September 10, 2013. Modern brain imaging has allowed researchers to start mapping networks of brain areas that carry out different tasks (for review, see Bressler and Menon, 2010). Of particular interest to schizophrenia researchers is the "salience network," which is thought to monitor a person’s immediate needs, select the stimuli most relevant to those needs, and feed them to the "central executive network" for further processing to ultimately guide behavior. These two networks do not appear to talk to each other as they should in schizophrenia, and two recent studies focus the blame for this on a key component of the salience network, the insular cortex.
By tracking the timing of brain activation with functional magnetic resonance imaging (fMRI) in people at rest, the two studies find a breakdown in how the anterior insula shapes activity in the dorsolateral prefrontal cortex (DLPFC)—a key member of the central executive network—in schizophrenia. Though fMRI studies often focus on functional connectivity between brain regions, as measured by simultaneous activity in any two regions, these studies examined “effective connectivity,” which measures how activity in one region precedes or predicts activity in another region.
The first study, led by Lauren Moran of the Maryland Psychiatric Research Center in Baltimore and published online April 25 in Biological Psychiatry, focused on selected brain regions and found that, in schizophrenia, the right anterior insula did not modulate the central executive network as it should. The second study, led by Lena Palaniyappan in collaboration with Peter Liddle at the University of Nottingham in the United Kingdom and published August 21 in Neuron, found much the same thing, but through a more agnostic, brainwide approach. Not only did the anterior insula fail to influence the DLPFC in schizophrenia as it did in controls, but feedback from the DLPFC to the anterior insula was also weaker than normal.
"This story is evolving rapidly and in an exciting direction," said Angus MacDonald, SRF advisor and researcher at the University of Minnesota, who was not involved in the studies. "There is specificity of the relationships as demonstrated by convergence of these studies."
The findings suggest a central role for the insula in schizophrenia. Though structural and functional anomalies in the insula are consistently found in the disorder (e.g., Ellison-Wright et al., 2008), these often take a back seat to findings in frontal or temporal regions. The two new studies suggest that well-known frontal dysfunctions in schizophrenia could stem from weak connections with the anterior insula.
The findings also bolster the idea that schizophrenia’s pathology involves disrupted interactions between large-scale networks within the brain. Disruptions in how networks communicate with each other in schizophrenia first emerged in studies of the default-mode network, a collection of brain regions activated when a person is resting quietly and suppressed when one is doing a task. In schizophrenia, this network is not adequately suppressed during tasks, suggesting a difficulty in disengagement from an introspective state (see SRF related news story).
The new research suggests this may be due to problems with the insula-containing salience network, which normally orchestrates the switch between the default-mode network and the central executive network, which enables task performance (Menon and Uddin, 2010). Consistent with this, another recent study found decreased functional connectivity within the anterior insula in schizophrenia, and this correlated with disturbed interactions between the salience network and the central executive network, as well as with severity of negative symptoms (Manoliu et al., 2013).
“The problem is not simply in one of these networks, but rather in the switching between the two networks,” Palaniyappan told SRF. “This is revealing because we know that people with schizophrenia find it very difficult to switch between their internal representations, and they find it hard to process external information as well.”
Centered on salience
“These studies build on and modify the leading hypothesis of this time, which, in my opinion, is aberrant salience," MacDonald wrote to SRF. The aberrant salience hypothesis of schizophrenia proposes that faulty dopamine signals incorrectly assign importance to innocuous elements of a person’s experience (Kapur, 2003), thus producing the disordered thinking and delusions of psychosis. The new findings may be seen as a brain network-based extension of this, in that a malfunctioning salience network does not flag the most relevant things in a person’s environment.
“The insula takes in signals from every part of the body and creates a kind of neuronal readiness to orient to the most appropriate things,” Palaniyappan said. For example, hunger signals from the gut would create a neuronal readiness in the brain to pay attention to any sign of food that might come along. Problems with this state of readiness, which Palaniyappan calls “proximal salience,” would mean that the things that matter most to people’s well being would not necessarily capture their attention.
Although effective connectivity requires looking at the timing of activations across the brain, the poor temporal resolution of fMRI precludes any firm conclusions about which area directly drives another. Nevertheless, researchers can get a sense of the direction of information flow between them.
In the Biological Psychiatry study, first author Moran and colleagues found a reduced outflow of activity from the right dorsal anterior insula. Based on brain activity measured in 44 people with schizophrenia and 44 healthy controls, the results were the same for two methods of measuring effective connectivity. In controls, activity in the anterior insula was followed by robust activity in regions belonging to the central executive network, including the DLPFC, as well as in regions of the default-mode network, including the posterior cingulate cortex (PCC) and the lateral parietal cortex. In schizophrenia, however, activity in the anterior insula did not switch these networks on. Specifically, the usual follow-up activity in the DLPFC, PCC, and lateral parietal cortex did not emerge.
As in previous studies of inter-network connectivity, these results came from people asked to simply rest in the scanner. This raises questions for some about how to interpret any differences found in brain activity in people with a mental illness, who may experience the scanner differently (Buckner et al., 2013). Moran’s study also measured brain activity while study participants were engaged in a sustained attention task, which might engage the brains of both groups more similarly. Even then, reductions in anterior insula outflow were found in schizophrenia compared to controls, particularly to the default-mode network, and these correlated with task performance.
In the Neuron study, first author Palaniyappan and colleagues imaged brain activity in 38 people with schizophrenia and 35 healthy controls while they rested. To find the region that was maximally influenced by right anterior insula activity, the researchers compared right anterior insula activity in one scan with activity in all brain voxels in the next scan, taken 2.5 seconds later. For both controls and people with schizophrenia, the DLPFC turned up as a region whose activity correlated with earlier anterior insula activity; in the schizophrenia group, however, this correlation was weaker than in controls, indicating that the anterior insula had a weaker than usual influence on the DLPFC. When the researchers looked across the brain for neural repercussions of DLPFC activity, the anterior insula turned up: In both groups, its activity was negatively correlated with earlier DLPFC activity (indicating a decrease in activity following DLPFC activity), but in schizophrenia, this correlation was weaker.
These data suggest that the right anterior insula and the DLPFC do not exert normal levels of control over each other in schizophrenia. This disruption may be a core feature of the disorder, as a measure of the extent of this “loop failure” varied with illness severity but not with antipsychotic dosage.
Palaniyappan noted that the resting-state paradigm he used avoids performance-related differences in brain activity that might occur between the two groups. Still, he said he expected a complementary task-based study would give a more pronounced effect than what was found here.
The researchers also reported a decrease in influence from the visual cortex onto the anterior insula in schizophrenia compared to controls. This suggests a chain reaction: The visual cortex doesn’t send enough signals to the insula, the insula doesn’t send enough signals to the frontal cortex, and the frontal cortex does not provide the usual feedback to the insula. Though all correlational, the results recall bottom-up theories of schizophrenia in which impaired sensory processing brings about its complex symptoms (Javitt, 2009, and see SRF related news story).
The new results put the insula at the center of a model of information processing in schizophrenia that can be further tested with methods such as magnetoencephalography (MEG), which better captures the timing of brain activations, or other task-based schemes. The intertwined nature of the insula and frontal cortex suggests that modulating one, possibly through non-invasive stimulation methods, may rehabilitate the other. The results highlight the interconnected nature of the brain and make steps toward understanding which connections matter to schizophrenia.—Michele Solis.
Moran LV, Tagamets MA, Sampath H, O'Donnell A, Stein EA, Kochunov P, Hong LE. Disruption of anterior insula modulation of large-scale brain networks in schizophrenia. Biol Psychiatry. 2013 Sep 15;74(6):467-74. Abstract
Palaniyappan L, Simmonite M, White TP, Liddle EB, Liddle PF. Neural primacy of the salience processing system in schizophrenia. Neuron. 2013 August; 79: 814-828. Abstract