Saul Jaime1,2, Hanbing Lu1, Jose Cavazos2,3, and Yihong Yang1
1Neuroimaging Research Branch, National Institute on Drug Abuse, NIH, Baltimore, MD, United States, 2Physiology, Uni. Texas Health Science Center-San Antonio, San Antonio, TX, United States, 3Neurology, Uni. Texas Health Science Center-San Antonio, San Antonio, TX, United States
Synopsis
Despite the high
clinical and pre-clinical value of fMRI, its success depends on the
identification of the underlying neurophysiological basis of the fMRI BOLD
signal. In previously reported simultaneous electrophysiology and fMRI studies,
experiments were limited by the poor spatial resolution or poor source
localization of intra-cranial or surface EEG techniques employed to record
neural activity. In order to overcome these limitations, we have developed a method that allows
concurrent whole brain fMRI acquisition and 32-channel intra-cerebral
electrophysiological recording in the rat brain. Introduction
Functional magnetic resonance imaging (fMRI)
is a potent noninvasive tool to study brain function in normal and disease
human populations and in animal models of disease. The success of this tool is
based on the understanding of the relationship between neuronal activity and
the fMRI signal, as measured by blood oxygenation level dependent (BOLD)
contrast. Concurrent electrophysiology and fMRI have been previously reported1-6.
Given the noninvasiveness nature, scalp EEG recording has been widely used in
human studies. However, this technique suffers from poor source localization of
neural activity, which is particularly an issue in studying spontaneous ongoing
brain activity. In studies involving intra-cerebral electrophysiological recording,
a limited number of micro-electrodes were typically employed, thus having low
spatial coverage. In order to overcome these limitations, we have developed a
method that allows concurrent whole brain fMRI acquisition and 32-channel
intra-cerebral electrophysiological recording in the rat brain.
Methods and Analysis
Electrode implantation: 250-300g Sprague Dawley rats were
anesthetized and chronically implanted with two 16-channel silicon micro-electrodes
(Neuronexus) into the striatum. Reference and ground wires were soldered to independent
brass screws anchored on the cerebellum. MRI: Experiments were conducted
on a Bruker 9.4T scanner. A single-turn surface coil was placed directly over
the skull encompassing the area of the electrode placement. A T1-weighted
sequence (FLASH; FOV: 35x35mm2) was used for slice localization.
Resting state fMRI data was acquired using a gradient echo-planar imaging (EPI;
FOV: 30x30mm2). Imaging data was processed and analyzed using AFNI
software. Electrophysiological recording: Local field potentials (LFP)
were recorded continuously using a multichannel data acquisition system (Plexon)
and amplified at a gain of 2000, notch filtered at 60 Hz, and digitally sampled
at 10 kHz. Data pre-processing: LFP data was pre-processed and analyzed
in MATLAB and EEGLAB software. MRI-induced artifacts in the LFP traces were
removed in the following steps: each 60-ms segment of artifact data was first
replaced by linear interpolation and then low-passed filtered to 400Hz and
down-sampled to 1kHz. The linear interpolation minimizes ringing effects from
high frequency artifact spikes. The artifact segments were then interpolated by
cubic-spline with data from 35ms before and 35ms immediately after the
artifacts. LFP data were then low-passed filtered to 110Hz and down-sampled to
250Hz. This artifact removal method was evaluated in clean EEG data. Spectra
with and without the artifact removal procedure were essentially identical,
suggesting that this method preserves the spectral content of the original LFP
signal.
Results and Discussion
The
chronic implantation of bilateral silicon electrodes, into the rat striatum, introduced
artifacts which were noticeable in T1-weighted anatomical images
(Fig1A) and in single-shot EPI images (Fig1B). There was apparent signal loss in
tissue surrounding the electrodes. The temporal signal-to-noise ratio (SNR) in
these regions is about 5%, while the temporal SNR in unaffected regions is
about 2%. There were no aberrant fluctuations on the temporal profile of the
time courses within the electrode track (Fig2A) in comparison to normal tissue
(Fig2B). The raw electrophysiological recordings were contaminated with typical
MRI-induced artifacts resulting from the strong alternations of the field
gradients (Fig3A; representative trace from a single channel) during imaging.
The artifacts were visible throughout all 32-channels and were readily removed
in a multi-step method we developed and described above (Fig3B). As can be
seen, the integrity of the recording was preserved after the denoising
procedure.
Conclusion
We
have developed a method that allows concurrent whole brain fMRI acquisition and
32-channel intra-cerebral electrophysiological recording in the rat brain. We
used commercial micro-electrodes and recording system, and did not require
customized software and hardware, thus this method can be readily adapted by
other labs. We have also developed a method in which the MRI-induced electrophysiological
artifacts can be removed. Importantly, the recording sites covered bilateral
striatum ranging from dorsal to ventral domain, and our preparation permitted
repeated experiments from the same animal. The ability of dense
electrophysiological recording coupled with concurrent fMRI opens the
possibility to further explore the neurophysiological basis of the resting
state fMRI signal.
Acknowledgements
This work was supported by
the Intramural Research Program of the National Institute on Drug Abuse, NIH.References
1. Logothetis NK et al.
Neurophysiological investigation of the basis of the fMRI signal. Nature. 2001;
412:150-157.
2. Shmuel A et al. Negative
functional MRI response correlates with decreases in neural activity in monkey
visual area V1. Nature Neuroscience. 2006; 9(4):569-577.
3. Vincent JL et al.
Intrinsic functional architecture in the anaesthetized monkey brain. Nature.
2007; 447:83-88.
4. Scholvinck ML et al.
Neural basis of global resting-state fMRI activity. PNAS. 2010;
107(22):10238-10243.
5. Pan WJ et al. Broadband
local field potentials correlate with spontaneous fluctuations in functional
magnetic resonance imaging signals in the rat somatosensory cortex under
isoflurane anesthesia. Brain Connectivity. 2011; 1(2):119-131.
6. Pan WJ et al. Infraslow
LFP correlates to resting-state fMRI BOLD signals. Neuroimage. 2013; 74:288-297.