Interconnections between different parts of the brain are dynamic and not static

Jun 1 2005

Using a newly released method to analyze functional magnetic resonance imaging (fMRI), Northwestern University researchers have demonstrated that the interconnections between different parts of the brain are dynamic and not static. This and other findings answer longstanding debates about how brain networks operate to solve different cognitive tasks. They are presented in the current (June 1) issue of the Journal of Neuroscience.

Equally important, the researchers discovered that the brain region that performed the integration of information shifted depending on the task their subjects performed. In this study, the subjects were assigned two language tasks. In both, subjects were asked to read individual words and then make a spelling or rhyming judgment.

"We found that one network takes different configurations depending on the goal of the task," said Tali Bitan, primary author of "Shifts of Effective Connectivity Within a Language Network during Rhyming and Spelling."

A post-doctoral fellow in the department of communication sciences and disorders, Bitan worked with Associate Professor James Booth of the same department and M-Marsel Mesulam, director of the Cognitive Neurology and Alzheimer's Disease Center in Northwestern's Feinberg School of Medicine.

Mesulam, who was among the first scientists to predict the existence of convergence zones within interconnected brain networks, said the study presents "the clearest and most convincing evidence to date" of the dynamics in effective connectivity.

To better understand dynamic effective connectivity, Mesulam compares the brain networks to a network of highways connecting different parts of a city. The highway is static. No matter how heavy the traffic load, it always has the same number of lanes. In the brain, there is a dynamic change that allows certain pathways to preferentially facilitate the demands of a given cognitive task. The brain highway in effect "adds lanes" to accommodate the requirements of the particular task.

Depending on the goal of the task -- whether subjects were asked to make an orthographic (spelling) judgment or a phonological (rhyming) judgment – the Northwestern researchers found that different convergence zones in the network were involved in the task.

"The existence and the identity of convergence zones --areas in which information from multiple sources meets in the brain -- have been debated since they were proposed in the late 20th century," said Bitan. "Now, with new techniques to analyze brain imaging data, we can examine the specific role played by different brain regions in the network that are required for any cognitive task. These techniques examining effective connectivity enable us to learn how the brain changes its interconnectivity according to the task at hand."

The Northwestern researchers also propose to explain the role of each brain region as it interacts within a complex network to achieve a specific cognitive goal.

The conventional method for analyzing fMRI data, which can only show which brain regions are active in a given task, showed two brain regions that were specifically active for each of the studied tasks: the lateral temporal cortex (LTC) for the rhyming task and the intraparietal sulcus (IPS) for the spelling task.

In addition to the task-specific regions, the inferior frontal gyrus (IFG) and the fusiform gyrus (FG) were engaged by both tasks. Dynamic Causal Modeling, the new method examining the influences between brain regions, indicates that each task preferentially strengthened the influences converging on the task specific regions (LTC for rhyming, IPS for spelling). This finding suggests that task specific regions serve as convergence zones that integrate information from other parts of the brain.

The results also show that switching between tasks -- in this case between rhyming and spelling -- led to changes in the influence of the IFG on the task specific regions. This finding suggests the IFG plays a pivotal role in "making" task specific regions more or less sensitive, depending on the task.

"Previous studies showed that the IFG is active in many different language tasks and suggested that the IFG was involved not only in the integration process but also in control of other brain regions," Bitan said. "Our study corroborates the role of the IFG in modulating other brain regions. In contrast, however, it shows that the integration process is done primarily in the task-specific regions."

In the 19th and early 20th century, scientists with a "localizationist" approach postulated that discrete brain regions were associated with specific functions of language and memory. By the end of the 20th century, a "connectionist" view stressing the importance of interconnected networks became the consensus.

The research presented in the Journal of Neuroscience effectively sets the stage for further development in our understanding of neuroscience. In their article, the Northwestern scientists provide evidence of the ways in which different cognitive goals are achieved from the interaction between different brain regions.

In addition to Bitan, Booth and Mesulam, co-authors of the article are Janet Choy and Douglas Burman of Northwestern's communication sciences and disorders department and Darren Gitelman, associate professor of neuology at Northwestern University Feinberg School of Medicine.