Ketamine induces multiple individually distinct whole-brain functional connectivity signatures

Article activity feed

1. 

   Version published to 10.7554/elife.84173 on eLife

   Apr 17, 2024
2. 

   eLife

   Dec 17, 2023

   eLife assessment

   This study presents valuable findings regarding inter-individual variability in the neural and behavioral effects of ketamine. The methodological approach used to characterize this variability is compelling, but the evidence to support the specificity of the changes and their genetic correlates is incomplete. The study would benefit from a more thorough examination of the specificity of the pharmacological and genetic results.

   Read the original source
3. 

   eLife

   Dec 17, 2023

   Reviewer #1 (Public Review):

   In this work, 40 healthy volunteers underwent a placebo followed by a ketamine infusion during a resting state fMRI scan. The authors use principal components analysis (PCA) of the difference in global brain connectivity (GBC) between the ketamine and placebo infusions as their summary neural measure. First, a GBC map is computed after processing with the HCP minimal pipeline and removal of the global brain signal for each scan (~4.5 min, TR=700ms). Then the significant PCA components of difference between ketamine and placebo GBC maps are taken as the neural effect of interest and compared to the mean delta GBC. The first two principal components account for 24.5% of the variance of the data and had correlations with SST and PVALB cortical gene expression patterns that were above chance. No significant …

   Reviewer #1 (Public Review):

   In this work, 40 healthy volunteers underwent a placebo followed by a ketamine infusion during a resting state fMRI scan. The authors use principal components analysis (PCA) of the difference in global brain connectivity (GBC) between the ketamine and placebo infusions as their summary neural measure. First, a GBC map is computed after processing with the HCP minimal pipeline and removal of the global brain signal for each scan (~4.5 min, TR=700ms). Then the significant PCA components of difference between ketamine and placebo GBC maps are taken as the neural effect of interest and compared to the mean delta GBC. The first two principal components account for 24.5% of the variance of the data and had correlations with SST and PVALB cortical gene expression patterns that were above chance. No significant correlations were found between mean change in GBC and these genes. Additionally, in comparison with the mean GBC the PCs were found to better correlate with behavioral measures.

   To further support their aim to establish the multi-dimensionality of the ketamine response using their neural measure, the PCA dimensionality was estimated in external datasets that used psilocybin and LSD with sample size matching using identical processing and found lower dimensionality in these datasets.

   Effective dimensionality was calculated using the participation ratio and dataset re-sampling was used to control for sample size in this calculation, but dimensionality is also affected by motion within the sample among other noise sources, which are not well discussed. In particular, each drug may affect physiological noise in different ways and this may in turn affect their dimensionality measurement.

   A PCA decomposition of the changes (ketamine-placebo) in 31 measured behavioral variables was also performed which resulted in two major PCs which accounted for 41.4% of the variance. Following prior work, behavioral PCs were mapped onto the neural PCs to create neuro-behavioral PCs. The weighing of the PCs at the individual level was explored to compare inter-individual variability.

   In an earlier fMRI study of the timeseries response to ketamine (De Simoni, 2013) it was clear that there are both individual and regional brain response differences. Behaviorally, there is known individual variability in the response to ketamine insofar as only about 60% of depressed people will experience symptom improvement and even then to varying levels. Thus, it is good to see that the compound summary measure of the PCA of the change in GBC after ketamine follows this pattern and shows inter-individual differences.

   A strength of this paper is that it brings together multimodal and external datasets and combines them in a linked analysis to support their investigational aims. Several sets of analyses are used to draw relations between fMRI results, genetics, and behavioral measures but the range of conclusions is limited by the understandably small sample size for this kind of drug challenge study. A weakness is that the chosen summary measure (delta GBC of ketamine-placebo, followed by a group-level PCA) that has been principally developed by this lab and has not seen wide replication. The presentation of the analyses could be simplified to increase readability and impact. Nevertheless, this study provides informative steps toward the development of markers for individualized drug response.

   Read the original source
4. 

   eLife

   Dec 17, 2023

   Reviewer #2 (Public Review):

   In this interesting work on the neuropharmacological effects of ketamine, the authors conducted a pharmacological functional magnetic resonance imaging (fMRI) study in 40 healthy participants receiving bolus and constant infusion of ketamine during resting-state fMRI. Data were preprocessed with the human connectome-based standard pipeline previously successfully used by the lab (FS parcellation and application of an atlas published by the group, HCP pipeline, FSL, global brain connectivity with and without global-signal regression). Briefly, GBC and principle component maps of the positive and negative syndrome scale (PANSS) were related to somatostatin and parvalbumin cortical gene expression patterns. In addition, the authors compared the effective dimensionality, i.e. eigenvalues of covariance matrices …

   Reviewer #2 (Public Review):

   In this interesting work on the neuropharmacological effects of ketamine, the authors conducted a pharmacological functional magnetic resonance imaging (fMRI) study in 40 healthy participants receiving bolus and constant infusion of ketamine during resting-state fMRI. Data were preprocessed with the human connectome-based standard pipeline previously successfully used by the lab (FS parcellation and application of an atlas published by the group, HCP pipeline, FSL, global brain connectivity with and without global-signal regression). Briefly, GBC and principle component maps of the positive and negative syndrome scale (PANSS) were related to somatostatin and parvalbumin cortical gene expression patterns. In addition, the authors compared the effective dimensionality, i.e. eigenvalues of covariance matrices of drug vs. placebo, and found higher complexity of responses in ketamine vs. LSD and psilocybin, which is very interesting. Also, there was substantial inter-individual variation in behavioral and neurobehavioral results, which was captured by PC and GBC maps. In supplementary results, the authors also showed that the principle component PS1 highly correlated with the fMRI global signal.

   Although a complex set of analyses is presented, the paper is written very clearly and understandable. The authors did a good job of outlining the steps of their analyses in supplemental diagrams and the source code is provided. As a general remark, I consider the main strength of this work, to acknowledge the very diverse inter-individual variation of ketamine's effects and to use advanced methodological approaches to disentangle these.

   Since the drug also exhibits strong variation in clinical antidepressant responses, the methodology applied here will very likely yield interesting results applied in clinical datasets of patients with major depressive disorder.

   Read the original source
5. 

   Version published to 10.1101/2022.11.01.514692v1 on bioRxiv

   Nov 1, 2022