Networks describing how brain regions interact with each other during the expression of a behavior (i.e, a functional connectome) are thought to capture the unique state of brain activation during task performance. Although it is often assumed that the network based properties of individual brain regions in functional connectomes are predictive of their importance for expression of the associated behavior, a systematic study of this relationship has not been undertaken.
To probe the relationship between brain network properties and behavior we examined a functional connectome describing the interactions of 84 brain regions during recall of a fear memory in mice (3). Initially, we developed a computational model inspired by how the brain is known to interact, and applied it to the fear memory functional connectome to make predictions about the importance of specific regions for memory recall (2). We then compared the model output to the behavioral effect of independently inactivating 21 brain regions and found that the model successfully predicted our experimental results. We also found that neuronal activity data alone or networks derived from anatomical connections were not predictive of behavior, suggesting that functional networks contain unique task-specific information.
In a related paper, we identified functional connectomes that underlie various aspects of social recognition memory in mice (1). Inhibition of protein synthesis in brain regions such as the anterior cingulate cortex, amygdala, and hippocampus prevents the establishment of long-term social recognition memory. A network based approach provided nuance to these experimental observations by identifying the anterior cingulate cortex and amygdala as important during the initiation of social interaction and the hippocampus as key to stabilizing long-term social memory.
Taken together, these findings establish and demonstrate the power of utilizing network concepts for understanding neuronal function. As the tools to process and analyze whole brain activity in animal models increases in prevalence in the coming years, our work provides important experimental evidence supporting network based approaches for gleaning insight into the interplay between brain activity and behavior.
Video: We developed a model to capture the spread of brain region inactivation beyond its immediate neighborhood using a fear memory functional connectome. As connections fall below a threshold, they are removed from the network and resulting global network properties are re-calculated. Examples of the model and associated behavioral output for two regions are shown (re: reunions thalamic nucleus, S2: secondary somatosensory cortex). Overall, we found that the model output was predictive of the behavioral results (r = -0.61).
References
1. Functional Connectivity of Multiple Brain Regions Required for the Consolidation of Social Recognition Memory. Tanimizu T, Kenney JW, Okano E, Kadoma K, Frankland PW, Kida S. The Journal of Neuroscience 37: 4103–4116, 2017.
2. Chemogenetic Interrogation of a Brain-wide Fear Memory Network in Mice. Vetere G*, Kenney JW*, Tran LM, Xia F, Steadman PE, Parkinson J, Josselyn SA, Frankland PW. Neuron 94: 363–374.e4, 2017.
Other References
3. Identification of a functional connectome for long-term fear memory in mice. Wheeler AL, Teixeira CM, Wang AH, Xiong X, Kovacevic N, Lerch JP, McIntosh AR, Parkinson J, Frankland PW. PLoS Computational Biology 9: e1002853, 2013.