The Resilience of the Phonological Network May Have Implications for Developmental and Acquired Disorders

Abstract

A central tenet of network science states that the structure of the network influences processing. In this study of a phonological network of English words we asked: how does damage alter the network structure (Study 1)? How does the damaged structure influence lexical processing (Study 2)? How does the structure of the intact network “protect” processing with a less efficient algorithm (Study 3)? In Study 1, connections in the network were randomly removed to increasingly damage the network. Various measures showed the network remained well-connected (i.e., it is resilient to damage) until ~90% of the connections were removed. In Study 2, computer simulations examined the retrieval of a set of words. The performance of the model was positively correlated with naming accuracy by people with aphasia (PWA) on the Philadelphia Naming Test (PNT) across four types of aphasia. In Study 3, we demonstrated another way to model developmental or acquired disorders by manipulating how efficiently activation spread through the network. We found that the structure of the network “protects” word retrieval despite decreases in processing efficiency; words that are relatively easy to retrieve with efficient transmission of priming remain relatively easy to retrieve with less efficient transmission of priming. Cognitive network science and computer simulations may provide insight to a wide range of speech, language, hearing, and cognitive disorders.

Using network science to examine audio-visual speech perception with a multi-layer graph

“…Decay (d) refers to the proportion of activation lost at each time step. This parameter ranges from 0 to 1 and was set to 0 in the simulations reported here to be consistent with the parameter settings used previously (e.g., [41]).…”

“…The dispersion of activation in such networks is more similar to simple diffusion models in physics than to the more complicated spreading activation models often seen in the linguistic and cognitive sciences, which might include facilitatory as well as inhibitory connections, decay of activation (although see the d parameter above), unidirectional spreading of activation, various thresholds, nodes with different resting activation levels, etc. Despite the simplicity of the diffusion of activation over the type of network employed in the present study, quite complex behaviors have been computationally reproduced via simulations (e.g., [39][40][41]).…”

Some challenges in using multilayer networks to bridge brain and mind.

“…An example of different network architectures having different properties can be seen in the structure of two computer networks: the internet is resilient to random damage but is vulnerable to attacks that target highly connected nodes (Albert et al, 2000), whereas the darknet (the internet of illegal activity), which has a slightly different structure, is resilient to both random damage and targeted attacks (De Domenico & Arenas, 2017). Another example can be found in the semantic network, which has a structure similar to the internet (Steyvers & Tenenbaum, 2005), and the phonological network, which has a structure similar to the darknet (Arbesman et al, 2010; Vitevitch et al, 2023). The potential differences in the structure of adjacent layers in a network of networks also suggest that the structure of one network may not inform us about the structure of an adjacent network.…”

Using Complex Networks in the Hearing Sciences

“…Recent work by Vitevitch et al (2023) manipulated several parameters in a computer simulation in a network model, or removed connections between nodes to simulate various speech, language, and hearing disorders. Similar approaches might help clinical audiologists predict how the perception of certain words might be affected at different levels of hearing loss (see the phoneme specific sentences developed by Huggins & Nickerson 1985).…”