Cortical network properties predict language learning ability

Findings by Sheppard et al. published in the Journal of Cognitive Neuroscience, reveal interesting brain functional network properties that make it easier for some to learn words of a new language. The authors of this study used functional magnetic resonance imaging to map brain hemodynamic responses of a group of volunteers during a pitch discrimination task. Subsequently, the volunteers participated in another experiment where they were to learn words of an artificial (spoken) language. Interestingly, results of network analysis of brain hemodynamic data obtained during the pitch discrimination task predicted individual differences in spoken language learning ability.

Brain networks were analyzed by the authors by reconstructing the cortical surface of each subject and by dividing the cortex into ~1000 nodes. Person’s correlation coefficients were then calculated between hemodynamic response time series of each of the nodes and correlations exceeding a certain threshold value were considered as a functional connection between two nodes. Network analysis across all the nodes revealed differences between successful and less successful learners. Successful learners had higher global efficiency, meaning that there were, on the average, fewer edges separating the nodes of their cortical networks from each other. On the other hand, local efficiency measure was higher in the less successful learners, suggesting that their local network connectivity was higher than in successful learners. When analyzed across specific anatomical regions, it was further observed that these network differences could be observed in prefrontal and parietal cortical areas bilaterally as well as in the right temporal cortex.

Network analysis offers a powerful alternative method that complements the more traditional functional neuroimaging data analysis methods. With a network analysis it can be effectively measured how cortical areas work together to give rise to perceptual and cognitive functions. In this particular study, it was very nicely observed that network properties of brain function predicted language learning capability, and I anticipate that we will see in the near future a wealth of highly interesting findings in cognitive neuroscience that are based on network analysis methodology. Furthermore, it would be interesting to see whether cortical functional network properties differ between healthy individuals and those suffering from language disorders such as dyslexia.

Reference: Sheppard JP, Wang JP, Wong PC. Large-scale cortical network properties predict future sound-to-word learning success. Journal of Cognitive Neuroscience (2012) 24: 1087-1103. http://dx.doi.org/10.1162/jocn_a_00210