Abstracts

Rationale: Absence seizures are associated with spike-and-wave discharge (SWD) on EEG, and often cause transient impairment of consciousness. There have been many efforts to pinpoint the underlying neuronal mechanisms behind the impairment of consciousness. Previous fMRI studies have correlated behavioral impairment with BOLD amplitude across the entire brain. However, interpretation of these fMRI changes is limited by lack of knowledge of the seizure-related hemodynamic response and the effect of seizure duration, as well as a lack of plausible neural activity models. It has been shown that a simple boxcar model of neural activity during SWD is inadequate. Convolving a canonical hemodynamic response function (cHRF) with a box car model of SWD duration does not match the observed fMRI changes in most brain regions. Our goal was to develop a more realistic model of neuronal activity during SWD to better explain these changes.  Methods: Over the past decade, we recorded BOLD fMRI signals during over 1000 absence seizures with simultaneous EEG in 39 human patients. Using the human data in conjunction with physiological recordings from animal studies, we created a physiologically realistic model of the electrical activity before/during/after the SWD. We then used the cHRF to fit the model to the BOLD activity in all voxels of the brain for seizures of varying duration.  Results: The model revealed distinct patterns in each networks of the brain and a close relationship between the neural activity amplitudes and the seizure duration. The default mode network (DMN) and task positive network (TPN) show a strikingly early background increase in activity as early as one minute before seizure onset, and a sustained large decrease that persists for >20s after seizure offset. The thalamus shows minimal activity before seizure onset, and a sustained increase in activity during the SWD. We also found a positive relationship between the seizure duration and amplitude of activity during seizures for the thalamus. For DMN and TPN (but not the thalamus), activity before/after seizures was also positively related to seizure duration. Conclusions: The model suggests a complex sequence of activity before and after SWD that may play a role in the physiological severity of the seizures. Long before and after EEG-based seizure duration, abnormal neuronal activity may appear in widespread cortical networks, particularly for longer-lasting seizures. In contrast, abnormal activity in the thalamus may be confined mainly to the time period in which spike-wave discharges appear on EEG. These findings shed new light on potential mechanisms that determine abnormal brain activity patterns associated with absence seizures, which could help develop improved treatments in the future. Funding: This work was funded by NIH RO1 NS055829, CTSA UL1 TR000142,F31 NS077540, T32 GM007205, the Loughridge Williams Foundation, and the Betsy and Jonathan Blattmachr Family.