Edited by: John J. Foxe, Albert Einstein College of Medicine, USA
Reviewed by: Christopher Marano, Johns Hopkins University School of Medicine, USA; Ciara McCabe, University of Oxford, UK
*Correspondence: Shawn Hayley, Department of Neuroscience, Carleton University, 327 Life Sciences Research Building, 1125 Colonel By Drive, Ottawa, ON, Canada K1S 5B6. e-mail:
This is an open-access article distributed under the terms of the
Major depressive disorder (MDD) is characterized by a dysregulated fronto-limbic network. The hyperactivation of limbic regions leads to increased attention and processing of emotional information, with a bias toward negative stimuli. Pathological ruminative behavior is a common symptom of depressive disorder whereby the individual is unable to disengage from internal mental processing of emotionally salient events. In fact, lower deactivations of the neural baseline resting state may account for the increased internal self-focus. The insular cortex, with its extensive connections to fronto-limbic and association areas has recently also been implicated to be a part of this network. Given its wide-reaching connectivity, it has been putatively implicated as an integration center of autonomic, visceromotor, emotional, and interoceptive information. The following paper will review recent imaging findings of altered insular function and connectivity in depressive pathology.
According to the World Health Organization, major depressive disorder (MDD) will be the second-leading cause of disability in the world by the year 2020, following heart disease (Möller and Henkel,
A myriad of scientific studies have examined the neural circuitry involved in emotional regulation, especially as it applies to mood disorders. It is believed that a dysregulation in brain regions associated with affect are central in the etiology of MDD (Kring and Bachorowski,
Brain imaging studies support the view that fronto-limbic networks are dysregulated in major depression. Indeed, reductions in gray matter density (Frodl et al.,
Recent imaging research suggests an altered basal neural resting state in MDD patients, characterized by increased neural activity in certain brain regions encompassing the default mode network (DMN; Lemogne et al.,
The ensuing sections will review behavioral markers and neurochemical alterations in MDD. As well, the neural correlates of MDD in the context of disrupted fronto-limbic connectivity will be discussed. Recently, the brain’s resting state (i.e., DMN) has garnered increasing attention as a dysregulated network in MDD. As the focus of this paper, the insular cortex, a region reported to be a part of the DMN and an important cortical structure implicated in emotional processing and regulation (among other functions) will be further examined in the context of MDD. Finally, recent imaging findings will be presented contrasting the role of the insula in depression across different information processing paradigms (cognitive, social, emotional, and memory systems). As well, the insula’s role in the resting state and following therapeutic treatment will be examined.
Emotional self-regulation is hindered in individuals suffering from MDD, causing a disruption in their attentional focus and creating memory biases toward negative life events. This results in an altered mood state due to negative sensory representations, which consequently modify cognitive processes (Siegle et al.,
Several meta-analyses have highlighted the role of the insula in depression with differing results. Hamilton et al. (
Emotional regulation studies seem to provide a more consistent view of the insula’s role in depressive disorder. A disruption in top-down processing is thought to mediate pervasive feelings of sadness and negative affect in depressed populations, whereby dorsolateral prefrontal and inferior parietal cortex activity is decreased and subgenual cingulate and anterior insula activity is increased (Mayberg et al.,
Several studies have linked insular function to depressive symptoms and behaviors. A principal component analysis of the Beck Depression Inventory revealed that psychomotor anhedonia symptom severity, characterized by decreased satisfaction and loss of interest in others, was negatively correlated with metabolism in the right insula (Dunn et al.,
Given that hereditary factors are often implicated in mood disorders, it should not be surprising that offspring of parents diagnosed with depression show altered neural functioning in mood processing and regulating brain areas. Using an incentive delay task whereby prizes are awarded upon hitting a target, Gotlib et al. (
It is plausible that a negative self-image (Northoff,
In major depression and in general anxiety states, the NE and serotonin (5-HT) systems are clearly imbalanced, with substantial data indicating that NE responses are increased to stressor exposure and fearful stimuli, whereas the 5-HT system is hypo-responsive. These neurotransmitter alterations contribute to the over-activation of the amygdala, hippocampal, and cortical pathways activating stress and fear responsiveness (Ressler and Nemeroff,
Atrophy of the prefrontal cortex likely decreases the ability of cortical modulation and inhibition of amygdala aversion pathways. It has been proposed that the prefrontal cortex is involved via connections with the amygdala in regulating affect, providing cognitive control over stress and fear responsiveness along with anger, anxiety, and frustration tolerance (Lang et al.,
Serotonergic synapses are most densely concentrated in limbic regions including the amygdala and the BNST, as well as the ventral striatum and hypothalamus. Serotonin from the raphe nuclei mediates tolerance to aversive experience in the amygdala, potentially decreasing the likelihood of a fight or flight response (Ressler and Nemeroff,
Other neurotransmitters have also been implicated in insula activity. For instance, reduced binding of a metabotropic glutamate receptor (mGluR5) was observed in the insula, cingulate, and hippocampal brain regions in depressed patients compared to healthy controls (Deschwanden et al.,
Neuroanatomically, depressive disorders are associated with changes in the emotional circuitry, encompassing the limbic structures as well as the prefrontal and parietal cortices. The amygdala, in association with other fronto-limbic structures, plays a pivotal role in the processing of emotional stimuli and arbitrates emotional influences in other brain regions related to attention, memory, and decision-making (Hariri et al.,
Ochsner et al. (
In addition to the functional alterations in the emotional processing and regulation centers of the brain, structural abnormalities have also been noted in hippocampal and prefrontal regions (see reviews by Fossati et al.,
The DMN has been defined as the baseline neural activity when one is disengaged from externally cued cognitive demands (Gusnard and Raichle,
The activities of the DMN have been linked to intrinsic processing (i.e., monitoring of one’s internal and external state Gusnard and Raichle,
Early studies using cortical stimulation techniques in humans and primates have elucidated the role of the insula with visceral phenomena, including abdominal sensations, gastric motility, respiratory, and cardiovascular functions (Penfield and Faulk,
Recently, the insula has garnered increasing attention as being involved in conscious awareness, whereby it has been inferred to hold a representation of one’s current subjective feeling of their body’s physiological state (Craig,
Emerging evidence from neuroimaging studies further lend support to the function of the insular cortex as a hub of meta-awareness, general arousal, and interoceptive processing (Bluhm et al.,
The insula has been further reported to be part of a “salience network” along with the anterior and mid-cingulate cortex (Sridharan et al.,
Neuroimaging studies have provided much evidence for a pathological activation of the brain’s limbic circuits (e.g., amygdala, anterior insular cortex) in depressed individuals in response to emotional events (Drevets,
Default mode networks have further been investigated as impaired neural targets in MDD pathology. As mentioned previously, the brain’s “resting circuitry” involving cortical midline structures (e.g., vmPFC, pACC, PCC) serves as a neural baseline of task-independent activity, showing greater activation in response to non-demanding self-related processing [e.g., reflection and unconstrained mental activity (mind wandering); Kjaer et al.,
Using positive and negatively valenced pictures, Grimm et al. (
Grimm et al. (
A recent meta-analysis by Delaveau et al. (
Paradigms exploring the brain at rest (e.g., staring at a fixation cross) provide a “state of the affairs” perspective of the neural baseline in MDD. To our knowledge, only two studies have explicitly examined insular activity in depressed populations using a resting state paradigm. The resting state in medication-free MDD patients has been shown to have a decreased functional connectivity in the left insula and amygdala with the resting state network (Veer et al.,
Study authors | Population | Paradigm | Imaging results | ||
---|---|---|---|---|---|
Hamilton et al. ( |
MDD patients ( |
One fMRI scan was taken at rest (eyes closed) | ↓ Activity in the right fronto-insular cortex in MDD patients in the DMN | ||
Healthy controls ( |
Observed default mode (DMN) and task-positive network (TPN) activations | ↑ Activity in the right fronto-insular cortex in healthy controls in the TPN | |||
Veer et al. ( |
Recently diagnosed (< 6mo), medication-free MDD patients ( |
Resting-state fMRI scan (eyes closed) | ↓ Functional connectivity between the left insula and the resting state network in MDD patients | ||
Healthy controls ( |
|||||
Yao et al. ( |
MDD in-patients ( |
Resting scan (eyes closed) | ↓ ReHo in right insula, ACC/PCC in MDD patients | ||
Healthy controls ( |
Anxiety severity, retardation, and hopelessness all positively correlated with ReHo in the right insula | ||||
Liu et al. ( |
First episode un-medicated MDD patients ( |
Resting scan (eyes closed) | ↓ ReHo in right insula in MDD patients compared to healthy controls |
||
Healthy controls ( |
|||||
Guo et al. ( |
Treatment resistant MDD patients ( |
Resting scan (eyes closed) | ↓ ReHo in left insula in treatment resistant MDD patients compared to healthy controls | ||
Healthy controls ( |
|||||
Lee et al. ( |
MDD patients ( |
Structural MRI scans were obtained to detect structural abnormalities using voxel-based morphometry (VBM) | ↓ Decreased gray matter concentration was found the insular gyrus (among other limbic, cerebellar, and frontal regions) in MDD patients compared to healthy controls | ||
Gray matter concentration in the left insular gyrus was negatively correlated with illness duration in MDD patients compared to healthy controls |
Hamilton et al. (
Regional homogeneity is a non-parametric analysis measure used to calculate local connectivity within a cluster, which is based on time course BOLD signal correlations between a voxel and its neighbors (Zang et al.,
In fact, voxel-based morphometry, a structural measurement of gray and white brain matter density, reveals that in MDD patients, gray matter concentration is decreased in the limbic system, the insular gyrus, and dorsal raphe nuclei (Lee et al.,
In order to evaluate the neural networks involved in depressive pathology, Lui et al. (
Study authors | Population | Paradigm | Imaging results | ||
---|---|---|---|---|---|
Lui et al. ( |
Un-medicated MDD patients, HAMD ≥ 18 on day of scan ( |
Participants underwent a baseline resting scan (eyes closed) and MDD patients were placed in one of 3 antidepressant medication conditions (tricyclics, SSRIs, SNRIs) | Non-refractory patients had a more distributed ↓ connectivity among ACC, bilateral PFC, HPC, amygdala, and bilateral insula compared to controls |
||
Kito et al. ( |
Treatment resistant MDD patients ( |
A PET scan was taken before and after TMS treatment (eyes closed). |
↓ rCBF in anterior and posterior insula following low-frequency TMS |
||
Martinot et al. ( |
Treatment resistant MDD patients ( |
Participants underwent a baseline resting scan (eyes closed) and were placed in one of 3 TMS conditions: |
Compared to healthy controls, non-responders to TMS treatment had |
These findings are surprising given that refractory patients, who showed a poor response to treatment, had increased connectivity in brain regions important for regulating mood states compared to MDD patients having successfully responded to antidepressant treatment. In fact, the refractory patients showed decreased connectivity in mainly prefrontal and thalamic areas in addition to limbic regions (Lui et al.,
Transcranial magnetic stimulation therapy is a non-invasive technique used typically in MDD patients otherwise not responding to conventional medication (Kim et al.,
In line with the TMS findings described above, using FDG-PET, Martinot et al. (
One of the main clinical symptoms in MDD pathology is the bias toward stimuli of negative valence. The inability to suppress the increased attention to these stimuli causes a disruption in normal emotional processing. In fact, Herwig et al. (
Study authors | Population | Paradigm | Imaging results |
---|---|---|---|
Herwig et al. ( |
Medicated MDD in-patients ( |
fMRI cueing task: Known condition (cue was presented showing emotional valence of picture) | ↑ bilateral insula activation in response to pending picture presentation with unknown emotional valence |
Healthy controls ( |
Unknown condition (absence of cues) | Positive correlation between depressive symptoms and right insula during unknown and negative picture expectation | |
Surguladze et al. ( |
MDD patients ( |
Two 6-min experiments: photos showing expressions of disgust and neutral faces | ↑ activation in the left insula in response to expressions showing 50 and 100% disgust, compared to neutral expressions |
Healthy controls ( |
Fearful expressions and neutral faces Participants were tested on accuracy of labeling the facial expressions. | ||
Samson et al. ( |
Un-medicated MDD patients ( |
Emotional paradigm: passively view sad facial expressions and neutral pictures (homes in Munich) |
Negative correlation between depressive symptom severity and bilateral insula activation of MDD patients following drug treatment when viewing sad facial expressions compared to the neutral pictures |
Demenescu et al. ( |
MDD out-patients ( |
Presented color facial expressions (angry, fearful, sad, happy neutral, scrambled faces) | ↑ activation in the right insula in response to happy vs. scrambled faces in MDD patients compared to healthy controls |
Anxiety disorder out-patients ( |
|||
MDD/Anxiety out-patients ( |
Participants were asked to rate the gender of each picture | ||
Healthy controls ( |
|||
Sprengelmeyer et al. ( |
Medicated MDD patients ( |
Experiment 1: facial recognition of emotion + morphed faces on a continuum. Participants had to attribute the emotion to the face | Experiment 1: average discrimination accuracy for MDD patients was positively correlated with gray matter volume in bilateral anterior insula |
Healthy controls ( |
Experiment 2: facial recognition at 4 different intensities (20, 40 60, 80%). Participants had to press a key as soon as they identified the emotion | Experiment 2: ↓ gray matter volume in insular cortex in MDD patients | |
Townsend et al. ( |
Un-medicated MDD patients ( |
Face matching paradigm: 3 conditions: |
No significant insular activation in MDD patients |
Increased insular activation in MDD patients is not limited to negative stimuli. It has also been detected when processing happy facial expressions compared to scrambled faces (Demenescu et al.,
Given the clinical symptoms associated with depressive pathology (e.g., anhedonia, impaired concentration, and working memory deficits), task-related paradigms aim to provide insight into the neural alterations associated with symptomatology in MDD patients. Videbech et al. (
Study authors | Population | Paradigm | Imaging results |
---|---|---|---|
Videbech et al. ( |
MDD in-patients with a HDRS score ≥ 17 ( |
PET scan taken at rest | ↓ Activity in the left insula in MDD patients when comparing verbal fluency to the resting state |
Healthy controls ( |
PET scan taken during phonological verbal fluency task. fMRI scan used as a template for spatial normalization of PET data | ↑ Activity in the bilateral insula in healthy controls when comparing verbal fluency to the resting state | |
Young et al. ( |
Un-medicated MDD patients in a current depressive episode ( |
Autobiographical memory task (fMRI): Cue words were presented (positive, negative, neutral) | ↓ BOLD signal in the anterior insula during memory recall in both groups, however, greater reduction in MDD patients |
Healthy controls ( |
Participants pressed a button once a memory was recalled and rated it on valence, arousal, and recency | ↓ BOLD signal in bilateral anterior insula and posterior insula during autobiographical memory recall in both groups (magnitude of reduction greater in MDD patients) | |
Strigo et al. ( |
Un-medicated MDD patients for a minimum of 30 days ( |
fMRI: 4 conditions: |
When shifting stimulus intensity (1 vs. 3 and 2 vs. 4), MDD patients show ↓ activation in the right anterior insula and bilateral mid-insula |
Using an autobiographical memory paradigm, Young et al. (
Major depressive disorder patients also show impairments in adapting to different environmental or homeostatic states. Strigo et al. (
Major depressive disorder is associated with deficits in emotional processing, with a bias toward negative stimuli. This may be due to the fronto-limbic alterations in MDD pathology, whereby limbic regions are hypoactivated and frontal executive regions show decreases in activity. As such, the top-down modulation of emotionally based information is impaired, leading to depressive symptomatology (e.g., rumination). These deficits in emotional processing could be further exacerbated by the morphological changes (e.g., gray matter volume reductions in fronto-limbic structures) and neurochemical imbalances in the brain.
Recently, much attention has focused on the neural basis of depressive pathology, particularly with respect to the baseline resting state in MDD patients. Disrupted connectivity in fronto-limbic structures is commonly reported as are abnormal deactivations in default mode regions. This supports the on-going mental processing of emotionally based information in MDD patients who have great difficulty disengaging from internal and external input. Of interest, the insular cortex with its many functions in emotional, sensorimotor and interoceptive processing, and extensive connectivity with DMN regions, also has been implicated in MDD pathology.
Review of recent literature points to decreases in connectivity of the insula with the amygdala and default mode regions (anterior and posterior cingulate cortex). As well, the insula shows greater activity in the resting state, supporting the inability of MDD patients to disengage from externally cued events, which may lead to pathological self-focused mental ruminative behaviors. However, therapeutic brain stimulation studies have shown positive results in regulating insular activity by decreasing its activation in MDD patients, pointing to an effective therapeutic strategy in re-establishing functional emotional processing.
Differing results have been obtained when looking at the processing of cognitive and emotionally based information in the insular region of MDD patients. Specifically, cognitive-based tasks requiring working memory, retrieval, and active processing showed decreases in the insula which may suggest impairments in integrating various cognitive processes required for mentally demanding activities. On the other hand, studies using emotionally based paradigms showed increases in insular activity. These findings may be due to the MDD patients’ greater susceptibility in attending to emotionally arousing stimuli. Indeed, greater insular activity was found primarily in response to negative emotions, pointing to impaired processing and regulation of negative mood and emotions.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.