Responsiveness variability during anaesthesia relates to inherent differences in brain structure and function of the fronto-parietal networks

  1. Feng Deng
  2. Nicola Taylor
  3. Adrian M. Owen
  4. Rhodri Cusack
  5. Lorina Naci

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Anaesthesia combined with functional neuroimaging provides a powerful approach for understanding the brain mechanisms of consciousness. Although propofol is used ubiquitously in clinical interventions that reversibly suppress consciousness, it shows large inter-individual variability, and the brain bases of this variability remain poorly understood. We asked whether three networks key to conscious cognition — the dorsal attention (DAN), executive control (ECN), and default mode (DMN) — underlie responsiveness variability under anaesthesia. Healthy participants (N=17) were moderately anaesthetized during narrative understanding and resting state conditions inside the Magnetic Resonance Imaging scanner. A target detection task measured behavioural responsiveness. An independent behavioural study (N=25) qualified the attention demands of narrative understanding. 30% of participants were unaffected in their response times, thus thwarting a key aim of anaesthesia — the suppression of behavioural responsiveness. Individuals with stronger functional connectivity within the DAN and ECN, between them, and to the DMN, and with larger grey matter volume in frontal regions were more resilient to anaesthesia. For the first time, we show that responsiveness variability during propofol anaesthesia relates to inherent differences in brain structure and function of the fronto-parietal networks, which can be predicted prior to sedation. Results highlight novel markers for improving awareness monitoring during clinical anaesthesia.

Key points

  • Moderate propofol anaesthesia showed highly variable effects across individuals

  • 30% of healthy participants’ response times were unaffected by anaesthesia and 70% had significantly delayed, fragmented, or fully omitted responses

  • Grey matter volume in frontal cortex and functional connectivity of the fronto-parietal networks predicted resilience to anaesthesia

Article activity feed

  1. eLife

    Reviewer #3:

    The authors have conducted a very challenging study. The paper is clearly written and the topic of neural function under anesthesia is interesting. However, a significant limitation is that many of the analyses presented here do not provide clear insights into the processes the authors are studying.

    -A key issue is that the authors aim to predict who is more or less sensitive to general anesthesia. However, each individual subject was given a different target plasma concentration of propofol, based on clinical scoring. So any difference in behavior may reflect different dosing rather than different behavioral sensitivity to a particular drug concentration.

    -The interpretation of increased functional connectivity is challenging in the context of anesthesia, which modulates vessel dilation and systemic physiology. These analyses would benefit from additional information about the fMRI signal characteristics, e.g. amplitude and physiological signals.

    -Fig. 3 is used to portray comparisons of wakefulness vs. sedation, implied in the text, but does not include direct statistical tests of the difference between the two conditions, and contrasting p<0.05 with p>0.05 does not indicate a significant difference. The suggestion of reduced cortical responses to auditory stimuli makes sense given that the participants are sedated, but the analysis does not seem to provide information about which aspect of auditory processing is modulated by sedation.

    -The statements about response time not being mediated by age may reflect an underpowered study, as age is a strong modulator of anesthetic sensitivity and one group has an n=6.

    -While many interesting MRI studies can be done with quite small n, depending on the question being asked (e.g. Midnight Scan Club, high-resolution individual studies), this study aims to conduct structure-based predictions of individual differences in behavior. This type of analysis requires more than the n=6 slow responders used for Fig. 5, as there are many other features that likely vary in a group this small. I appreciate that the authors have conducted a very challenging study, and it is not easy to collect more data, but while many interesting analyses can be done on this type of data, this is not an appropriate sample size for assessing GMV-individual differences associations. Larger samples sizes or within-subjects analyses are needed for robust GMV effects.

    -Cluster correction method in 'Analyses of fMRI data' should be specified (and checked, Eklund et al.). The precise statistical method used to assess FDR corrected activity correlations with individual subject response times is not clear; it seems that the ANOVA resulted in non-significant results that are nevertheless being reported as differences using Hedges d?

    -The presented evidence does not sufficiently support the authors' conclusion that they "provided very strong evidence that individual differences in responsiveness under moderate anaesthesia related to inherent differences in brain function and structure within the executive control network, which can be predicted prior to sedation.". I would commend the authors on their interesting and challenging experiment, and recommend refocusing the analyses.

  2. eLife

    Reviewer #2:

    In this study, Deng and colleagues have sought to assess the neural correlates of individual differences in responsiveness variability across wakefulness and moderate levels of propofol-induced anaesthesia. In addition to resting state scanning and an auditory story task, the participants underwent behavioural assessments including memory recall and a target detection task. Furthermore, the auditory story task was independently rated by a separate group for its "suspensefulness". Focusing their analysis on three major large-scale brain networks, the group-level results first indicated significant differences in the between network interactions of the chosen networks across wakefulness and sedation, specifically in the narrative condition. Furthermore, during the same condition, there was reduced cross-subject correlation between wakefulness versus sedation centred mainly on the sensorimotor brain regions. Moreover, based on the responses in the target detection task, the participants were grouped into fast and slow responders which then showed significant differences in gray matter volume as well as connectivity differences in the wakeful auditory story task condition within the executive control network.

    Overall, this is a well-written manuscript. However, my initial enthusiasm about the question of interest was hampered by major theoretical and methodological concerns related to this study. Below I outline these points in the hopes that they improve this study and its outcomes.

    First and foremost, the authors state that their major interest in this study was to assess individual differences in sedation-induced response variability and its potential brain bases. Despite the attractiveness of this topic, which is undoubtedly of interest both to the academic community and the general public, I do not believe that the current study design would allow the authors to answer this question. First of all, although I completely appreciate the difficulty in recruiting participants to take part in such pharmacological studies, I do not think that a group of 17 participants is enough to be able to assess "individual differences". For this to work, there has to be a large enough sample based on adequate power calculations, keeping in mind all the spurious false positive effects that are generated by pharmacological interventions and their downstream effects on connectivity estimates (e.g. motion, global signal etc.). Second, though it is perfectly valid to carry out the initial within-group connectivity and whole-brain activity analyses for the task/rest (which I believe are the only statistically and experimentally sound sections), following these results, the authors mainly carry out multiple exploratory analyses that aim to infer what happened to 3 non-respondent participants (or 6 slow responders). This to me is closer to a case study rather than an experimental study with proper statistics. Overall this fast/slow responder split only comes as an afterthought and does not seem to be the main intention behind the study. This is at odds with the major goal stated in the introduction that the main aim of the authors was to assess inter-individual differences. As such, I do not think that the analyses highlighted by the authors provide enough evidence to support their claims. More detailed points are provided below:

    • The introduction is well-written, citing as much of the relevant literature on this topic as possible. Having said that, I am not really convinced about the justification for selecting the dorsal attention, executive control and default mode networks as the sole focus of the authors' analysis. Although it is true that there is a strong a priori basis that these associative networks play an important role in maintaining consciousness, the references that the authors refer to are equally biased in focusing their analyses on specific higher-order networks, creating a circular argument. In light of evidence highlighting the importance of sensorimotor networks in this context, as well as the balance in their interactions with associative cortices, I would argue that a whole-brain approach would be better suited. Furthermore, as indicated by the whole-brain analysis during the auditory story task, most alterations were centered on the primary somatosensory regions. This is at odds with the justification of the authors on focusing their connectivity analyses solely on associative brain networks.

    • Given the wide age range (and its potential influence on the obtained results), it would be great for the authors to provide the mean and standard deviation of age within groups, and whether the groups were age-matched (though the range seems similar).

    • The authors state that only the reaction time was measured in the auditory target detection task, but later in the results section they mention "omissions". Given that such omissions might be strongly indicative of unresponsiveness/sleep, it is unclear how one can interpret the observed brain-based effects solely from the perspective of reduced information processing (especially when the data was collected under eyes-closed conditions).

    • The authors provide a thorough description of the sedation administration procedures, which is excellent. Nevertheless, I was wondering whether the blood plasma propofol concentrations could be used to explain some of the results in individual differences or even a nuisance regressor to show that the effects were not simply driven by this factor.

    • I failed to find any information in the methods section as to why/how the authors have decided on a mean-split of the participants to fast/slow responders. Given the already small sample size, further reducing degrees of freedom by a split of 11 versus 6 participants makes it very problematic in terms of any statistical tests that can be carried out.

    • Line 441 - Results should not be reported if it did not reach statistical significance.

    • Line 448 - For the two analyses on this page the authors indicate that although in the wakefulness condition there were significant brain activity that correlated with (not "predicted") task stimulus, no significant effects were observed in the sedation condition. This absence of evidence should not be then taken as evidence of absence. In other words, such lack of evidence can be explained by a variety of factors not attributable to the effect of sedation on brain activity (e.g. simply by the fact that the participants were not paying attention to the story or falling asleep).

    • Line 484 - I do not think it is acceptable/justifiable to carry out post-doc tests, when there was no significant difference in the main ANOVA.

    • Line 503 - I am not really sure about the justification behind the assessment of gray matter volume. Besides the issues related to small sample size, the observed differences in functional connectivity may then simply be due to differences in the quality of the data that can be extracted from the defined ROIs in a subset of participants. Was this analysis corrected for age (as a continuous variable)? In any case, as far as I am aware, there is no simple relationship between gray matter volume and functional connectivity (i.e. greater/smaller gray matter volume does not necessarily mean greater/smaller functional connectivity). Hence, once cannot make the conclusion that: "These results lend support to the functional connectivity results above, and together they strongly suggest that connectivity within the ECN, and especially the frontal aspect of this network, underlies individual differences in behavioural responsiveness under moderate anaesthesia."

    • Line 509 - Again, I am not really sure about the justification behind the analysis carried out here. The authors state that the ROIs that were found in the gray matter volume analysis overlapped with a priori ROIs which they suggest explain differences observed in functional connectivity. They then select a subset of these ROIs and again show that there are differences in connectivity. This seems rather circular.

    • The authors state that "Rather, only the functional connectivity within the ECN during the wakeful narrative condition differentiated the participants' responsiveness level, with significantly stronger ECN connectivity in the fast, relative to slow responders." I apologise if I am missing something, but I do not see any evidence for such a strong claim. All that the authors have found was that there were significant functional connectivity differences in the executive control network in the wakefulness condition between fast and slow responders (which was defined and grouped by the authors themselves), with no significant effect of condition or state. I fail to understand why this one result from a multitude of exploratory analyses that were conducted was picked out as the "main finding" when one cannot make any inferences about its direct relation to sedation.

    • Overall, I would urge the authors to re-think their analysis strategy and the corresponding discussion of their results.

  3. eLife

    Reviewer #1:

    Deng et al. studied the mechanisms underlying the wide propofol effect-site concentration range associated with loss of responsiveness. Data was acquired from two centers (MRI, Canada; Auditory, Ireland). This is a well conducted study. The results could also explain why older patients (with presumably smaller gray matter volume) are more sensitive to propofol. My major concerns relate to precision in language.

    1. The authors studied mechanisms underlying why patients lose consciousness at a wide range of propofol effect-site concentration. This behavioral phenomenon is known and well described (Iwakiri H, Nishihara N, Nagata O, Matsukawa T, Ozaki M, Sessler DI. Individual effect-site concentrations of propofol are similar at loss of consciousness and at awakening. Anesth Analg. 2005;100:107-10). I would suggest that the. authors position their paper as such. They did not study general anesthesia per se, and the allusions to awareness under anesthesia may not be relevant.

    2. Per comment 1 above. Please reword the intro and discussion section i.e., " Anaesthesia has been used for over 150 years to reversibly abolish consciousness in clinical medicine, but its effect can vary substantially between individuals." What type of anesthesia are you referring to? Anesthetic vapors? Please provide a reference for this statement or make it propofol specific. Awareness under general anesthesia is related to numerous factors, many of which are iatrogenic as detailed in the NAP 5 study "The incidence of awareness rose from 1 out of 135,000 general anaesthetics to 1 out of 8,200 general anaesthetics when neuromuscular blockers were used" (https://pubmed.ncbi.nlm.nih.gov/25204697/). Further, it is unclear when dreaming occurs (during induction which is reasonable to expect/during emergence which is also reasonable to expect versus during the anesthesia). My suggestion is to qualify your statements by stating that this should be further studied in the context of possible genetic predisposition to awareness (Increased risk of intraoperative awareness in patients with a history of awareness. Anesthesiology 2013;119:1275-83).

    3. The term "moderate anaesthesia" is confusing to me, and would be to most clinicians. Please cite the description of what comprises moderate anesthesia. My interpretation is that the study was about sedation. Did you mean moderate sedation? (https://www.asahq.org/standards-and-guidelines/continuum-of-depth-of-sedation-definition-of-general-anesthesia-and-levels-of-sedationanalgesia).

    4. "the antagonistic relationship between the DMN and the DAN/ECN #and# was reduced during moderate anaesthesia, with a stronger and significant result in the narrative condition relative to the resting state." Anticorrelation?

    5. The suggestion that fMRI can be used to improve the accuracy of awareness monitoring is, in my opinion, not necessary and a stretch.

  4. eLife

    Summary: There was general enthusiasm for the topic of study and the general approach of using neuroimaging to study brain function under anesthesia. However, the reviewers also shared a number of significant concerns, particularly regarding whether the data which has been collected is sufficient to answer the core questions being asked (for example, whether the number of participants supports a robust individual differences analysis).

    Jonathan E Peelle (Washington University in St. Louis) served as the Reviewing Editor.

  5. Version published to 10.1101/2020.06.10.144394 on bioRxiv