Relating layer fMRI signals to acoustics and intracranial neuronal activity in the human auditory cortex in a naturalistic design
Curation statements for this article:-
Curated by eLife
eLife Assessment
This study presents a valuable finding on linking the frequency of neural activity to cortical depths of blood flow in a naturalistic setting of participants listening to music. The presentation of evidence in the version of the original submission is incomplete, as further clarifications in methods and results, as well as performing additional analyses, would strengthen the study. The work will be of interest to cognitive neuroscientists working on multimodal recordings, auditory perception and music.
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (eLife)
Abstract
Naturalistic auditory perception engages feedforward and feedback dynamics across cortical depths, yet how these are organized in the human auditory cortex has been difficult to verify noninvasively. Here, we examine depth-dependent coupling between neuronal activity and fMRI during passive music listening. Depth-specific fMRI responses were modeled using neuronal oscillation envelopes elicited by the same naturalistic stimuli from a separate group of patients under intracranial EEG monitoring. From deep toward superficial depths, the relationship between oscillatory power and fMRI responses systematically changed: alpha/beta activity (8-30 Hz) was increasingly associated with negative fMRI responses, mapping top-down feedback, while gamma band (>30 Hz) oscillations showed increasingly positive associations. Relative to a purely acoustical fMRI baseline, broadband high-frequency activity (>70 Hz), a proxy for neuronal firing, showed the strongest coupling to BOLD signals at intermediate cortical depths receiving feedforward inputs from earlier auditory pathways. Our findings reveal a spectrolaminar organization of neurovascular coupling in the human auditory cortex.
Article activity feed
-
eLife Assessment
This study presents a valuable finding on linking the frequency of neural activity to cortical depths of blood flow in a naturalistic setting of participants listening to music. The presentation of evidence in the version of the original submission is incomplete, as further clarifications in methods and results, as well as performing additional analyses, would strengthen the study. The work will be of interest to cognitive neuroscientists working on multimodal recordings, auditory perception and music.
-
Reviewer #1 (Public review):
Summary:
In their submitted work, Lee and colleagues examine the correlation between electrophysiological activity as measured by SEEG, and layer-specific activation patterns, as measured through 7T fMRI. This analysis was performed using patients undergoing monitoring for epilepsy surgery guidance, as well as healthy controls, as they both listened to music.
They find that, in general, higher-frequency SEEG activity correlated positively with the fMRI signal, while lower frequencies correlated negatively. Across cortical depth, higher-frequency activity correlated positively with middle-to-upper layers, whereas lower frequencies showed their strongest negative correlations in superficial layers.
Strengths:
This is an interesting physiological study in that, to the best of this reviewer's knowledge, it has …
Reviewer #1 (Public review):
Summary:
In their submitted work, Lee and colleagues examine the correlation between electrophysiological activity as measured by SEEG, and layer-specific activation patterns, as measured through 7T fMRI. This analysis was performed using patients undergoing monitoring for epilepsy surgery guidance, as well as healthy controls, as they both listened to music.
They find that, in general, higher-frequency SEEG activity correlated positively with the fMRI signal, while lower frequencies correlated negatively. Across cortical depth, higher-frequency activity correlated positively with middle-to-upper layers, whereas lower frequencies showed their strongest negative correlations in superficial layers.
Strengths:
This is an interesting physiological study in that, to the best of this reviewer's knowledge, it has not been done before with auditory stimuli using the combination of iEEG (as opposed to scalp EEG) and fMRI. The framework fits well with models of layer-specific feedforward versus feedback processing (e.g., Bastos et al., 2012).
Weaknesses:
Its main limitations are a lack of specificity to the acoustic stimuli, the absence of correction for venous draining, and the fact that it is largely a replication/port of prior work.
-
Reviewer #2 (Public review):
Summary:
The authors present an investigation of the relationship between the iEEG frequency bands signal and hemodynamic responses at different cortical depths. Based on this, the authors aim to uncover the layered origin of iEEG signals at different frequencies. The authors then interpret their results in terms of feedforward and feedback processing, arguing that the correlations between fMRI and iEEG signals reflect the interaction between both processes. In addition, the authors aim to infer the extent to which these processes are involved during naturalistic music processing.
Strengths:
This study combines the neural recording methodologies yielding the highest spatio-temporal precision achievable in humans, while using naturalistic auditory stimuli. This combination of recording methods and …
Reviewer #2 (Public review):
Summary:
The authors present an investigation of the relationship between the iEEG frequency bands signal and hemodynamic responses at different cortical depths. Based on this, the authors aim to uncover the layered origin of iEEG signals at different frequencies. The authors then interpret their results in terms of feedforward and feedback processing, arguing that the correlations between fMRI and iEEG signals reflect the interaction between both processes. In addition, the authors aim to infer the extent to which these processes are involved during naturalistic music processing.
Strengths:
This study combines the neural recording methodologies yielding the highest spatio-temporal precision achievable in humans, while using naturalistic auditory stimuli. This combination of recording methods and experimental design offers key insights regarding the precise origin of iEEG signals, which is necessary to improve the interpretability of future iEEG studies.
Weaknesses:
(1) The current framing of the paper leads the authors to interpret their findings in ways that are not warranted by the data. The main analysis of the paper consists of correlating the hemodynamic responses from different layers with iEEG signals from different frequency bands, which enables us to infer the relationship between the two signals. It does not, however, enable us to draw inferences regarding the extent of feedforward and feedback processing and the interaction between the two during naturalistic auditory processing. This would require comparing hemodynamic responses in different cortical layers or frequency bands activation against some baseline condition. Based on the presented analysis, statements such as "our frequency-specific results demonstrate that naturalistic music perception seamlessly integrates both feedforward and feedback processing streams" should be removed.
(2) The presentation of existing literature omits key details and findings, making it difficult to fully understand the research question the authors are trying to address. For example, the author mentions studies showing that feed-forward and feedback processing are segregated across cortical layers and that feed-forward and feedback processing have distinct time-frequency signatures (lines 47-58). However, the authors do not mention which cortical layer or which frequency band is associated with which kind of processing. As a result, it is difficult for the reader to determine what exact hypothesis the author is trying to test in the study. This might also relate to the confusion raised in (1).
(3) The method section omits key details. When describing the paradigm, the authors do not describe how the tones were presented, nor how the signals were synchronized. Similarly, there is no mention of the pipeline used for iEEG electrodes localization. In addition, the exact regressors that entered the generalized linear model of hemodynamic responses are not clearly stated: were all regressors (frequency bands + acoustic signal + HFA) entered together in a single model or in separate models? The mention of a cubic spline is also not sufficient for the reader to understand what was done and for which purpose. Finally, the exact tests used for some comparisons are omitted (in Figure 2, for example, no mention of the exact test used to compare betas between A1 and A2). The current structure of the method section is also quite difficult to follow: the authors switch back and forth between describing acquisition protocols and participant counts, for example.
(4) The lack of methodological details (as described in point 3 above) casts doubts about the validity of some of the statistical tests reported. Throughout the paper, the authors present quantitative statements and statistical tests comparing the fitted beta parameters between brain regions (A1 and A2) and cortical depths. However, the authors do not mention any normalization procedure taken to ensure that the scale of the signals being compared was equated. If the overall magnitude of the signals in A1 differs from that of A2, the mean of the beta distribution is expected to differ as well. Similarly, if the signal-to-noise ratio differs between brain regions or cortical layers, so should the variance of the beta parameters across subjects, which might break the homoscedasticity assumption of some tests, which might or might not be a problem depending on the exact test the authors used (hence the importance of reporting them).
-
-
-
