(J-381) Cortico-cortical evoked potentials suggest focal bipolar stimulation is not as localized as monopolar stimulation

Introduction::

    Cortico-cortical evoked potentials (CCEPs) elicited by single pulse electrical stimulation are used to investigate brain connectivity. CCEPs can be invoked with either monopolar or bipolar stimulation. Bipolar stimulation is hypothesized to elicit a more focal response, however how electrode configuration affects CCEP amplitudes and spread has not been well characterized.

    The hypothesized mechanism of a more focal response during bipolar stimulation is current flowing between two adjacent electrodes rather than spreading out to a distant ground, as with monopolar stimulation. We hypothesized that bipolar stimulation’s focal activation of tissue results in smaller amplitude CCEPs, with smaller local neural activity, fewer number of CCEPs, all clustered closer to the stimulating electrodes as compared to monopolar stimulation.

CCEP amplitudes have been characterized by three features: early peak between 10-50ms representing monosynaptic activity called A1, a second peak between 50-250ms representing polysynaptic activity called A2, and overall area under the curve between 10-250ms. Meanwhile, high gamma power has been correlated with local neural activity. CCEPs represent effective connectivity of the ongoing dynamical brain network, and thus can be measured with outdegree, the number of CCEPs identified after stimulating an electrode. Lastly, the spread of CCEPs from stimulation can be quantified by distance between the stimulating and recording electrodes.

Materials and Methods::

    Bipolar and Monopolar CCEPs were recorded from 21 patients undergoing stereoelectroencephalography for long-term epilepsy monitoring. 7.5mA, 500µs per phase, 50µs interphase interval, randomized 2.5-3.5s intertrial interval stimulations in varying brain regions. CCEPs were identified as per Kundu et al 2020. CCEPs recorded from electrode contacts with high interictal epileptiform discharges, or located in white matter or in seizure onset zones, were removed. The evoked responses were common median re-referenced and analyzed for changes in amplitude, in high gamma power, in degree, and spread.

    For an electrode pair, CCEPs were compared if they were evoked by both bipolar stimulation and a monopolar stimulation. If two monopolar-evoked CCEPs were available, the max of the following metrics was used. Amplitude metrics were calculated as the mean absolute value for early (A1; 10-50 ms) and late peaks (A2; 50-250ms) and the area under the absolute value of the time series (AUC; 10-250 ms). High gamma power (70-150 Hz) was extracted from a fourth order Butterworth filtered, Hilbert transformed, multiplied by the complex conjugate, averaged across trials and across time, 5-100 ms poststimulation data.

   Outdegree was calculated as the number of CCEPs. For Monopolar, the max between the two electrodes was taken, for Bipolar the number of CCEPs, and for Monopolar-Monopolar, the unique set of CCEPs. Spread, or the Euclidean distance between, the average of the stimulating electrode(s), and the recording electrode, was calculated for CCEPs evoked by Bipolar or Monopolar stimulation and CCEPs evoked exclusively by either Bipolar stimulation or by Monopolar stimulation.

Results, Conclusions, and Discussions:: At 7.5 mA stimulation, averaged across 2525 stimulation-response sites, Bipolar CCEPs were larger and more variable than Monopolar CCEPs (Figure 1). Bipolar-evoked CCEPs increased 3% in A1 amplitude (Kruskal-Wallis p < 0.001), increased 82% in A2 amplitude, and increased 70% in AUC (Figure 2; Wilcoxon-rank sum p < 0.001 for each). There were 32% more CCEPs evoked by Bipolar stimulation than monopolar stimulation (n = 124 pairs of stimulating electrodes, Wilcoxon-rank sum p< 0.01) and 13% more than unique responses of two monopolar stimulations (Figure 3; n = 124 pairs of stimulating electrodes, Wilcoxon-rank sum p < 0.05). High gamma power was not significantly different (Figure 4; Wilcoxon-rank sum p > 0.05). CCEPs evoked by eitherr configurations were not spread significantly differently, but CCEPs evoked exclusively by bipolar stimulation were 17% closer to stimulated electrodes than those exclusively by monopolar stimulation (Figure 5A; Figure 5B, n­_{Bipolar CCEPs} = 1686, n_{Monopolar CCEPs} = 590, Wilcoxon rank sum p < 0.01);

    CCEPs are useful for measuring the brain’s effective connectivity, but are sensitive to electrode configuration. We controlled for factors affecting CCEPs by removing responses in white matter tissue, in seizure onset zone, or containing large interictal epileptiform discharges. As a result, we find that Bipolar stimulation elicits larger amplitudes, greater degrees, and that CCEPs elicited from bipolar stimulation are closer to the stimulating electrode than those elicited by Monopolar stimulation.
    We hypothesized incorrectly that Bipolar stimulation would more focally activate tissue when compared to monopolar stimulation. We could be wrong because bipolar stimulation leads to bigger artifacts or that Monopolar stimulations resulted in relatively noisier channels, both of which misidentify CCEPs. Or the impedance between two adjacent electrodes is larger than the impedance between an electrode and a distant, low impedance ground - resulting in larger current densities, activating a larger volume of tissue, accounting for the higher amplitudes and degree.
    Future work can investigate the impact of effect of the volume of tissue activated using biophysical models and density of tissue activated with in-vitro neural tissue models.

    In conclusion, these results may change stimulation protocols for quantifying brain dynamics.

Acknowledgements (Optional): :

References (Optional): : Kundu B, Davis TS, Philip B, Smith EH, Arain A, Peters A, Newman B, Butson CR, Rolston JD. A systematic exploration of parameters affecting evoked intracranial potentials in patients with epilepsy. Brain Stimul. 2020 Sep-Oct;13(5):1232-1244. doi: 10.1016/j.brs.2020.06.002. Epub 2020 Jun 3. PMID: 32504827; PMCID: PMC7494632.