Threat of shock increases excitability and connectivity of the intraparietal sulcus

Abstract

Anxiety disorders affect approximately 1 in 5 (18%) Americans within a given 1 year period, placing a substantial burden on the national health care system. Therefore, there is a critical need to understand the neural mechanisms mediating anxiety symptoms. We used unbiased, multimodal, data-driven, whole-brain measures of neural activity (magnetoencephalography) and connectivity (fMRI) to identify the regions of the brain that contribute most prominently to sustained anxiety. We report that a single brain region, the intraparietal sulcus (IPS), shows both elevated neural activity and global brain connectivity during threat. The IPS plays a key role in attention orienting and may contribute to the hypervigilance that is a common symptom of pathological anxiety. Hyperactivation of this region during elevated state anxiety may account for the paradoxical facilitation of performance on tasks that require an external focus of attention, and impairment of performance on tasks that require an internal focus of attention.

Trial registration: ClinicalTrials.gov NCT00047853.

Keywords: alpha; anxiety; fMRI; global connectivity; human; magnetoencephalography; neuroscience; startle.

Figure 1.. Schematic of experimental paradigm.

(A) Subjects underwent alternating blocks of threat and safety. (B) Visual display present on the screen during the experiment. During the experiment subjects saw two circles. The color of the outer circle indicated the block type. The color of the inner circle was controlled by the subject, and reflected the subject’s then-current anxiety level. DOI: http://dx.doi.org/10.7554/eLife.23608.003

Figure 2.. Behavioral results from both experiments.

(A) Anxiety ratings during the MEG study. (B) Startle magnitude during the MEG study. (C) Anxiety ratings during the fMRI study. Bars represent the mean ± within-subject SEM (Cousineau, 2005). (D) Correlations between anxiety potentiated startle (APS) and differential anxiety ratings. The black squares represent the correlation between APS and ratings during the MEG session. The red dots represent the correlation between APS during the MEG study and anxiety ratings during the fMRI study in the subset of subjects who participated in both studies. DOI: http://dx.doi.org/10.7554/eLife.23608.005

Figure 3.. Overview of global brain connectivity (GBC) measure.

(A) Map showing average GBC across all safe and threat TRs. (B) Cartoon schematic of a correlation matrix. The 43204 voxel x 43204 voxel cross correlation matrix was calculated separately for each subject and each condition. Correlations were averaged across rows for the entire grey matter mask, to create a single map reflecting the average correlation between each voxel and all other voxels in the mask. (C) Graph representing the mean GBC following the Fisher’s Z transformation for safe and threat averaged across the entire grey matter mask. Bars represent the mean ± within-subject SEM (Cousineau, 2005). DOI: http://dx.doi.org/10.7554/eLife.23608.007

Figure 4.. Results from voxelwise global brain connectivity (GBC) analysis.

(A) Statistical map showing results from a threat vs. safe paired-sample t-test. (B) Graph representing average GBC values after applying the Fisher’s Z transformation for clusters shown in panel A. Bars represent the mean ± within-subject SEM (Cousineau, 2005). DOI: http://dx.doi.org/10.7554/eLife.23608.010

Figure 5.. Results from bilateral IPS seed-based connectivity analysis.

(A) Statistical map showing results from a threat vs. safe paired-sample t-test. (B) Graph representing average IPS connectivity values for clusters shown in panel A. Bars represent the mean ± within-subject SEM (Cousineau, 2005). DOI: http://dx.doi.org/10.7554/eLife.23608.013

Figure 6.. Overview of MEG analyses.

(A) Spectrogram representing power averaged across all subjects and all sensors with peak in the alpha frequency band. (B) Graph showing the frequency of peak alpha (individual alpha frequency) averaged across subjects. Bars represent the mean ± SEM. (C) Example of single subject alignment with sensors (black dots) source grid (green dots) and headmodel (surface) plotted together. DOI: http://dx.doi.org/10.7554/eLife.23608.015

Figure 7.. Alpha results from threat vs. safe t-test.

(A) Statistical map in sensor space showing a significant reduction in alpha power. Black symbols represent clusters of sensors showing significant threat vs. safe differences. (B) Graph showing average alpha power for safe and threat conditions in the largest cluster of sensors in panel A. (C) Statistical map in source space showing a significant reduction in alpha power. (D) Graph showing average alpha power for safe and threat conditions in the cluster in panel C. Bars represent the mean ± within-subject SEM (Cousineau, 2005). DOI: http://dx.doi.org/10.7554/eLife.23608.017

Figure 8.. Conjunction map from voxelwise fMRI GBC analysis and MEG alpha power differences.

Colors represent significant safe vs. threat differences from the fMRI analysis (yellow), MEG analysis (blue), and both analyses (green). DOI: http://dx.doi.org/10.7554/eLife.23608.019

Author response image 1.

DOI: http://dx.doi.org/10.7554/eLife.23608.022

Author response image 2.

DOI: http://dx.doi.org/10.7554/eLife.23608.023

Author response image 3.

DOI: http://dx.doi.org/10.7554/eLife.23608.024

Author response image 4.. Α connectivity across AAL regions during 2 second baseline prior to startle probe.

A) Mean α connectivity across AAL regions during safe periods. B) Mean α connectivity across AAL regions during threat periods. C) Unthresholded T-test results comparing α connectivity during safe and threat conditions. Thresholded T-test results comparing α connectivity during safe and threat conditions. Labels on Y-axis correspond to regions of the AAL axis. Labels on the X-axis correspond to groups from AAL atlas (frontal, limbic, occipital, parietal, subcortical, temporal, cerebellum). Boxes in A and B represent α connectivity in the occipital cortices. DOI: http://dx.doi.org/10.7554/eLife.23608.025

Author response image 5.. Β connectivity across AAL regions during 2 second baseline prior to startle probe.

A) Mean β connectivity across AAL regions during safe periods. B) Mean β connectivity across AAL regions during threat periods. C) Unthresholded T-test results comparing β connectivity during safe and threat conditions. Thresholded T-test results comparing β connectivity during safe and threat conditions. Labels on Y-axis correspond to regions of the AAL axis. Labels on the X-axis correspond to groups from AAL atlas (frontal, limbic, occipital, parietal, subcortical, temporal, cerebellum). DOI: http://dx.doi.org/10.7554/eLife.23608.026

Author response image 6.. Adjacency matrices constructed from downsampled timeseries for the α and β bands.

DOI: http://dx.doi.org/10.7554/eLife.23608.027

Author response image 7.. Adjacency matrices constructed from non-downsampled timeseries for the α and β bands.

DOI: http://dx.doi.org/10.7554/eLife.23608.028

Author response image 8.. Adjacency matrices constructed from downsampled timeseries that have been converted to z-scores for the α and β bands.

DOI: http://dx.doi.org/10.7554/eLife.23608.029

Author response image 9.. Adjacency matrices constructed from non-downsampled timeseries that have been converted to z-scores for the α and β bands.

DOI: http://dx.doi.org/10.7554/eLife.23608.030