The Neural Dynamics of Fronto-Parietal Networks in Childhood Revealed using Magnetoencephalography

Abstract

Our ability to hold information in mind is limited, requires a high degree of cognitive control, and is necessary for many subsequent cognitive processes. Children, in particular, are highly variable in how, trial-by-trial, they manage to recruit cognitive control in service of memory. Fronto-parietal networks, typically recruited under conditions where this cognitive control is needed, undergo protracted development. We explored, for the first time, whether dynamic changes in fronto-parietal activity could account for children's variability in tests of visual short-term memory (VSTM). We recorded oscillatory brain activity using magnetoencephalography (MEG) as 9- to 12-year-old children and adults performed a VSTM task. We combined temporal independent component analysis (ICA) with general linear modeling to test whether the strength of fronto-parietal activity correlated with VSTM performance on a trial-by-trial basis. In children, but not adults, slow frequency theta (4-7 Hz) activity within a right lateralized fronto-parietal network in anticipation of the memoranda predicted the accuracy with which those memory items were subsequently retrieved. These findings suggest that inconsistent use of anticipatory control mechanism contributes significantly to trial-to-trial variability in VSTM maintenance performance.

Keywords: cognitive control; cognitive development; development; executive control; magnetoencephalography.

Figure 1.

(A) Trial schematic showing an example of a VSTM Load-4 target present trial, with the timings for both children and adults. (B) The behavioral data, shown as d′ values for the 2 groups, across the 2 VSTM load conditions. The error bars refer to the standard error of the mean.

Figure 2.

The right lateralized fronto-parietal identified using the ICA and GLM in both adults (upper panel) and children (lower panel). The left hand images show the spatial extent of the component networks (in terms of the absolute Pearson Correlation values between each brain location and this component); the right hand images show the time course of the GLM contrast for these networks. The black line reflects the effect of VSTM load, and the cyan line reflects the effect of subsequent accuracy.

Figure 3.

The supplementary analysis performed on the children's data. Activity in the right fronto-parietal network was used as a continuous trial-wise regressor, the effect of which upon the subsequent processing of the memoranda can be seen here.

Figure 4.

A set of additional components produced by the temporal ICA that were present in both adults and children. (A and B) Lateralized motor-visual networks that were right and left lateralized, respectively. (C and D) Lateralized frontal-visual networks that were right and left lateralized, respectively. As in Figure 2, the spatial maps are produced by calculating the absolute Pearson Correlation values between each brain location and each component.