If two associations share an item, one may be remembered at the expense of the other (BC recalled but not AB). Here, we identify the neural processes by which this competition materializes and is resolved. We analyzed fMRI signal while participants studied sets of pairs that reliably induced pair-to-pair associative interference, but which participants could not fully resolve. Precuneus activity tracked retrieval of previous pairs during study of later overlapping pairs. This retrieval apparently produced interference by diverting study resources from the currently displayed pair. However, when activity in ventromedial prefrontal cortex, as well as anterior subregions of the hippocampus, was present while the earlier pair had been studied, interference was reversed, and both pairs were likely to be recalled. Angular gyrus and mid-frontal activity were related to interference resolution once the participant had seen both pairs. Taken together, associations compete via precuneus-mediated competitive retrieval, but ventromedial prefrontal cortex may neutralize this by ensuring that when the earlier association is remembered while studying the later pair, memories of the two pairs can overcome interference likely via activity in mid-frontal cortex and angular gyrus.

Prevailing theories of implicit or unaware learning propose a developmental invariance model, with implicit function maturing early in infancy or childhood despite prolonged improvements in explicit or intentional learning and memory systems across childhood. Neuroimaging studies of adult visuomotor sequence learning have associated fronto-striatal brain regions with implicit learning of spatial sequences. Given evidence of continued development in these brain regions during childhood, we compare implicit sequence learning in adults and 7- to 11-year-old children to examine potential developmental differences in the recruitment of fronto-striatal circuitry during implicit learning. Participants performed a standard serial reaction time task. Stimuli alternately followed a fixed 10-step sequence of locations or were presented in a pseudorandom order of locations. Adults outperformed children, achieving a significantly larger sequence learning effect and showing learning more quickly than children. Age-related differences in activity were observed in the premotor cortex, putamen, hippocampus, inferotemporal cortex, and parietal cortex. We observed differential recruitment of cortical and subcortical motor systems between groups, presumably reflecting age differences in motor response execution. Adults showed greater hippocampal activity for sequence trials, whereas children demonstrated greater signal during random trials. Activity in the right caudate correlated significantly with behavioral measures of implicit learning for both age groups, although adults showed greater signal change than children overall, as would be expected given developmental differences in sequence learning magnitude. These results challenge the idea of developmental invariance in implicit learning and instead support a view of parallel developments in implicit and explicit learning systems.

In recent years, three attentional networks have been defined in anatomical and functional terms. These functions involve alerting, orienting, and executive attention. Reaction time measures can be used to quantify the processing efficiency within each of these three networks. The Attention Network Test (ANT) is designed to evaluate alerting, orienting, and executive attention within a single 30-min testing session that can be easily performed by children, patients, and monkeys. A study with 40 normal adult subjects indicates that the ANT produces reliable single subject estimates of alerting, orienting, and executive function, and further suggests that the efficiencies of these three networks are uncorrelated. There are, however, some interactions in which alerting and orienting can modulate the degree of interference from flankers. This procedure may prove to be convenient and useful in evaluating attentional abnormalities associated with cases of brain injury, stroke, schizophrenia, and attention-deficit disorder. The ANT may also serve as an activation task for neuroimaging studies and as a phenotype for the study of the influence of genes on attentional networks.