Event-related modulation of alpha rhythm explains the auditory P300 evoked response in EEG

Reviewer #1 (Public Review):

This EEG study probes the prediction of a mechanistic account of P300 generation through the presence of underlying (alpha) oscillations with a non-zero mean. In this model, the P300 can be explained by a baseline shift mechanism. That is, the non-zero mean alpha oscillations induce asymmetries in the trial-averaged amplitudes of the EEG signal, and the associated baseline shifts can lead to apparent positive (or negative) deflections as alpha becomes desynchronized at around P300 latency. The present paper examines the predictions of this model in a substantial data set (using the typical P300-generating oddball paradigm and careful analyses). The results show that all predictions are fulfilled: the two electrophysiological events (P300, alpha desynchronization) share a common time course, anatomical sources (from inverse solutions), and covariations with behaviour; plus relate (negatively) in amplitude, while the direction of this relationship is determined by the non-zero-mean deviation of alpha oscillations pre-stimulus (baseline shift index, BSI). This is indicative of a tight link of the P300 with underlying alpha oscillations through a baseline shift account, at least in older adults, and hence that the P300 can be explained in large parts by non-zero mean brain oscillations as they undergo post-stimulus changes.

Reviewer #2 (Public Review):

The authors attempt to show that event-related changes in the alpha band, namely a decrease in alpha power over parieto/occipital areas, explain the P300 during an auditory target detection task. The proposed mechanism by which this happens is a baseline-shift, where ongoing oscillations which have a non-zero mean undergo an event-related modulation in amplitude which then mimics a low frequency event-related potential. In this specific case, it is a negative-mean alpha-band oscillation that decreases in power post-stimulus and thus mimics a positivity over parieto-occipital areas, i.e. the P300. The authors lay out 4 criteria that should hold if indeed alpha modulation generates the P300, which they then go about providing evidence for.

Strengths:
- The authors do go about showing evidence for each prediction rigorously, which is very clearly laid out. In particular, I found the 3rd section connecting resting-state alpha BSI to the P300 quite compelling.
- The study is obviously very well-powered.
- Very well-written and clearly laid out. Also, the EEG analysis is thorough overall, with sensible analysis choices made.
- I also enjoyed the discussion of the literature, albeit with certain strands of P300 research missing.

Weaknesses:
In general, if one were to be trying to show the potential overlap and confound of alpha-related baseline shift and the P300, as something for future researchers to consider in their experimental design and analysis choices, the four predictions hold well enough. However, if one were to assert that the P300 is "generated" via alpha baseline shift, even partially, then the predictions either do not hold, or if they do, they are not sufficient to support that hypothesis. This general issue is to be found throughout the review. I will briefly go through each of the predictions in turn:

1. The matching temporal course of alpha and P300 is not as clear as it could be. Really, for such a strong statement as the P300 being generated by alpha modulation, one would need to show a very tight link between the signals temporally. There are many neural and ocular signals which occur over the course of target detection paradigms: P300, alpha decrease, motor-related beta decrease, the LRP, the CNV, microsaccade rate suppression etc. To specifically go above and beyond this general set of signals and show a tighter link between alpha and P300 requires a deeper comparison. To start, it would be a good idea to show the signals overlapping on the same plot to really get an idea of temporal similarity. Also, with the P300-alpha correlation, how much of this correlation is down to EEG-related issues such as skull thickness, cortical folding, or cognitive issues such as task engagement? One could perhaps find another slow wave ERP, e.g. the Lateralised Readiness Potential, and see if there is a similar strength correlation. If there is not, that would make the P300 relationship stand out.

In Figure 3, it is clear that alpha binning does not account for even 50% of the variance of P300 amplitude. Again, if there is such a tight link between the two signals, one would expect the majority of P300 variance to be accounted for by alpha binning. As an aside, the alpha binning clearly creates the discrepancy in the baseline period, with all alpha hitting an amplitude baseline at approx. 500ms. I wonder if could you NOT, in fact, baseline your slow wave ERP signal, instead using an appropriate high pass filter (see "EEG is better left alone", Arnaud Delorme, 2023) and show that the alpha binning creates the difference in ERP at the baseline which then is reinterpreted as a P300 peak difference after baselining.

2. The topographies are somewhat similar in Figure 4, but not overwhelmingly so. There is a parieto-occipital focus in both, but to support the main thesis, I feel one would want to show an exact focus on the same electrode. Showing a general overlap in spatial distribution is not enough for the main thesis of the paper, referring to the point I make in the first paragraph re Weaknesses. Obviously, the low density montage here is a limitation. Nevertheless, one could use a CSD transform to get more focused topographies (see https://psychophysiology.cpmc.columbia.edu/software/csdtoolbox/), which apparently does still work for lower-density electrode setups (see Kayser and Tenke, 2006).

3. Very nice analysis in Figure 6, probably the most convincing result comparing BSI in steady state to P300, thus at least eliminating task-related confounds.

4. Also a good analysis here, wherein there seem to be similar correlation profiles across P300 and alpha modulation. One analysis that would really nail this down would be a mediation analysis (Baron and Kenny, 1986; https://davidakenny.net/cm/mediate.htm), where one could investigate if e.g. the relationship between P300 amplitude and CERAD score is either entirely or partially mediated by alpha amplitude. One could do this for each of the relationships. To show complete mediation of P300 relationship with a cog task via alpha would be quite strong.

One last point, from the methods it appears that the task was done with eyes closed? That is an extremely important point when considering the potential impact of alpha amplitude modulation on any other EEG component due to the well-known substantial increase in alpha amplitude with eyes closed versus open. I wonder, would we see any of these effects with eyes opened?

Overall, there is a mix here of strengths of claims throughout the paper. For example, the first paragraph of the discussion starts out with "In the current study, we provided comprehensive evidence for the hypothesis that the baseline-shift mechanism (BSM) is accountable for the generation of P300 via the modulation of alpha oscillations." and ends with "Therefore, P300, at least to a certain extent, is generated as a consequence of stimulus-triggered modulation of alpha oscillations with a non-zero mean." In the limitations section, it says the current study speaks for a partial rather than exhausting explanation of the P300's origin. I would agree with the first part of that statement, that it is only partial. I do not agree, however, that it speaks to the ORIGIN of the P300, unless by origin one simply means the set of signals that go to make up the ERP component at the scalp-level (as opposed to neural origin).

Again, I can only make these hopefully helpful criticisms and suggestions because the paper is very clearly written and well analysed. Also, the fact that alpha amplitude modulation potentially confounds with P300 amplitude via baseline shift is a valuable finding.

Author Response

We are grateful for the constructive feedback and the possibility of further improving our manuscript in terms of quality and clarity. Below, we have prepared a brief answer to the points raised in the reviewers’ feedback. We plan to address all these issues fully in the revised version of the manuscript.

We agree that some of our claims were overly enthusiastic. We will rewrite parts of the manuscript to tame our statements. Additionally, we are thankful for the comments on the use of language, which we will certainly apply while editing the manuscript. Below, we focus on the main comments.

Both reviewers: We appreciate advice on possible confounding factors. We should note here that there is substantial evidence on the effects of alpha rhythm amplitude on the excitability of a neuronal network and, as a consequence, on the amplitude of evoked responses (Baumgarten et al., 2016 Cerebral Cortex; Iemi et al., 2017 eLife; Stephani et al., 2021 eLife). This effect is due to changing the gain for evoked responses, and it is quite different compared to the baseline-shift mechanism (BSM). In BSM, the changes in the amplitude of evoked responses occur due to the generation of an additional evoked response component, which we tried to reveal in our current work. Still, we agree with suggestions to test additional factors, such as earlier evoked responses, baseline window, and head size, and we will test those.

Reviewer #2 Comment 2: Certainly, for low-density recordings, some method of data transformation is required. Here we would like to show our reasoning for why we did not use current-source density (CSD) but rather utilised other approaches. First, the CSD transform performs well for spatially localised activities since it is a spatial high-pass filter. In our case, P300 and alpha amplitude dynamics are fairly widespread with low spatial frequency, and we believe we would not benefit from applying CSD. Second, CSD has been shown to be more sensitive to surface sources in the crowns of gyri. For activity in the P300 window, we have no reason to believe that this is the case. Third, as we completely agree that low density montage is a limitation, we used source reconstruction with eLoreta (Fig. 5) to refine the spatial localisation of potential sources of P300 and alpha amplitude change.

Reviewer #1 Comment 4: Our study is indeed based on a sample of older participants. However, in our previous work (Studenova et al., 2022), we compared young and elderly participants using resting-state data. There, we measured the baseline-shift index (BSI). We found that BSIs for elderly participants were lower in comparison to those for young participants. Therefore, despite these limitations, in the current study, we were still able to detect a correspondence between BSIs and evoked responses in elderly participants. Therefore, we believe that for a sample of young participants, the results should not be different.

Reviewer #2 Comment 4: We agree that mediation analysis will provide additional insights, and we will add it to the revised version of the manuscript.

Overall, we found the reviewer's comments very helpful. We will update the manuscript accordingly.