Abstracts

Rationale: Much remains to be understood about the complex reciprocal relationship of sleep and epilepsy. In particular with respect to focal epilepsies, the reason for increased seizure frequency and likelihood of secondary generalization during sleep in remains unknown. We hypothesized that the alterations of cortical network structure responsible for sleep create an increase in focal cortical excitability and enhanced seizure-propagation tendency. We explored our hypothesis by comparing the pattern of cortico-cortical evoked potentials (CCEPs) during wakefulness and sleep, in two patients undergoing presurgical evaluation with subdural electrodes (SDE). Methods: Two patients with SDE over the left hemisphere were studied. SDE coverage included lateral frontal and temporal, basal and posterior temporal, and medial occipital and parietal areas, for a total of 219 contacts. A total of 23 electrode pairs were stimulated following recording of all seizures and reinstitution of anticonvulsants in two sessions: the first during wakefulness, and the second set during ensuing light sleep later the same day. Each stimulation at an electrode pair comprised 50 charge-balanced pulses at 0.5 Hz frequency, 10 mA current and 500 μs pulse width. CCEPs were readily visualized in the raw running subdural EEG. Formal data processing was carried out offline in MATLAB and results visualized on cortical surfaces constructed with Brainsuite and Curry 7. Results: Widely distributed CCEPs were seen in wakefulness and sleep seen following stimulation of most electrode pairs over the lateral brain surface; responses were less well seen with basal and medial cortex stimulation. There was a trend for CCEPs to be more widely distributed in wakefulness compared to sleep (p = 0.3). There was an obvious visual stereotypy to CCEP morphology across adjacent electrodes in sleep compared to wakefulness. An index of seizure ‘transmissibility' across each electrode contact was computed by convolving adjacent CCEPs in a pair-wise fashion across the long axis of each grid. Considered as a group, transmissibility indices were significantly higher in sleep than in wakefulness for both patients (p < 0.01). Conclusions: By analogy with the theory of linear time-invariant (LTI) systems, CCEPs may be modelled as the ‘impulse response' of the cortex. This analogy, though highly approximate, permits estimation of how disturbances may propagate along a cortical network by the successive convolution of adjacent CCEPs. While a fuller elucidation awaits further work, we report here on the significant stereotypy between adjacent CCEPs in sleep, compared to wakefulness, that yield significantly larger pair-wise convolutions. Evolving oscillatory disturbances (seizures) thus spread more readily from electrode-to-electrode in sleep. More generally, we these preliminary observations are an initial step in understanding how the network architecture of the sleeping brain lends itself to increased vulnerability to seizures in focal epilepsy.