Abstracts

Rationale: Infantile spasms is a severe epileptic syndrome of early childhood. The epileptic spasms (ES) are brief stiffening of arms and legs lasting a few seconds and tend to occur in clusters. The neurophysiological mechanisms underlying infantile spasms are unknown. Our lab has developed a rat model of infantile spasms induced by chronic infusion of tetrodotoxin (TTX) into the neocortex on postnatal day 11 or 12. The ES recapitulate the electrophysiological characteristics of human condition including ictal slow wave/decremental events and interictal hypsarrhythmic-like activity. The animals respond to the FDA approved drugs, ACTH and vigabatrin. Clinical observations have been unable to definitively determine where spasms are generated. We hypothesized that spasms originate in the neocortex. Methods: Linear 16 channel microelectrode arrays were implanted into the somatosensory cortex of rats to determine the distribution across the cortical laminae of the neuronal activity that occur at the onset of ES. Three concurrent neurophysiological events were analyzed: the slow wave at the onset of ictal events, coincident pathological high frequency oscillations (pHFOs) and multiunit activity (MUA). We used current source density (CSD) analysis to locate the local generator of the initiating slow wave. In other experiments we used simultaneous microwire recordings to compare unit activity from the somatosensory cortex and ventrobasal nucleus of thalamus at spasm onset. Results: In 9 animals, CSD analysis revealed large current sinks for the ictal event slow wave in neocortical layers V/VI. Analysis of MUA supports this finding since unit firing at spasm onset was 3-fold higher in infragranular versus supragranular layers. Surprisingly, we routinely recorded a dramatic pause in MUA firing immediately preceding spasm onset. Since this pause was very reminiscent of the down state of neuronal firing during slow wave sleep (SWS), we undertook an analysis of this sleep state. In both epileptic and control animals, we recorded alternating clustering (upstate) and silencing (downstate) of unit firing during SWS. However, the pHFOs and MUA spikes were clustered at the peak of the LFP slow wave in epileptic animals, while they were more widely distributed across the phase of slow waves in controls. In addition, the power of pHFOs was significantly higher in the upstate in epileptic animals than in their controls. This leads us to believe that epileptic spasms are triggered by an intense upstate. Simultaneous MUA recordings in the cortex and the ventrobasal nucleus of thalamus at the onset of ictal events showed that on average MUA firing in the cortex preceded that of the thalamus by 60 ms. Moreover, ictal events in the neocortex preceded the onset of behavior spasms by 230 ms suggesting that the neocortical ictal discharges likely initiate behavioral spasms. Conclusions: Our results suggest that the neocortex contributes in critical ways to the generation of epileptic spasms in this animal model and the spasm generators are abnormal microcircuits in infragranular layers V and VI.  Funding: NIH-NINDS