Ketamine and Attentional Bias Toward Emotional Faces: Dynamic Causal Modeling of Magnetoencephalographic Connectivity in Treatment-Resistant Depression

This double-blind, placebo-controlled, within-subjects study (n=34) used magnetoencephalographic (MEG) recordings to find a correlation between anti-depression effects (for those suffering from depression; TRD & BD) and neurological changes (e.g. faster GABA, AMPA, and NMDA transmission in specific brain areas).

Abstract

“The glutamatergic modulator ketamine rapidly reduces depressive symptoms in individuals with treatment-resistant major depressive disorder (TRD) and bipolar disorder. While its underlying mechanism of antidepressant action is not fully understood, modulating glutamatergically-mediated connectivity appears to be a critical component moderating antidepressant response. This double-blind, crossover, placebo-controlled study analyzed data from 19 drug-free individuals with TRD and 15 healthy volunteers who received a single intravenous infusion of ketamine hydrochloride (0.5 mg/kg) as well as an intravenous infusion of saline placebo. Magnetoencephalographic recordings were collected prior to the first infusion and 6–9 h after both drug and placebo infusions. During scanning, participants completed an attentional dot probe task that included emotional faces. Antidepressant response was measured across time points using the Montgomery-Asberg Depression Rating Scale (MADRS). Dynamic causal modeling (DCM) was used to measure changes in parameter estimates of connectivity via a biophysical model that included realistic local neuronal architecture and receptor channel signaling, modeling connectivity between the early visual cortex, fusiform cortex, amygdala, and inferior frontal gyrus. Clinically, ketamine administration significantly reduced depressive symptoms in TRD participants. Within the model, ketamine administration led to faster gamma aminobutyric acid (GABA) and N-methyl-D-aspartate (NMDA) transmission in the early visual cortex, faster NMDA transmission in the fusiform cortex, and slower NMDA transmission in the amygdala. Ketamine administration also led to direct and indirect changes in local inhibition in the early visual cortex and inferior frontal gyrus and to indirect increases in cortical excitability within the amygdala. Finally, reductions in depressive symptoms in TRD participants post-ketamine were associated with faster α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) transmission and increases in gain control of spiny stellate cells in the early visual cortex. These findings provide additional support for the GABA and NMDA inhibition and disinhibition hypotheses of depression and support the role of AMPA throughput in ketamine’s antidepressant effects.“

Authors: Jessica R. Gilbert, Christina S. Galiano, Allison C. Nugent & Carlos A. Zarate

Summary

INTRODUCTION

Ketamine’s rapid antidepressant effects have galvanized research into the neurobiological underpinnings of mood disorders, and many studies have now demonstrated that ketamine can relieve depressive symptoms in individuals with both major depressive disorder and bipolar depression, including those who are treatment-resistant.

Ketamine is a non-competitive N-methyl-D-aspartate (NMDA) receptor antagonist, but several studies suggest that NMDA receptor antagonism may not be the direct mechanism underlying ketamine’s antidepressant effects. Furthermore, ketamine may also activate plasticity mechanisms and promote synaptic potentiation.

A growing body of evidence suggests that altering the ratio of cortical excitation/inhibition balance could underlie a host of disorders, including depression. Therapeutic-dose ketamine increases gamma power in TRD participants.

Individuals with MDD show a bias toward negative emotional information, including a bias toward faces demonstrating negative emotions compared to positive emotions. Ketamine has been shown to normalize brain activation in TRD patients in regions of frontal cortex and the amygdala.

This study used magnetoencephalography (MEG) and dynamic causal modeling (DCM) to model effective connectivity between regions of interest (ROIs) in a group of participants with TRD and healthy volunteers who underwent both ketamine and placebo saline infusions. A network of regions activated by the task was modeled using DCM and compared to baseline and placebo scans. Ketamine administration was predicted to increase gamma power in the amygdala, a key region involved in the emotional processing of face stimuli.

Participants

Participants were 19 individuals with a DSM-IV-TR diagnosis of TRD without psychotic features and 15 healthy volunteers who had failed on average 3.8 antidepressant trials across their lifetime. The present study used data drawn from a larger clinical trial that assessed ketamine’s antidepressant effects. The TRD study involved hospitalized participants who were drug-free for at least 2 weeks prior to MEG testing. Healthy volunteers completed study procedures as inpatients but were otherwise outpatients.

Clinical Measurements

The MADRS was administered 60 min prior to infusions and at multiple time points following infusions. The difference between ketamine and placebo was estimated at 230 min, the time point closest to the MEG scan.

MEG Acquisition and Preprocessing

Participants completed a dot probe task with emotional face stimuli presented using E-Prime presentation software. The task used a mixed block/event-related design and was randomized for emotion, gender of face, side of emotional face, and side of probe. Trials were additionally blocked into two “angry blocks” and two “happy blocks”, resulting in four emotional face trial types: angry congruent, angry incongruent, happy congruent, and happy incongruent.

Neuromagnetic data were collected using a 275-channel CTF system with SQUID-based axial gradiometers. The data were filtered from 1 to 58 Hz and epoched from 100 to 1,000 ms peristimulus time, and then analyzed using the academic freeware SPM12.

Source Localization and Source Activity Extraction

The multiple sparse priors routine implemented in SPM12 was used to identify gamma frequency sources of activity from participant-level data over a peristimulus event time window from 100 to 1,000 ms. The main effect of infusion was tested using a more liberal criterion of p 0.05, uncorrected.

In this study, the activation of the bilateral early visual cortices, fusiform cortex, amygdala, and inferior frontal gyrus was modeled using DCM. The activation of these regions was modeled using stimulus-induced event-related potentials.

Dynamic Causal Modeling

A biophysical model of neural responses based on neural mass models was used to predict recorded electrophysiological data features. The model included connection parameters for AMPA- and NMDA-mediated glutamatergic signaling as well as GABA signaling.

Thalamic input drove activity in early visual cortex, which sent signals to the fusiform cortex, amygdala, and inferior frontal gyrus. Two models of message-passing were constructed, one with traditional bottom-up processing, and the other with presumed magnocellular projections to frontal cortex.

For the DCM analyses, an ERP model was fitted to the MEG activity over 1 – 500 ms peristimulus time in a wide frequency band from 1 to 50 Hz. The best fitting model was selected by comparing the log-model evidence of Model 1 and Model 2 across participants.

A second-level modeling extension of DCM called parametric empirical Bayesian analysis (62) was applied to determine the mixture of parameters that mediated ketamine’s effects. Group by drug interactions were of particular interest.

After examining the group effect, drug effect, and group by drug interaction using parametric empirical Bayesian analysis, post-hoc classical statistical tests were conducted to determine whether any parameters were associated with antidepressant response.

Clinical and Behavioral

Ketamine reduced MADRS score by 5.37 points compared to placebo at 230 min post-infusion. This was due to a reduction in reaction time bias and accuracy rates on the emotional dot probe task.

Although no significant behavioral effects were observed on reaction time bias scores, participants were more accurate during the baseline session than during the ketamine session and the placebo session.

Source-Level

We source-localized MEG data to identify the primary generators of the signal, and found that the left-lateralized early visual cortex, fusiform cortex, amygdala, and inferior frontal gyrus were all activated in response to the dot probe task.

Dynamic Causal Modeling

Two plausible models were constructed to account for connectivity between ROIs. Model 2 was found to have the strongest model evidence.

Parametric empirical Bayes was used to test for parameters contributing to the group effect, drug effect, and group by drug interactions. Four receptor time constants showed meaningful group by drug interactions, including the GABA time constant in the early visual cortex and the NMDA time constant in the early visual cortex.

Five connections were identified in the early visual cortex, amygdala, and inferior frontal gyrus that showed meaningful group by drug interaction effects. Ketamine reduced self-inhibitory drive on spiny stellate cells, inhibitory interneurons, and superficial pyramidal cells in the early visual cortex.

Parameters Associated With Antidepressant Response

We explored whether any parameters associated with the group effect, drug effect, or group by drug interactions were also associated with antidepressant response. Two parameters were found to be associated with antidepressant response.

DISCUSSION

This study used MEG recordings to probe ketamine’s effects in individuals with TRD and healthy volunteers. It found that ketamine increased connectivity to positive faces and decreased connectivity to negative faces.

Clinically, ketamine was found to reduce depressive symptoms in the TRD sample, consistent with previous findings (6, 9). Behaviorally, ketamine participants were more accurate than healthy volunteers during the task, and the best performance occurred during the baseline session.

We modeled gamma-band activity during the dot probe task and found increased gamma power in the amygdala post-ketamine vs. placebo for both TRD participants and healthy volunteers. This suggests that increased cortical excitation in this key emotional face processing region may be related to normalization of emotional processing following drug administration.

Two plausible models of message passing were subsequently fit, and a Bayesian modeling extension of DCM was used to test for meaningful parameters contributing to the group effect, drug effect, and group by drug interactions. Four modeled receptor time constants showed group by drug interactions, including faster GABA and NMDA transmission estimates in the early visual cortex following ketamine administration, and slower NMDA transmission estimates in the fusiform cortex and amygdala following ketamine administration.

Group by drug interactions were found for modeled intrinsic connectivity within the early visual cortex, amygdala, and inferior frontal gyrus. Ketamine decreased inhibitory drive on self-connections, increased excitatory drive on self-connections, and reduced inhibitory drive on self-gain on superficial pyramidal cells. These findings reflect changes in intrinsic connectivity that regulate or modulate inhibition locally, and may reflect a link between increased pyramidal cell excitability and increased gamma power.

We tested whether any parameters identified in our analysis of group effects, drug effects, or group by drug interactions were associated with antidepressant response in our TRD participants. Two parameters, AMPA time constant and inhibitory self-gain on spiny stellate cells, were found to be associated with antidepressant response.

We collected MEG data 6–9 h after ketamine administration to avoid side effects while measuring therapeutic drug effects. We cannot comment on acute changes in modeled parameter estimates, but previous findings have demonstrated associations between AMPA parameters and antidepressant response in TRD.

CONCLUSIONS

Ketamine administration leads to key changes in estimates of GABA and NMDA time constants, as well as excitatory and inhibitory intrinsic connectivity within key regions important for visual processing of emotional faces.

AUTHOR CONTRIBUTIONS

JG designed the study, CG conducted the literature search, AN conceptualized the study, and CZ edited the manuscript. All authors approved the submitted version.

FUNDING

This work was supported by the Intramural Research Program at the National Institute of Mental Health, National Institutes of Health, and by the NARSAD Independent Investigator Award.