Human hippocampal responses to network intracranial stimulation vary with theta phase

  1. Sarah M. Lurie
  2. James E. Kragel
  3. Stephan U. Schuele
  4. Joel L. Voss

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This project has substantial potential for better explaining the physiological basis of how to best use electrical stimulation on the cortical surface to modulate the hippocampal memory system. This would be an important task translationally and practically because it could lead to methods for modulating activity in deep brain structures noninvasively. However, in its current form the paper has weaknesses that make the results hard to trust and interpret. In its current form, it is not clear if the data clearly support the paper's strong conclusions.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Hippocampal-dependent memory is thought to be supported by distinct connectivity states, with strong input to the hippocampus benefitting encoding and weak input benefitting retrieval. Previous research in rodents suggests that the hippocampal theta oscillation orchestrates the transition between these states, with opposite phase angles predicting minimal versus maximal input. We investigated whether this phase dependence exists in humans using network-targeted intracranial stimulation. Intracranial local field potentials were recorded from individuals with epilepsy undergoing medically necessary stereotactic electroencephalographic recording. In each subject, biphasic bipolar direct electrical stimulation was delivered to lateral temporal sites with demonstrated connectivity to hippocampus. Lateral temporal stimulation evoked ipsilateral hippocampal potentials with distinct early and late components. Using evoked component amplitude to measure functional connectivity, we assessed whether the phase of hippocampal theta predicted relatively high versus low connectivity. We observed an increase in the continuous phase-amplitude relationship selective to the early and late components of the response evoked by lateral temporal stimulation. The maximal difference in these evoked component amplitudes occurred across 180 degrees of separation in the hippocampal theta rhythm; i.e., the greatest difference in component amplitude was observed when stimulation was delivered at theta peak versus trough. The pattern of theta phase dependence observed for hippocampus was not identified for control locations. These findings demonstrate that hippocampal receptivity to input varies with theta phase, suggesting that theta phase reflects connectivity states of human hippocampal networks. These findings confirm a putative mechanism by which neural oscillations modulate human hippocampal function.

Article activity feed

  1. eLife

    Author Response

    Reviewer #1 (Public Review):

    1 - Problems with the analysis of stimulation latency

    The data in this paper show a variable latency in signal propagation from stimulation sites to hippocampal recording electrodes. In an attempt to measure this latency, the authors examine the theta phase offset between each pair of stimulation and recording electrodes (Figure 9). They interpret their results as showing a consistent 90-degree phase offset. However, their data do not support this interpretation because in fact their measurements show a bimodal distribution of phase differences with peaks at 0 and 180 degrees. It is not valid to interpret the circular mean of a bimodal distribution because the result is not well defined. Further, individual electrodes do not show a mean difference of 90 degrees.

    Because the results do not reliably support the claim of a consistent 90 phase difference between the hippocampus and cortex, it is a substantial problem for the paper, given the importance of hippocampal-cortical timing in their interpretation. In particular, the authors should reconsider how they frame their results in relation to the Siegle and Wilson work and others.

    We no longer emphasize the phase difference between hippocampus and neocortex in the revised manuscript. This phase difference was computed to attempt to address the possibility that there was some latency in the propagation of stimulation effects from lateral temporal cortex to hippocampus, which would affect our interpretation of which theta phase angles evoked minimal versus maximal hippocampal response (i.e., “peak” stimulation trials may actually have involved stimulation propagating to hippocampus sometime after its peak). However, as noted above in response to Essential Revisions #1, we cannot fully rule out the possibility that volume conduction influenced our estimates of phase lag. We no longer emphasize this analysis and have moved it to the appendix (Appendix 1-Figure 4), along with a new analysis using bipolar rereferencing to address the volume conduction issue.

    The manuscript is now focused on the main finding of the experiment, of a 180-degree separation between theta phases associated with minimal versus maximal evoked responses. We analyzed this via circular-linear models of phase versus evoked amplitude, as suggested by the reviewers, rather than the phase-binning analyses emphasized in the original manuscript. Circular-linear analyses are indifferent to the specific phase values associated with minimal/maximal response. We have also expanded our Introduction with further discussion of homologies to the rodent literature, including to the Siegle and Wilson paper. Our revised Discussion section emphasizes that the central homology is that there is 180-degree separation between hippocampal theta phase angles associated with minimal versus maximal responsiveness to input, with less emphasis placed on the specific angles (i.e., peak versus trough), given difficulties in comparing specific phase angles across species and recording approaches.

    2 - Problems with the figures

    Some figures in the paper were hard to interpret and I felt it would benefit readers for many to be combined. The results from Figures 3 through 7 would be helpful to see side by side, as they show various investigations of the same data. In Figure 4, it would be helpful to see both plots from (a) on the same axis, as is in (b). I did not find that the accuracy estimation paper in Figure 2 was important to include in the main paper. It would be better suited for the supplement, in my view, unless I am missing something.

    We have substantially revised the figures for clarity. The analyses presented in original Figures 2, 6, and 9 have been moved to the appendix (as revised Appendix Figures 1, 3, and 4). Figure 3 has been combined with Figure 1 into the revised Figure 1. Figures 4 and 7 have been combined in order to show EP data from all four phase bins side-by-side (Figure 3). We did not combine a) and b) from the original figure onto the same axis, as we found it difficult to interpret the four overlaid traces (i.e., 2 EP traces and 2 phase-matched stimulation-free traces). However, these data are now shown side-by-side and on equal axes. We have updated all EP visualizations to improve readability. Figure 5 has been expanded to include component amplitudes comparisons for both peak versus trough and rising versus falling phases, in keeping with the expanded Figure 3.

  2. Version 2 published on bioRxiv
  3. eLife

    Evaluation Summary:

    This project has substantial potential for better explaining the physiological basis of how to best use electrical stimulation on the cortical surface to modulate the hippocampal memory system. This would be an important task translationally and practically because it could lead to methods for modulating activity in deep brain structures noninvasively. However, in its current form the paper has weaknesses that make the results hard to trust and interpret. In its current form, it is not clear if the data clearly support the paper's strong conclusions.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public Review):

    Lurie et al. investigate the effects of stimulation applied to the lateral temporal lobe on the ipsilateral hippocampus with respect to endogenous hippocampal theta phase. They find that the magnitude of evoked hippocampal potentials correlated with hippocampal theta phase at the time of stimulation. The experiments are novel and could be valuable for showing how to use cortical stimulation to modulate the hippocampus. However, I did not find the paper to be suitable in its current form because of the high variance in hipp-cortical stimulation latency between subjects, which is not properly measured. This and other concerns make the findings challenging to interpret.

    1 - Problems with the analysis of stimulation latency
    The data in this paper show a variable latency in signal propagation from stimulation sites to hippocampal recording electrodes. In an attempt to measure this latency, the authors examine the theta phase offset between each pair of stimulation and recording electrodes (Figure 9). They interpret their results as showing a consistent 90-degree phase offset. However, their data do not support this interpretation because in fact their measurements show a bimodal distribution of phase differences with peaks at 0 and 180 degrees. It is not valid to interpret the circular mean of a bimodal distribution because the result is not well defined. Further, individual electrodes do not show a mean difference of 90 degrees.

    Because the results do not reliably support the claim of a consistent 90 phase difference between the hippocampus and cortex, it is a substantial problem for the paper, given the importance of hippocampal-cortical timing in their interpretation. In particular, the authors should reconsider how they frame their results in relation to the Siegle and Wilson work and others.

    2 - Problems with the figures
    Some figures in the paper were hard to interpret and I felt it would benefit readers for many to be combined. The results from Figures 3 through 7 would be helpful to see side by side, as they show various investigations of the same data. In Figure 4, it would be helpful to see both plots from (a) on the same axis, as is in (b). I did not find that the accuracy estimation paper in Figure 2 was important to include in the main paper. It would be better suited for the supplement, in my view, unless I am missing something.

  5. eLife

    Reviewer #2 (Public Review):

    Lurie and colleagues attempted to assess whether electrical inputs to the human hippocampus are affected by theta phase. Building off a long history of rodent and computational research showing that hippocampal theta phase affects a multitude of hippocampal processes, including evoked excitatory potentials, the authors examined this effect in a group of human epilepsy patients. Each patient had recording wires in the hippocampus and stimulating electrodes in the lateral temporal cortex. Subjects were given stimulating pulses at random and then the evoked hippocampal potentials were compared according to which theta phase stimulation occurred on. After accounting for conduction delays between the lateral temporal stim site and the hippocampal recording site, they found that evoked potentials delivered on the failing phase of hippocampal theta had larger evoked responses for both an early (~70-110ms) and late (~120-200ms) components. Similar protocols were tried for evoked potentials in both the amygdala and orbitofrontal cortex, and only the only theta phase-dependent effect was found for early components in amygdalar evoked potentials. The current work is consistent with a large body of rodent research and also adds interesting new wrinkles to how we should consider oscillatory linked neural interactions.

    The data set is large (8 subjects with hundreds of stimulation events each) and is well analyzed. The approach to theta phase estimation is well thought out and consistent with past efforts. The combination of oscillatory synchrony offset based phase lag estimation with stimulation provides a new window in which we should conceptualize neural interactions. While conduction delays are obviously well known, most rodent experiments studying evoked activity rely on internal hippocampal stimulation, so no phase lag. In contrast, most studies of area-to-area theta based communication rely upon hippocampal theta phase and fail to consider any possible conduction delays, which could significantly alter the impact of phase-locked activity in distant areas. This is an important point that neural network simulations need to carefully consider and the current publication provides a blueprint for how to conceptualize (and quantify) these effects.

  6. eLife

    Reviewer #3 (Public Review):

    The authors are interested in understanding how increased input to the hippocampus from entorhinal cortex at particular phases of the theta rhythm alters hippocampal function. Rodent studies have shown memory enhancement from phase specific stimulation of the hippocampus. The authors examine hippocampal responses to intracranial electrical stimulation of the lateral temporal lobe in patients with medically refractory epilepsy. Specifically, they examine the amplitude of the resulting evoked potentials relative to stimulation at different phases of the hippocampal LFP between 3-8 Hz. They are not stimulating entorhinal cortex, but the electrodes they stimulate were determined based on their ability to evoke hippocampal responses in the LFP. They hypothesize that this in-network stimulation should modulate hippocampal responses. The authors find that hippocampal responses are indeed enhanced, but in an opposite pattern to what they expected. Responses are larger when hippocampus stimulation occurs at the peak of the LFP, which they suggest is likely due to a ~90 degree phase difference between the lateral temporal input region and the hippocampus. The effects are small but significant and would be bolstered by 1) controlling for the timing at which stimulation was delivered, 2) showing data for every subject, 3) controlling for re-referencing scheme, and 4) showing results in a richer frequency and phase context. With adequate controls and supportive data, these results will provide an important contribution to understanding how stimulation in concert with the brain's natural rhythms can be used to more effectively modulate brain activity, which is an important step towards understanding how such precise modulation of brain activity can be used to alter memory, perception, and behavior.

  7. Version 1 published on bioRxiv