Network INtegration

Assumptions of the EFEC model 

One central prediction of the EFEC model  is that functional integration of information stored in different components of the brain network occurs through temporal binding. This prediction was based on evidence on the existence of temporal binding mechanisms to explain holistic subjective experiences of percepts especially within the visual system (Engel & Singer, 2001: http://www.sciencedirect.com/science/article/pii/S1364661300015680; Tononi & Edleman 1998: http://www.sciencemag.org/content/282/5395/1846.short). Further, the relevance of temporal binding across distributed networks to integration of word meaning has been discussed and tested  by Pulvermueller et al. (Pulvermueller, 1999: http://www.ncbi.nlm.nih.gov/pubmed?term=F.Pulvermuller%20(1999)%20Words%20in%20the%20Brain's%20Language%2C%20Behavioral%20and%20Brain%20Sciences%2C%2022%2C%20253-336).

Studies testing network integration

Green et al., (2010) http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2941398/ used the time course of fMRI signal in a right superior anterior temporal seed region while healthy participants experience guilt, indignation (according to post-scan ratings), or fixate a visual pattern. This anterior temporal lobe signal is then used to see in which other brain regions there are correlated increases or decreases in signal across the time course of the experiment ("physiological" correlation). Further, we tested whether this temporal correlation changes its strength depending on the type of experienced moral feeling (guilt or indignation). This is called a "psychophysiological interaction (PPI)" and is seen as an index of functional integration between two brain regions that is selective for a particular experimental condition (Friston et al., 1997: http://www.fil.ion.ucl.ac.uk/spm/doc/papers/karl_ppi.pdf). We were able to confirm the prediction that while participants felt guilt relative to indignation, functional integration was stronger between the right superior anterior temporal lobe region, previously shown to represent Social conceptual knowledge, and the septal / subgenual cingulate area, previously shown to be selectively activated for guilt vs. indignation. Reversely, while participants felt indignation, the same anterior temporal region showed stronger functional integration with the lateral orbitofrontal cortex, previously shown to be more strongly activated for indignation vs. guilt (Zahn et al., Cer.Cor. 2009, free pdf: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2733324/pdf/bhn080.pdf).

Synthesis

This study provides evidence for temporal integration of neural activation in anterior temporal-frontolimbic networks while participants experience moral feelings. This finding was, however, obtained at a slow time scale (the fMRI scanner provided images every two seconds).  Another conundrum of this study was that content-specific network integration was not consistent across subjects, but could only be observed when modelling individual differences in guilt- and indignation-proneness during the experiment. We have hypothesized that individuals with low levels of functional integration across the network are more prone to overgeneralizations of moral feelings (Green et al., 2010). Since functional integration could be selectively disrupted for different types of moral feelings, this study provides a mechanism by which one could explain selective overgeneralization of guilt relative to indignation, postulated as a central feature of classical forms of major depressive disorder. This prediction has since been confirmed (Green et al. 2012; Lythe et al. 2015), but these two studies threw up another conundrum. Namely, that self-blame-selective over- and under-connectivity of temporo-limbic networks occur in different network parts and may be differentially associated with primary vulnerability, resilience or vulnerability to recurrence.

Future studies

Electrophysiological measures (e.g. EEG) of temporal binding are needed to further test the network integration hypothesis derived from the EFEC model (this was one of the aims of Jennifer Gethin's PhD, co-supervised by Wael El-Deredy) showing some interesting changes in theta power for self-blaming emotions in medication-free remitted MDD, but with small effect sizes, probably due to the difficulty in using EEG to assess such complex cognitive functions.