(a)Background
One of the greatest challenges in epileptology is the identification and localisation of those brain areas responsible for eliciting and maintaining epileptiform activity. Furthermore, it is also important to ascertain whether there is any dynamic interaction between those different brain areas. Research has, to date, focused primarily on EEG. Although EEG has high temporal resolution, it lacks spatial resolution due to the inverse problem of dipole source localization. On the other hand, fMRI has great spatial resolution but a relatively poor temporal resolution due to the delayed haemodynamic response that we image, the so called BOLD (Blood Oxygenation Level Dependent) signal. To take advantage of the temporal resolution of EEG and spatial resolution of fMRI it would be desirable to combine both methods in a single technique. Even though several experiments have been reported in recent years using simultaneous acquisition of EEG and fMRI in humans, very few experiments have endeavoured to combine these techniques in animal studies. Here we describe our results using simultaneous acquisition of EEG and fMRI to study the picrotoxin model of epilepsy.
(b)Methods
Male adult Wistar rats (330 g) were anaesthetized with urethane (140 mg/kg, i.p.) and carbon fibre EEG surface electrodes were placed over the skin of both cerebral hemispheres. The signals from the electrodes were amplified inside the bore of the fMRI scanner by a x2000 pre-amplifier, placed 25 cm away from the head of the rat to decrease “magnetic gradient interference”. The signals were then taken out of the magnet via optical fibre cables, reconverted into electrical signals, conditioned (CyberAmp380-Axon Instruments; amplification=200; high-pass filter=0.15 Hz; low-pass filter=60 Hz; Notch filter 50 Hz) and then converted to digital data (MP100-Biopac Systems; sampling rate=500 spl/s). Two raw signal channels were recorded and two digitally filtered channels were calculated in real time (low-pass filter at 60 Hz with Q=1 + band-stop filters in 38, 76 and 114 Hz). Physiological monitoring (respiration and temperature) was performed throughout the experiment. Dynamic MR images were acquired using a 4.7 T imaging system (Varian; Gradient-echo; TR=1050 ms; TE=10 ms; 40 slices per volume affording one whole brain scan per minute for a total scan duration of 45 mins.). Picrotoxin injection (8 mg/kg, i.p.) was remotely performed after acquisition of the first 10 volumes, without interruption of imaging or EEG recording.
(c)Results
we were able to acquire good quality EEG and fMRI data from all subjects. During the experiments all animals presented a typical evolution of the EEG morphology: electrodecremental response followed by isolated spikes, poli-spikes and spike-and-waves, culminating in status epilepticus. The MR images acquired showed a robust negative BOLD response (global minimum) in the caudate and nucleus accumbens and a positive response (global maxima) in the amygdale.
(d)Conclusion
The methodology described above is suitable for simultaneous acquisition of EEG and fMRI in anaesthetized animals. The analysis of the images acquired during the electrodecremental response showed robust negative BOLD in the caudate and nucleus accumbens. The negative BOLD signal change observed during the electrodecremental response suggests a correlation between amplitude variation of the EEG and the BOLD contrast.