Babiloni Slides

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 1/29

Brain Computer Interface forcommunication and control

Fabio Babiloni

Dept. Human Physiology and PharmacologyUniversity of Rome, “La Sapienza”

Rome, Italy

IRCCS “Fondazione Santa Lucia”, Rome, Italy

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 2/29

Human computer interfaces

In the classical Star Wars third movie (the returnof Jedi) Darth Vader reveals a connection between

his neural system and the computer

Today, such high levelof integration betweenman and machine seemsreally yet too far fromthe common practice

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 3/29

Overview of the presentation

Future trends

Definition of a Brain Computer Interface

Principalneurophysiologicalsignals that can be

used to do the job

The most active

research groups inthe BCI field andtheir achievements

Nicolelis, Nature 20

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 4/29

Brain-Computer communication

through EEG

“Brain–computer interfaces (BCI’s) give their userscommunication and control channels that do notdepend on the brain’s normal output channels of

peripheral nerves and muscles.”

“A BCI changes the electrophysiological signals frommere reflections of CNS activity into the intendedproduct of the activity: messages and commands

that act on the world”Wolpaw, 2002

Feedback and biological adaptation Nicolelis, Nature 2001

Acquisition

or estimationof thecorticalactivity

Processing andclassification ofcortical signals

Actuation inthe real world

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 5/29

The most downloaded paperfrom Clinical Neurophysiology

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 6/29

Variations of EEG waves are

correlated with some mental states

8-12 Hertz,

alpha EEGwaves

8-12 Hertz,mu EEGwaves

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 7/29

Movement-related thoughts elicited

specific cortical patterns

Several EEG studies havebeen also demonstrated thatimagined movements eliciteddesynchronization patterns

different for right and leftmovement imaginations

Neuroscientific studies withfMRI have demonstrated thatmotor and parietal areas areinvolved in the imagination ofthe limb movements

Imagined left movement Executed left movement

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 8/29

Motor cortical activity in tetraplegics

Shoam et al. Nature vol 413 2001

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 9/29

MRPs Right finger movement alpha ERD

A closer look into the brain

dynamics underlying the movementpreparation and execution

From –1 before (movie start) to +0.1 sec post-movement

Where: centro-parietal scalp area

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 10/29

On the use of neurophysiological signals

to control devicesEEG, EMG, EOG

– Quality of sensors

– SNR (EMG >>10, EEG ≈ 1)

Actuators

Featureextraction

PatternRecognition

-Time-dependent features- Times series values

-Frequency dependent features– AR, FFT, Wavelet

-LDA, MDA-Non linear classifier

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 11/29

Present-days BCIs

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 12/29

Threshold classifiers for the Brai

Computer Interface (Tubingen)

Institute of Medical Psychology andBehavioural NeurobiologyDepartment chair: Prof. Niels Birbaumer

Dr. Andrea Kübler -biologist

Nicola Neumann - psychologistSlavica Coric - assistantDr. Thilo Hinterberger - physicistDr. Jochen Kaiser - psychologistDr. Boris Kotchoubey - psychologist, physician

Dr. Jouri Perelmouter - mathematician

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 13/29

Patient HPS using the

Thought Translation Device

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 14/29

Present-days BCIs

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 15/29

Left Right

Unbalance of ERD for imagined

left and right movements

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 16/29

EEG patterns related tocognitive tasks Power spectrum increase/decrease of EEG data recorded when

subject imagines or performs a movement of his middle finger.

δ

θ

α

β

γ

Babiloni et al., IEEE Tr. Rehab. Eng., 2000

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 17/29

Brain Computer Inter aces

at the Graz UniversityProf. Gert Pfurtscheller 

Mu-rhythms patternrecognition by linear andnon linear classifiers

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 18/29

The Adaptive Brain Interface

José del R. Millán

Josep MouriñoMarco Franzè

Fabio TopaniAdriano Palenga

Fabrizio Grassi

Maria Grazia Marciani

Donatella Mattia

Febo Cincotti

Fabio Babiloni

Markus VarstaJukka Heikkonen

Kimmo Kaski

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 19/29

ABI Training

6.40-7.30

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 20/29

Brain-operated Virtual Keyboard

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 21/29

A gameapplication

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 22/29

Finalist to the Descartes prize 2001

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 23/29

Present-days BCIs

l ’ d h

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 24/29

Wolpaw’s Wadsworth

Center

Spelling device(2.25)Aid screen

P300 spelling

device

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 25/29

BCI controlled by

estimated cortical activity

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 26/29

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 27/29

Future trends: increase awareness

of controlled devices

BCI is a slow communication channel– Best performance with virtual keyboard: 3characters per minute

Need for “smart” devices, e.g.:– T9 programs for SMS on cellular phones– Trajectory aware weelchairs or robotic

arms

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 28/29

EEG Based BCI inrehabilitationFocus: degree of Autonomy

– Partially restoring the abilities, mostly usingalternative strategies

– Communication aid-> Controlling device

Focus: degree of Functional Recovery

– Tuning of the rehabilitation actions to maximize levelof recovery– Cortical plasticity->Rehabilitation device

F t t d

8/9/2019 Babiloni Slides

http://slidepdf.com/reader/full/babiloni-slides 29/29

Future trends Identification of those signals, whether evoked potentials,

spontaneous rhythms, or neuronal firing rates, that users arebest able to control independent of activity in conventionalmotor output pathways;

Development of training methods for helping users to gainand maintain that control

Delineation of the best algorithms for translating thesesignals into device commands;

Identification and elimination of artifacts such aselectromyographic and electro-oculographic activity;Adoption of precise and objective procedures for evaluating

BCI performance;

Identification of appropriate BCI applications andappropriate matching of applications and usersAttention to factors that affect user acceptance of

augmentative technology, including ease of use, cosmesis, and

provision of those communication and control capacities thatare most important to the user