This is Part 2 of a two-part series. See Part 1.
July 7, 2014. Re-energized after a coffee break, the audience at the Schizophrenia International Research Society meeting returned for round two of the April 6, 2014, opening plenary session, which was devoted to connectomics. Instead of looking at the brain part by part, speakers would discuss "a whole new way of looking at the brain," explained moderator Ren Kahn, University Medical Center Utrecht, the Netherlands. The focus would now be on views of the brain as a whole and the abnormalities in the connections between various areas, both structurally and functionally, in schizophrenia (see SRF related news report; SRF news report).
Ed Bullmore, University of Cambridge, UK, provided an overview of the topological (structural) properties of human connectomes. The fast-paced advances of the physics of complex networks—from social networks to the Internet, to the brain—have revealed that many share important organizational properties. As the network science movement has "spilled over" into neuroscience, it has been met with a growing quantity and quality of data that afford unprecedented access to images of brain networks and other large-systems neuroscience databases, Bullmore added.
For the first time, it is possible to "reasonably believe that we might be able to map out the complete connectome architecture," he said.
Bullmore went on to describe a number of topological properties that characterize all human connectomes. One property is the existence of "hubs," nodes that have a higher than average number of connections with the rest of the network. Another property is the so-called rich club ordering, in which highly connected (or high-degree) nodes tend to be tightly interconnected with each other. These hubs and clubs are highly expensive in terms of blood flow and glucose metabolism, but also support higher order, integrative behaviors such as executive function, Bullmore said.
New data from his laboratory suggest that it is exactly these high-cost/high-value components of the network that are most likely to be damaged or lesioned in brain disorders, said Bullmore. A meta-analysis of more than 20,000 patients identified nine brain disorders, including schizophrenia, with lesions more likely to be in hubs than other regions of the network (Crossley et al., 2014). The affected hubs differ among illnesses, he added, and are consistent with symptomatology. In Alzheimer's disease, temporal lobe hubs are more likely to be damaged, while in schizophrenia, frontal lobe hubs are most affected.
Structure vs. function
Deanna Barch, Washington University, St. Louis, Missouri, focused more specifically on the schizophrenia connectome. Although the focus in years past has been on abnormalities in individual brain regions, we now know that a host of brain regions are impacted in the illness, she said. "This is a tale of many regions, not just one," Barch added.
However, it is important to make the distinction between whether all regions are affected or just some, she said. A global difference in brain connectivity suggests a different etiological mechanism than more focal abnormalities.
When important confounds, such as increased motion in patients compared to controls, are accounted for, overall network structure seems to be intact in schizophrenia, said Barch. In addition, connectivity within known networks appears to be normal in the illness. In contrast, connectivity between networks, particularly the frontal-parietal, cerebellar, and cingulo-opercular networks, is reduced in schizophrenia. The degree of disrupted connectivity predicts cognition, IQ, working memory, and symptoms, she added.
This disconnectivity is not just present at a functional level; structural connectivity is also disturbed in schizophrenia, said Barch. A reduced strength of rich club connections suggests that there are key hubs in the brain that are impaired in psychosis, and supports the idea of schizophrenia as a "disconnection" syndrome (see SRF related news report).
Lessons from animals
The final speaker of the session was Holly Moore of Columbia University, New York City, who also emphasized that in schizophrenia, some circuits are more affected than others. Moore focused primarily on data gleaned from neuroanatomical and electrophysiological studies conducted in animals. More so than human studies, animal work has highlighted the role of subcortical structures in behaviors that are affected in schizophrenia, she said.
Moore highlighted the striatum as a "place where there is considerable opportunity for convergence between limbic, temporo-limbic, and prefrontal cortical connections," that are affected in schizophrenia.
She proposed several principles to keep in mind while reading the connectomics literature. Structural connectivity can be affected by the direction of a connection, the density and myelination of the tract, the synaptic density within the target region, the neurotransmitter used, and the neuronal class(es) receiving synapses. Similarly, when thinking about functional connectivity, it is important to consider the neurotransmitter action on target neurons, the activity pattern of a projection, and the synchrony across populations, said Moore.—Allison A. Curley.