effects
Related Technologies
Electronics
Signal Processing
HCI
Cognitive Neurosci.
Communications
Medicine
Psychology
Manufacturing
Wearable Computing
ExG Sensors
25
Future directions
Future progress hinges on recognition that BCI
development is an interdisciplinary problem,
involving neurobiology, psychology, engineering,
mathematics, computer science, and clinical
rehabilitation Wolpaw et al.,
2002. Communications, linguistics, HCI, and
human factors are also important.
26
Acknowledgements
Thanks to Chris Agocs, Jacqueline Boccanfuso, Ben
Chi, Adriane Davis, Umang Dua, Bernhard Graimann,
Axel Graeser, David Leland, Thorsten Lueth, Luke
McCampbell, Dennis McFarland, Melody Moore
Jackson, Jaime Pineda, John Polich, Samir Ramji,
Gerv Schalk, and Diana Valbuena for help with
work presented here. Thanks to my colleagues in
the BCI community!
Attendees of the BCI conference in New York
27
Websites with more info
www.bci-info.org http//www.cis.gsu.edu/brainlab/
bci.ucsd.edu www.cyberkineticsinc.com www.neurals
ignals.com Additional videos are available from
the Wolpaw lab www.bciresearch.org More videos
from the Pfurtscheller lab http//bci.tugraz.at/m
ovies.html Other BCI websites are easy to find
online. The 60 Minutes video shown on 2 November
is easy to find on the CBS site. Please contact
the author for recent BCI articles, updates on a
summer 2009 class on BCIs or Cog Sci 17, specific
references for articles or other
materials allison_at_iat.uni-bremen.de
28
Thank you
Thats all, folks!