Winning the Lottery With Neural Connectivity Constraints: Faster Learning Across Cognitive Tasks With Spatially Constrained Sparse RNNs.

Recurrent neural networks (RNNs) are often used to model circuits in the brain and can solve a variety of difficult computational problems requiring memory, error correction, or selection (Hopfield, 1982; Maass et al., 2002; Maass, 2011). However, fully connected RNNs contrast structurally with their biological counterparts, which are extremely sparse (about 0.1%). Motivated by the neocortex, where neural connectivity is constrained by physical distance along cortical sheets and other synaptic wiring costs, we introduce locality masked RNNs (LM-RNNs) that use task-agnostic predetermined graphs with sparsity as low as 4%. We study LM-RNNs in a multitask learning setting relevant to cognitive systems neuroscience with a commonly used set of tasks, 20-Cog-tasks (Yang et al., 2019). We show through reductio ad absurdum that 20-Cog-tasks can be solved by a small pool of separated autapses that we can mechanistically analyze and understand. Thus, these tasks fall short of the goal of inducing complex recurrent dynamics and modular structure in RNNs. We next contribute a new cognitive multitask battery, Mod-Cog, consisting of up to 132 tasks that expands by about seven-fold the number of tasks and task complexity of 20-Cog-tasks. Importantly, while autapses can solve the simple 20-Cog-tasks, the expanded task set requires richer neural architectures and continuous attractor dynamics. On these tasks, we show that LM-RNNs with an optimal sparsity result in faster training and better data efficiency than fully connected networks.

Identification of the pore-forming and binding domains of the Sneathia vaginalis cytopathogenic toxin A.

The association between Sneathia vaginalis and preterm birth is emerging. The Gram-negative anaerobe produces a large exotoxin, the cytopathogenic toxin A (CptA), that forms pores in human epithelial cells and red blood cells. The structure of the toxin has not been determined, but in silico analysis predicts that a large amino-terminal region of the protein is globular and separated from the carboxy-terminal tandem repeats by a disordered region. We found that a recombinant protein consisting of the predicted structured amino-terminal portion of CptA and devoid of the repeat region was sufficient to permeabilize epithelial cells and red blood cells. The repeat region was capable of binding to epithelial cells but did not permeabilize them or lyse red blood cells. CptA is the only S. vaginalis virulence factor that has been examined mechanistically to date, and this analysis sets the foundation for an understanding of how this novel pore-forming toxin exerts its activity.

The Effects of Prime Supporters within a College Student's Support Network.

College students face academic and emotional challenges within and outside of the classroom. There is not a unique way to deal with all of these challenges, however obtaining support from other individuals in their network can be an important factor in their success in dealing with these challenges. Exploratory social support network data from 38 undergraduate students were collected and analyzed to find common network structures and examine the relationship between these structures and student characteristics, GPA, perceived social support, and use of other help resources. We found half of the students had what we defined to be a novel network characteristic, prime supporters, which are individuals who provided academic and emotional support to students for both routine and intense academic and emotional problems. We found students with a prime supporter in their social network tended to have higher GPAs and perceived social support than those without a prime supporter. Students with prime supporters also had differences in academic help-seeking behavior. These findings suggest the need for further research on the social networks of college students, particularly the presence of individuals who are providers of multiple sources of support.