Abstracts

Rationale: Transcranial electrical stimulation (TES) has been shown to produce frequency-specific changes in cortical network activity as well as changes in cognition. However the mechanism underlying the observed changes is poorly understood due to the limited spatial resolution of electroencephalogram (EEG) which is often used in conjunction with TES. Stimulating the surface of cortex directly should provide a better control over the region of stimulation in comparison to TES. Also, when the stimulation is carried out in conjunction with electrocorticogram (ECoG), activity and network dynamics can be studied at a more localized scale due to enhanced spatial resolution. Using this approach, the effect of periodic stimulation on cortical activity during two different cognitive states was studied. Methods: This study is funded by the pilot grant of TTSA Phase I, School of Medicine, UNC. In this study, 2 patients undergoing long-term monitoring for surgical resection of epileptic seizure foci at UNC Neuroscience Hospital were stimulated intracranially during two different cognitive states (data from one patient shown here). In the first state, the patients were asked to relax with their eyes closed (resting state) and in the second state they performed a visual working memory task. Bipolar current pulses of 2mA and 200µS duration were applied for 5 seconds at a frequency of 10 Hz and an inter-trial duration of ~40 seconds between subsequent stimulations. During the visual working memory task, the pulse was applied time-locked to individual trials. The stimulation was carried out at 14 different electrode pairs once during the resting state and once during the working memory task. A principal component analysis (PCA) based method was used to remove stimulation artifacts. Trials were divided into 3 epochs - before, during and after stimulation and power spectral densities were computed. Modulation index was calculated to determine how power spectral densities changed during different epochs. Results: Spectral analysis of epochs before application of pulses revealed differences at different frequency bands for the two tasks suggesting that the patients were indeed in two different cognitive states. Electrical stimulation tended to suppress activity in lower frequency bands during the working memory task whereas there was no significant effect of electrical stimulation during the resting state. However stimulation tended to enhance activity in gamma bands during both states. Modulation indices at different frequency bands were significantly different between the two states during stimulation suggesting state dependent effect of stimulation. Effect size as measured by Cohen's d was 0.94 for theta; 0.93 for delta; 0.64 for alpha; 0.47 for beta; and 0.06 for gamma. Paired t-test for significance of difference revealed p <0.01 for theta, delta, alpha and beta. The difference between the states was not significant in gamma frequency band. Conclusions: Rhythmic electrical stimulation of cortex produced discernible changes in oscillatory activity dynamics which importantly depended on cortical state.