Concurrent visual learning of adjacent and nonadjacent dependencies in adults and children.

Abstract

Concurrent learning of adjacent and nonadjacent dependencies has been shown in adults only. This study extended this line of research by examining dependency-specific learning for both adjacent and nonadjacent dependencies concurrently in both adults and children. Seventy adults aged 18 to 64 (40 females, 30 males; Experiment 1) and 64 children aged 10 to 11 years (40 girls, 24 boys; Experiment 2) were tested with a new serial reaction time (SRT) task in which they were trained for 6 -8 minutes on materials comprising equally probable adjacent and nonadjacent dependencies. They were then asked to discriminate between trained and untrained dependencies in a familiarity task. Both adults and children showed implicit concurrent learning of both adjacent and nonadjacent dependencies. The two dependency types were learnt to the same extent. However, adults showed a rapid, sustainable and dependency-specific sensitivity throughout the SRT task while children only showed a dependency-specific sensitivity to violations of expectations after exposure. When the two groups were statistically compared, only adults showed a dependency-specific learning effect after exposure. These findings are in line with the age-related improvement model of dependency learning.

Dynamic Motion and Human Agents Facilitate Visual Nonadjacent Dependency Learning

“…Beyond this increase in resource demands, the computational problem increases because there are multiple nonadjacent relationships that the learner could consider. There is also evidence that learners learn adjacent patterns even as they are learning nonadjacent ones (Deocampo, King, & Conway, 2019; Iao, Roeser, Justice, & Jones, 2021; Romberg & Saffran, 2013; Wang & Mintz, 2018). Thus, given that the nonadjacent relationship itself is less constrained than adjacency, and given the correlated increase in resource demands, learning NADs would logically benefit from constraints that limit the domain over which nonadjacent patterns are considered.…”

Incidental Auditory Category Learning and Visuomotor Sequence Learning

“…Or, simultaneous regularities may interacteven interfere -with learning if learning in each domain competes for limited, common cognitive resources, or if the combination of regularities lowers the informational value with which each regularity predicts behavior. There are many demonstrations that adult learners can track multiple statistical regularities simultaneously and use them to guide behavior (Conway & Christiansen, 2006;Conway et al, 2020;Deocampo et al, 2019;Goschke & Bolte, 2012;Iao et al, 2021;Idemaru & Holt, 2014;Jiménez & Vázquez, 2011;Mitchel & Weiss, 2011;Vuong et al, 2016;Zhang & Holt, 2018). But less is known about the extent to which learning across multiple regularities may introduce interference, for example in competition for common cognitive resources.…”