966IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 51, NO. 6, JUNE 2004
Principles of a Brain-Computer Interface (BCI)
Based on Real-Time Functional Magnetic Resonance
Nikolaus Weiskopf*, Klaus Mathiak, Simon W. Bock, Frank Scharnowski, Ralf Veit, Wolfgang Grodd,
Rainer Goebel, and Niels Birbaumer
Abstract—A brain-computer interface (BCI) based on func-
tional magnetic resonance imaging (fMRI) records noninvasively
activity of the entire brain with a high spatial resolution. We
present a fMRI-based BCI which performs data processing and
feedback of the hemodynamic brain activity within 1.3 s. Using
this technique, differential feedback and self-regulation is feasible
as exemplified by the supplementary motor area (SMA) and
parahippocampal place area (PPA). Technical and experimental
aspects are discussed with respect to neurofeedback. The method-
ology now allows for studying behavioral effects and strategies of
local self-regulation in healthy and diseased subjects.
Index Terms—Brain-computer interface, blood oxygen level-de-
pendent, neurofeedback, self-regulation.
regulation of electrical brain activity , . Specific be-
havioral effects dependent upon the functional role of the
cephalography (EEG) have been used to train voluntary
Manuscript received July 11, 2003; revised February 6, 2004. This work was
supported in part by the WIN-Kolleg of the Heidelberg Academy of Sciences
and Humanities (Germany), in part by the Deutsche Forschungsgemeinschaft
(DFG) under Grant SFB 550/B5+B1 and Grant Th 812/1-1, and in part by the
National Institute of Health (NIH). Asterisk indicates corresponding author.
*N. Weiskopf is with the Institute of Medical Psychology and Behavioral
Neurobiology, University of Tübingen, Tübingen, Germany and also with the
Section of Experimental MR of the CNS, Department of Neuroradiology, Uni-
versity of Tübingen, Tübingen, Germany (e-mail: nikolaus.weiskopf@uni-tue-
K. Mathiak is with the Center for Neurology, University of Tübingen,
Tübingen, Germany (e-mail: email@example.com).
S. W. Bock is enrolled in the Graduate School of Neural and Behavioral
Sciences and the International Max Planck Research School, University of
Tübingen, Tübingen, Germany (e-mail: firstname.lastname@example.org).
F. Scharnowski is enrolled in the Graduate School of Neural and Behav-
ioral Sciences and the International Max Planck Research School, University
of Tübingen, Tübingen, Germany (e-mail: email@example.com-
R. Veit is with the Institute of Medical Psychology and Behavioral Neurobi-
ology, University of Tübingen, Tübingen, Germany (e-mail: ralf.veit@uni-tue-
W. Grodd is with the Section of Experimental MR of the CNS, University
of Tübingen, Tübingen, Germany (e-mail: firstname.lastname@example.org-
R. Goebel is with the Department of Cognitive Neuroscience, Faculty of
Psychology, University of Maastricht, Maastricht, The Netherlands (e-mail:
N. Birbaumer is with the Institute of Medical Psychology and Behavioral
Neurobiology, University of Tübingen, Tübingen, Germany, and also with the
Center for Cognitive Neuroscience, University of Trento, Trento, Italy (e-mail:
Digital Object Identifier 10.1109/TBME.2004.827063
regulated brain area have been shown . However, the re-
gional specificity of the feedback is restricted by the relatively
low spatial resolution of the EEG. In contrast, functional mag-
netic resonance imaging (fMRI) measures the blood oxygen
level-dependent (BOLD) signal with high spatial resolution
and can access the whole brain (for a review cf. ). Electric
brain activity as measured by EEG and the BOLD signal seem
to be highly correlated (e.g., ).
In order to increase the spatial specificity of neurofeedback
training, different studies investigated the feasibility to use
fMRI for operant training of the BOLD-response with feedback
of the BOLD signal as reward. Two previous studies applied
delayed feedback (by approx. 60 s) of BOLD-responses by
ratings of the experimenter based on the local BOLD signal
. In two other studies, immediate feedback of local BOLD
signals was provided to further facilitate learning , .
We will discuss the technical and experimental basis of a
brain-computer interface (BCI) based on real-time fMRI and
present an implementation for neurofeedback.
II. TECHNICAL AND EXPERIMENTAL ASPECTS
In this section, we consider the demands on the fMRI-BCI
as concerns speed, signal-to-noise ratio (SNR), artifact suppres-
sion, and usability of the feedback signal.
A. Temporal Properties of BOLD-Response Feedback
is facilitated by minimum delays between physiologic activity
and feedback . However, the BOLD signal reflects vascular
effects and only indirectly the neuronal signal. The hemody-
namic response function (see inset in Fig. 1) introduces a physi-
ological delay of 3 to 6 s before signal changes can be observed.
Finally, the cognitive processing of the feedback signal in-
troduces an additional delay into the regulatory loop. We sug-
gest to reduce the data, e.g., from an appropriate combination
of regions of interest (ROI’s), to one dimension for easy and
unequivocal design of a fMRI-BCI and to facilitate cognitive
processing (see “BCI” in Fig. 1).
B. Signal- and Contrast-to-Noise Ratio
1) Field Strength: Contrast-to-noise ratio (CNR), i.e., the
sensitivity to the BOLD effect, and SNR determine the relia-
bility of the feedback signal and indirectly how fast changes in
0018-9294/04$20.00 © 2004 IEEE
WEISKOPF et al.: PRINCIPLES OF A BCI BASED ON REAL-TIME fMRI 967
activity is measured by fMRI using the blood oxygen level-dependent (BOLD)
effect which resembles the neurovascular response to electric brain activity
(red curve superimposed over the rendered brain). The measured hemodynamic
response is delayed by approx. 3–6 s from the neuronal activity. Images
are reconstructed, undistorted, and averaged on the “Trio” scanner console.
Turbo-BrainVoyager retrieves the data via local area network (LAN), and
performs data preprocessing (including 3D motion correction) and statistical
analysis. The signal time-series of interactively selectable regions of interest
are exported via LAN to the custom-made visualization software “BCI” which
provides feedback to the subject using video projection. Feedback is presented
with a delay of less than 1.3 s from time of image acquisition.
Setup and data-flow of the fMRI brain-computer interface. Local brain
the feedback signal can be observed. At higher field strengths,
the magnetization of the probe increases, and, thus, the MR
signal and SNR might increase linearly with magnetic field
strength. However, higher field strengths lead to more distor-
tions and signal dropouts in the image due to larger frequency
offsets in the object. Moreover, a shorter apparent transversal
and a longer longitudinal relaxation time
ameliorate the theoretical gain as well as safety problems
such as radio frequency (RF) deposition. In general, the range
of 3–4 T seems to be an appropriate compromise for BOLD
sensitive imaging .
2) Single-Shot Multi-Echo Acquisition: The BOLD signal
depends upon the echo time (TE). For the apparent transversal
, the expected contrast is proportional to
and maximal for TE close to
pling of multiple images after a single RF excitation pulse and
their appropriate combination, thus, increases CNR without the
disadvantageous effects of increased readout bandwidth .
. The sam-
C. Real-Time Artifact Control
1) Head Motion: Even the movement of a minor fraction
of a resolution element, i.e., below 1 mm, can induce signal
changes comparable to the one achieved by maximal brain ac-
tivation. Efficient real-time motion correction should be robust
and should not add significant noise and artifact sources .
Mathiak et al.  showed that at least 3 slices should be ac-
quired in order to reduce movement noise to less than 1% of the
voxel size in case of a 64
64 image matrix.
2) Image Distortions: Single-shot image encoding as used
for echo-planar imaging (EPI), i.e., one RF excitation pulse
per image, suffers from significant image distortions in case
of off-resonances caused by inhomogeneities of the static
magnetic field which can be observed close to air-tissue inter-
faces. These distortions depend upon motion and, thus, could
simulate BOLD signal changes by head motions which cannot
be corrected by motion correction. We recently developed
a multi-echo EPI technique that dynamically reduces image
distortions in a single-shot . This technique is particularly
immediate and does not depend on previous (time demanding)
reference scans (e.g., ).
3) Signal Dropout: Inhomogeneities of the magnetic field
may also cause signal dropouts and lead to a reduced BOLD
by adapted MR pulse sequences , but, in general, speed or
SNR is traded for signal recovery. We focused on areas that
were unaffected by large signal dropouts such as supplementary
motor area (SMA) and parahippocampal place area (PPA).
A. Participants and Experimental Paradigms
area (SMA) and the parahippocampal place area (PPA) using
the fMRI-BCI. Three of the subjects were members of the re-
search group and experienced in fMRI and neurofeedback ex-
periments. Subject 2 was naïve with respect to fMRI and the
task. Written informed consent was obtained prior to the study
in accordance with the local ethics committee.
followed by feedback sessions. The first localizer session was
run to delineate the SMA. It consisted of five baseline blocks
separated by three blocks of bimanual finger tapping (50 s each;
e.g., ). The second localizer session determined the location
were presented to the subject . Beginning and ending with
a baseline block, three PPA blocks alternated with three FFA
blocks (37.5 s each) and were separated by baseline blocks (25
s each). During the baseline blocks, the subject counted from
During feedback sessions, the difference of the mean BOLD
signal of two regions of interest (ROI) approximating SMA and
Each rectangular ROI [Fig. 2(a)] was based on the preceding
functional localizers and enclosed in average about 32 voxels
for the SMA and 29 voxels for the PPA. During up-regulation
blocks, the subjects should raise the curve, and during down-
regulation blocks, they had to lower the curve. Three up-regu-
lation blocks alternated with three down-regulation blocks (50
s each) and were separated by baseline blocks [31 s; see also
Fig. 2(b)]. Subjects counted down from 100 during baseline
jects1and2, sessionswereacquired ontwo consecutivedays(3
and 4 feedback sessions on the first day, respectively). Subjects
2 and 3 saw inverted curves to rule out effects of the direction.
The participants should use an individual strategy to control
the differential BOLD signal. However, visual imagery (e.g.,
968 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 51, NO. 6, JUNE 2004
to the experimenter. On the left panel statistical maps were superimposed
over one oblique echo-planar image. Green spots indicated areas which were
activated during up-regulation blocks, and blue spots indicated areas which
were activated during down-regulation blocks. Regions of interest were marked
as rectangles which approximated the supplementary motor area (red), and
parahippocampal place area (green). On the right upper and middle panel
time courses of SMA and PPA, respectively, were plotted as white curves
superimposed over the experimental block design which was represented by
grey (baseline), green (up-regulation), and blue (down-regulation) stripes.
The right lower panel displayed the estimated and corrected head motion.
Statistical maps and time courses were updated continuously within 1.3 s
after image acquisition. (b) Feedback screen presented to the subject. The
differential BOLD signal was presented as a constantly updated yellow curve
on a color-coded background (signal of SMA minus PPA). The tasks were
presented as colored stripes: grey indicated the baseline, during green blocks
subjects had to raise the curve (up-regulation), and during blue blocks subjects
had to decrease the curve (down-regulation). The red curve is the low-pass
filtered (Gaussian ???? ? ?? stime points) time-series. Arrows and
filtered curve were not presented during on-line feedback. Mean and standard
deviation (std)were estimated based on the data of the first baseline block.
On-line statistical analysis and neurofeedback. (a) Display presented
enteering) or motor imagery (e.g., fist clenching, dancing; )
served as a general starting point. Moreover, subjects were in-
structed that the feedback signal was delayed by approx. 6 s and
that they should not move any body parts and should breathe
B. fMRI Data Acquisition and Image Reconstruction
For all functional scans, 10 oblique transversal-coronal
slices were acquired using a single-shot multi-echo EPI
(echo-planar imaging) sequence (repetition time
, effective echo times
kHz/pixel). The echoes
and parahippocampal place area (PPA) across feedback sessions. In all subjects
significant task correlated changes in the differential feedback signal were
observed (t-value ????? is equivalent to ? ? ????). In subjects 1 and 2, the
effect increased across feedback sessions (? ? ????; one-sided Pearson).
Differential regulation of activity in supplementary motor area (SMA)
were recorded with alternating phase encoding direction, cor-
rected for geometric distortions  and averaged to improve
BOLD sensitivity in real-time , . The first ten volumes
of each session were discarded to suppress
effects. To improve BOLD sensitivity, the experiments were
performed at a high static magnetic field (3 T) on a whole body
scanner equipped with a volume head coil (Magnetom Trio,
Siemens, Erlangen, Germany).
C. On-Line Statistical Analysis and Visual Feedback
Statistical analysis was performed by Turbo-BrainVoyager
rate personal computer retrieving the image files as soon as they
were created by the image reconstruction system via local area
network (LAN). The program performed data preprocessing,
statistical analysis, and export of ROI time-courses to hard disk
in real-time [Fig. 2(a)]. Preprocessing of the data included in-
cremental linear detrending of the time-series and 3D motion
correction. Statistical analysis based on a general linear model
(GLM) was performed cumulatively using the recursive least
squares regression algorithm.
For separate treatment of visual feedback of brain activity
from statistical analysis we developed custom-made software
(“BCI”; Fig. 1) based on Matlab 6.5 (The MathWorks, Natick,
Voyager which included information about the paradigm and
imaging parameters. During the fMRI session, it accessed the
and displayed the feedback curve [Fig. 2(b)].
IV. RESULTS AND DISCUSSION
All subjects achieved significant differential BOLD am-
plitudes between SMA and PPA (Fig. 3) as determined by
off-line ROI analysis . This analysis of the differential
feedback signal was performed with custom-made software
based on SPM99 (Wellcome Department of Imaging Neu-
roscience, Queens Square, London, U.K.); time-series were
WEISKOPF et al.: PRINCIPLES OF A BCI BASED ON REAL-TIME fMRI 969
tation) were included into the GLM . In subjects 1 and 2,
control of the feedback signal increased significantly across
training sessions (
; one-sided Pearson; Fig. 3).
Taken together, the present study showed the feasibility of
on-line differential feedback and regulation of local BOLD ac-
tivity using a fMRI-BCI. However, potential learning effects
have to be further assessed in a larger sample size across more
sessions to allow for conclusions on the efficiency of this neu-
s) and low-pass filtered
s), and motion parameters (3 translation, 3 ro-
Several extensions of the current technical and experimental
approach are possible. Recently, fast imaging sequences
reducing signal dropouts in areas with large magnetic field
inhomogeneities became available , . Thus, activity of
brain regions such as amygdala and orbitofrontal cortex can be
measured. Moreover, prospective motion correction techniques
can further reduce artifactual feedback signals by continuously
tracking the brain with the measured volume . Real-time
cortex-based statistical analysis focuses on the grey matter of
the cortex with high spatial precision and will possibly improve
the spatial specificity and sensitivity even further .
Subcortical and orbitofrontal structures which are involved
in emotional regulation could be a target for a fMRI based neu-
rofeedback training. In anxiety disorders or other mood disor-
ders, dysregulation in areas not accessible to other neurofeed-
back techniques such as EEG-BCI were reported. The training
itself might be integrated in a broader therapeutical approach.
cortical potentials (SCP), which are successfully used by pa-
amus and basal ganglia may improve the control of SCP .
Moreover, spatially specific feedback might reveal optimized
strategies in the individual and in general for SCP self-regula-
In general, subjects might be trained by fMRI feedback to
control local brain activity. This offers the opportunity to study
behavior, e.g., mental strategies, physiological measures, affec-
tive state, dependent on self-regulated activity in circumscribed
brain regions. This approach adds to the current neuroimaging
methodology which usually measures brain activity dependent
The authors would like to thank M. Erb, B. Kardatzki, and
U. Klose for technical support and helpful discussions, and
B. Newport for technical assistance. They would also like to
thank Siemens Medical (Erlangen, Germany) and especially S.
Thesen for their support with sequence programming.
 N. Birbaumer, N. Ghanayim, T. Hinterberger,I. Iversen, B. Kotchoubey,
A. Kubler, J. Perelmouter, E. Taub, and H. Flor, “A spelling device for
the paralyzed,” Nature, vol. 398, pp. 297–298, 1999.
 B. Rockstroh, T. Elbert, N. Birbaumer, and W. Lutzenberger, “Biofeed-
back-produced hemispheric asymmetry of slow cortical potentials and
its behavioral effects,” Int. J. Psychophysiol., vol. 9, pp. 151–165, 1990.
 N. K. Logothetis, “MR-imaging in the nonhuman primate: Studies of
function and of dynamic connectivity,” Curr. Opinion Neurobiol., vol.
13, pp. 630–642, 2003.
 S. S. Yoo and F. A. Jolesz, “Functional MRI for neurofeedback: Feasi-
 S. Posse, D. Fitzgerald, K. Gao, U. Habel, D. Rosenberg, G. J.
Moore, and F. Schneider, “Real-time fMRI of temporolimbic regions
detects amygdala activation during single-trial self-induced sadness,”
NeuroImage, vol. 18, pp. 760–768, 2003.
 N. Weiskopf, R. Veit, M. Erb, K. Mathiak, W. Grodd, R. Goebel,
and N. Birbaumer, “Physiological self-regulation of regional brain
activity using real-time functional magnetic resonance imaging (fMRI):
Methodology and exemplary data,” NeuroImage, vol. 19, pp. 577–586,
 R. C. deCharms, K. Christoff, G. H. Glover, J. M. Pauly, S. Whitfield,
and J. D. E. Gabrieli, “Learned regulation of spatially localized brain
activation using real-time fMRI,” NeuroImage, vol. 21, pp. 436–443,
T: Comparison of oxygenation-sensitive magnetic resonance imaging,”
Magn. Reson. Med., vol. 45, pp. 595–604, 2001.
 R. Deichmann, O. Josephs, C. Hutton, D. R. Corfield, and R. Turner,
“Compensation of susceptibility-induced BOLD sensitivity losses in
echo-planar fMRI imaging,” NeuroImage, vol. 15, pp. 120–135, 2002.
 K. Mathiak, A. Rapp, T. T. J. Kircher, W. Grodd, I. Hertrich, N.
Weiskopf, W. Lutzenberger, and H. Ackermann, “Mismatch responses
to randomized gradient switching noise as reflected by fMRI and
whole-head magnetoencephalography,” Human Brain Map., vol. 16,
pp. 190–195, 2002.
 K. Mathiak and S. Posse, “Evaluation of motion and realignment for
functional magnetic resonance imaging in real time,” Magn. Reson.
Med., vol. 45, pp. 167–171, 2001.
 N. Weiskopf, U. Klose, M. Erb, W. Grodd, N. Birbaumer, and K.
Mathiak. Distortion correction in fMRI using single-shot multi-echo
EPI. presented at 9th Int. Conf. Functional Mapping of the Human
Brain. [CD-ROM] [abstract]. Available in NeuroImage
 P. Jezzard and R. S. Balaban, “Correction for geometric distortion in
echo planar images from B0 field variations,” Magn. Reson. Med., vol.
34, pp. 65–73, 1995.
 M. Lotze, P. Montoya, M. Erb, E. Hülsmann, H. Flor, U. Klose, N.
Birbaumer, and W. Grodd, “Activation of cortical and cerebellar motor
areas during executed and imagined hand movements: An fMRI study,”
J. Cogn. Neurosci., vol. 11, pp. 491–501, 1999.
 K.M. O’CravenandN.Kanwisher,“Mentalimageryoffacesandplaces
activates corresponding stimulus-specific brain regions,” J. Cogn. Neu-
rosci., vol. 12, pp. 1013–1023, 2000.
 R. Goebel, “Cortex-based real-time fMRI,” NeuroImage, vol. 13, p.
 K. J. Worsley and K. J. Friston, “Analysis of fMRI time-series revis-
ited–Again,” NeuroImage, vol. 2, pp. 173–181, 1995.
 S. Thesen, O. Heid, E. Mueller, and L. R. Schad, “Prospective acquisi-
tion correction for head motion with image-based tracking for real-time
fMRI,” Magn. Reson. Med., vol. 44, pp. 457–463, 2000.
H. Flor, and N. Birbaumer, “Brain areas activated in fMRI during self-
regulation of slow cortical potentials (SCP),” Exp. Brain Res., vol. 152,
pp. 113–122, 2003.
Nikolaus Weiskopf studied physics at the University
of Miami, Coral Gables, FL, and University of
Tübingen, Tübingen, Germany. He received the
diploma degree in physics in 2000. He developed
techniques to measure slowly varying magnetic
fields and to analyze epileptic activity at the
Magnetoencephalography Center at the University
of Tübingen. Since 2000, he is working towards the
Ph.D. degree at the Institute of Medical Psychology
and Behavioral Neurobiology and is enrolled in the
Graduate School of Neural and Behavioral Sciences
and the International Max Planck Research School, University of Tübingen.
His research is focused on the development of magnetic resonance imaging
(MRI) techniques, real-time analysis of functional MRI data, and brain-com-
970 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 51, NO. 6, JUNE 2004 Download full-text
Klaus Mathiak studied medicine and mathematics
at the Universities of Berlin, Berlin, Germany. He re-
ceived the diploma in applied mathematics in 1993
and the M.D. degree in 1996. He did his thesis on
medical statistics in 1997 and received the Ph.D. de-
gree and habilitation in behavioral neuroscience in
Since 1998, he is with the Center for Neurology,
University of Tübingen, Tübingen, Germany. He
developed techniques for real-time artefact handling
in fMRI, nonparametric statistics in neuroimaging,
and neurophysiological-based warning systems. His research is focused on the
neurophysiology of the auditory system and the development of neuroimaging
Simon W. Bock received the B.Sc. degree in
physics. He finished medical studies at the Uni-
versity of Tübingen Medical School, Tübingen,
Germany, in 2002. He is enrolled in the Graduate
School of Neural and Behavioral Sciences and
International Max Planck Research School, Uni-
versity of Tübingen. He received medical training
at the University Hospital in Tübingen, at Brown
University, Providence, RI, and at the Universidad
Mayor de San Simón, Cochabamba, Bolivia.
In 2003, he completed research projects at the
Institute of Medical Psychology and Behavioral Neurobiology, University of
Tübingen, and at the Institute of Pharmacology and Toxicology, University of
Zürich, Zürich, Switzerland. Since 2004, he is with the Center for Neurology,
Department of Cognitive Neurology, University of Tübingen, focusing on
awake behaving monkey fMRI and social attention.
Frank Scharnowski received the B.Sc. degree in
cognitive science from the University of Osnabrück,
Osnabrück, Germany, in 2001.
In 2001 and 2002, he was a Research Assistant at
the Max Planck Institute for Biological Cybernetics,
Tübingen, Germany. He did a full-time practical
training at the Institute of Medical Psychology and
Behavioral Neurobiology, University of Tübingen,
in 2003. Currently, he is at the Graduate School
of Neural and Behavioral Sciences and Interna-
tional Max Planck Research School, University of
Mr. Scharnowski received a Certificate of Honour donated by the
Riedel-de-Haen Study Foundation in recognition of extraordinary achieve-
ments in the Cognitive Science course of study.
Ralf Veit studied psychology at the University of
Tübingen, Tübingen, Germany. He received the
diploma in 1992, and the Ph.D. degree in 1997.
From 1992-1997, he worked on psychological
influences on cardiovascular disorders. Since 1998,
he is member of the Institute of Medical Psy-
chology and Behavioral Neurobiology, University
of Tübingen. He is Lecturer in medical psychology.
His research is focused on personality disorders,
emotional regulation, and neurofeedback using
fMRI/EEG and peripheral measures.
Wolfgang Grodd studied biology and medicine at the University of Tübingen,
Tübingen, Germany. He received his diploma in biology in 1976 and the M.D.
degree in medicine in 1984.
He did an internship at the Department of Medical Radiology University of
Tübingen (Chairman: Prof. Dr. W. Frommhold) and was 1984-1985 Postdoc-
toral Fellow at the contrast media laboratory of R.C. Brasch at the Department
of Radiology at the University of California at San Francisco (Chairman: Prof.
Dr. A. Margulis). Since 1995, he is Professor of Neuroradiology and Head of
the section on experimental magnetic resonance of the CNS at the University of
Tübingen. His research interest focuses on proton spectroscopy and functional
imaging of the brain with special emphasis on brain development and cerebellar
Dr.Grodd certified as a radiologist in 1986 and as a neuroradiologist in 1990.
Rainer Goebel studied psychology and computer
science at the University of Marburg. He received
the Ph.D. degree in cognitive psychology from
the Technical University of Braunschweig, Braun-
schweig,Germany, in 1995.
From 1995-1999, he was Postdoctoral Fellow at
furt/Main, Germany, and founded its functional neu-
roimaging group. In 19997/1998, he was a Fellow at
the Institute for Advanced Studies, Berlin, Germany.
Since January 2000, he is a full Professor of Cog-
nitive Neuroscience in the Department of Psychology, Maastricht University,
Maastricht, The Netherlands. He is also board member of the F. C. Donders
interests focus on the visual system including attention and imagery, and ad-
vanced brain imaging methods including cortex-based data analysis and visual-
ization tools and real-time fMRI.
Niels Birbaumer was born 1945. He received the
Ph.D. degrees in biological psychology, art history,
and statistics from the University of Vienna, Vienna,
Austria, in 1969.
In 1975-1993, he was Full Professor of Clin-
ical and Physiological Psychology, University of
Tübingen, Tübingen, Germany. In 1986-1988, he
was Full Professor of Psychology, Pennsylvania
State University, University Park. Since 1993, he
is Professor of Medical Psychology and Behavioral
Neurobiology with the Faculty of Medicine of
the University of Tübingen and Professor of Clinical Psychophysiology,
University of Padova, Padua, Italy. Since 2002, he is Director of the Center
of Cognitive Neuroscience, University of Trento, Trento, Italy. His research
topics include neuronal basis of learning and plasticity; neurophysiology and
psychophysiology of pain; and neuroprosthetics and neurorehabilitation. He ha
authored more than 450 publications in peer-reviewed journals and 12 books.
Among his many awards Dr. Birbaumer has received the Leibniz-Award of
the German Research Society (DFG), the Award for Research in Neuromus-
cular Diseases, Wilhelm-Wundt-Medal of the German Society of Psychology,
and Albert Einstein World Award of Science. He is President of the European
Association of Behavior Therapy, a Fellow of the American Psychological As-
of Science and Literature.