PowerPointPrsentation

1
Berlin Neuroscience Forum 2004 April 22-24
Liebenwalde
High-frequency oscillations in human
somatosensory evoked potentials during
simultaneous EEG and fMRI recordings Frank
Freyer, Matthias Moosmann, Petra Ritter, Gabriel
Curio and Arno Villringer Berlin NeuroImaging
Center, Charité, Humboldt University Berlin,
Germany
Introduction
Discussion
For further analysis the recording channel with
the highest HFO peak-to-peak amplitude was chosen
(CP5 in both subjects). In order to isolate HFOs
during scan acquisition, the artifact containing
EEG-periods (1100ms per TR) were eliminated
prior to segmentation and averaging.
Functional magnetic resonance imaging (fMRI)
based on the blood oxygenation level dependent
(BOLD) signal is widely used for neuroimaging.
However, the exact relationship between the
measured fMRI signal and the underlying neuronal
activity is still unclear. Recently simultaneous
acquisition of fMRI and EEG became feasible
allowing for the additional recording of direct
neuronal signals. Although EEG mainly represents
slow postsynaptic neuronal activity, a high
frequency (600Hz) component of the
somatosensory-evoked potential (SEP) superimposed
on the primary cortical response N20 has been
shown recently, supposedly indexing bursting
action potentials or firing of neurons (Curio
et al., 1994).  Recording of these high-frequency
oscillations (HFO) as a third modality additional
to the low frequency SEP and the BOLD signal
would provide important additional information
regarding the relationship between neuronal
activity and the BOLD signal. However,
considering the low amplitude (0.15µV) of HFOs
and the contamination of the EEG-signal by MR
induced artifacts-, it is questionable whether
HFOs are detectable during simultaneous EEG-fMRI
acquisition. Here we addressed this question
using a simultaneous EEG-fMRI setup.
For the first time, we report the possibility for
non-invasively monitoring indexes of neuronal
synaptic activity as well as action potentials
during fMRI, opening a new approach to the
investigation of brain function. Functional
dissociation of low frequency SEP component N20
and the superimposed HFOs can be induced by
variation of stimulus intensity (Klostermann et
al., 1998) and frequency (Klostermann et al.,
1999). In addition, an inverse relationship of
N20 and HFOs during the sleep-wake cycle has been
shown (Hashimoto et al., 1996). As a next step,
these dissociating paradigms could be applied to
a combined EEG-fMRI setup in order to search for
correlations of BOLD response with SEP and HFOs,
respectively. Concurrent acquisition of HFOs as
an additional modality during simultaneous
EEG-fMRI recordings will help to further clarify
the relationship between neuronal activity and
the measured BOLD response.
Results
HFO superimposed on the N20 component were
clearly visible in both subjects in the static
magnetic field data as well as in the
artefact-free non-acquisition periods recorded
with simultaneous fMRI. Single subject data are
shown in Fig. 1.
Abbreviations
BOLD blood oxygen level dependent EEG
electroencephalography fMRI functional magnetic
resonance imaging HFO high-frequency
oscillation SEP somatosensory evoked
potential TR repetition time TA acquisition time
Methods
Two healthy right-handed subjects (one male, one
female, 25 years) participated in the experiment.
Subjects were lying in the static field of the
MR-scanner, while SEPs were evoked by
conventional electrical stimulation of the right
median nerve (repetition rate 8,1 Hz, 50 above
motor threshold). Subsequently, during ongoing
stimulation, 200 fMRI scans (1.5 T, TR 4400ms,
TA 1100ms, 10 slices) were acquired.
Concurrently, continuous EEG (32-channel,
Brainproducts, reference electrode between Cz and
Fz, digital sampling rate 5000Hz) was recorded
during the entire experiment. SEPs were sampled
over a 80ms (20ms pre-, 60ms post-) period and
averaged over 8000 trials in order to obtain an
optimal signal-to-noise-ratio. Offline digital
bandpass filtering was performed for isolation of
low- and high- frequency SEP-components
(0.53-400Hz and 400-700Hz, respectively).
References
Curio G et al. Electroencephalogr Clin
Neurophysiol. 91, 483-7 (1994) Hashimoto I,
Mashiko T and Imada T Electroencephalogr Clin
Neurophysiol. 100, 189-203 (1996) Klostermann F,
Nolte G, Losch F and Curio G Neurosci Lett 256,
101-04 (1998) Klostermann F, Nolte G and Curio G
Neuroreport 10, 1625-29 (1999)
Contact
Fig.1 SEP during simultaneous fMRI recording in a
single subject. top row low-frequency-SEP, with
indicated N20 bottom row corresponding HFOs,
with indicated peak-to-peak amplitude range (max
0.159µV, min -0.158µV) Note the different
magnitude of the HFO compared to the
low-frequency SEP.
Frank Freyer Berlin-Neuroimaging
Center, Schumannstraße 20/21, 10098 Berlin Email
f_freyer_at_web.de