Extended Data Fig. 2: Graph network of topic model.
From: Collective memory shapes the organization of individual memories in the medial prefrontal cortex
To illustrate the semantic model and the connections between words derived from our topic models, we computed a Lemmas x Lemmas correlation matrix using the estimated distribution of probabilities over 10 topics. This correlation matrix was then thresholded and transformed into a binary adjacency matrix by keeping the top 10% of the strongest connections between lemmas. The adjacency matrix is visualized here using a force vector algorithm proposed with the Gephi software (https://gephi.org/). Each node represents one of the 6,240 lemmas. The color and the size of the node is determined by its maximal topic assignment and probability, respectively. Amongst the 6,240 nodes, only key words describing Memorial pictures and specifically translated into English, are displayed here for visualization purpose. The size of the label is proportional to its topic probability. The distribution of topic probabilities for each Memorial pictures is directly derived from the topic probabilities associated with these key words.
