Rationale
fMRI signals are normally related to cerebral blood flow (CBF) metabolism and neuronal activity. During absence seizures, patients experience brief episodes of staring, accompanied by spike-wave discharges (SWD) on electroencephalography (EEG). Neuroimaging studies of absence seizures have yielded contradictory results. Some have shown increased CBF or metabolism during SWD, while others have shown the opposite. Limited spatio-temporal resolution, and varying measurement conditions may be important causes of this variability. Therefore, the purpose of this study was to directly measure neuronal activity and neuroimaging signals at high spatio-temporal resolution during SWD through multi-modal recordings in WAG/Rij rats, an established animal model of human absence seizures. Establishing the relationship between neuronal activity and neuroimaging signals in SWD has significant practical applications for localizing and treating brain regions responsible for seizure generation.
Methods
fMRI measurements were performed in a 7 T or 9. 4 T horizontal bore spectrometer in WAG/Rij rats under fentanyl/haloperidol anesthesia and neuromuscular blockade, during spontaneous SWDs. Simultaneous EEG recording was performed with carbon filament electrodes to determine the timing of seizures, and artifacts were removed. fMRI signals were then analyzed by comparing images acquired during seizures to baseline images. Results were superimposed on high-resolution anatomical images in the same coronal plane. We also measured physiological changes in brain regions identified to show fMRI signal changes during SWD, using a custom-built combination probe. The probe simultaneously recorded CBF using tissue laser Doppler flow (LDF), and extracellular multiunit activity at high time resolution from the same region during SWD and at baseline.
Results
Comparison of ictal and interictal epochs revealed focal bilateral increases in fMRI signal in both cortical and subcortical structures during SWD. We found that fMRI increases were localized mainly to primary somatosensory cortex and to specific thalamic and brainstem nuclei (n=12 WAG/Rij rats studied). The occipital cortex was spared. Electrophysiology mapping on 35 WAG/Rij rats, demonstrated strong correspondence between fMRI and electrophysiology for regions involved and spared during SWD. Combined LDF and electrophysiology studies were then performed, concentrating on the primary somatosensory (S1BF) and primary visual (V1 M) cortex (respectively, involved and spared on fMRI). We found that in S1BF, each SWD produced a transient increase in the rate of neuronal firing, and with a delay of 2–3 seconds, a transient increase in CBF (n=12 WAG/Rij rats). V1 M showed no significant changes in neuronal firing rate or CBF.
Conclusions
Multi-modal high spatio-temporal measurements in the same system demonstrated that fMRI signals, CBF and neuronal firing all increased in regions intensely involved in SWD and did not change in regions spared by seizures. fMRI and CBF changes corresponded closely in space and time to changes in neuronal activity. These findings demonstrate that fMRI can be used to accurately map regional changes in brain function during seizures, which may help the development of improved therapies targeted at the involved regions.