The effects of multi-task learning and time-varying hemispheric asymmetry on lateralisation in a neural network model

While various functional and cognitive capabilities appear to differ in both degree and direction of lateralisation, the factors underlying these differences are poorly understood. It is hypothesised that time-varying asymmetry in plasticity between homologous regions in the cerebral hemispheres, coupled with asynchronous development of capabilities, may account for the lateralisation differences in two ways. First, the lateralisation of an earlier acquired behaviour may influence the lateralisation of a later developing behaviour. Second, temporal changes in the underlying plasticity asymmetry may also result in differences in lateralisation for functions acquired at different times. This study examines the plausibility of these hypotheses using a computational neural network model consisting of two interacting hemispheric regions and capable of learning two tasks. Lateralisation was measured while learning rates for the two hemispheres were changed independently over time to create a time-varying asymmetry in plasticity, and while the initiation of learning for each task was also varied over time. The results suggest that the lateralisation of one behaviour/function can affect the lateralisation of another in a fundamental way, and that experimentally observed temporal differences in hemispheric development can result in functional lateralisation differences, providing support for past theories of lateralisation based on asymmetric hemispheric growth and maturation.