Cognitive Skills
Research in this domain follows three strands. First, we have been analyzing data from the FLEX study, a large training study on task switching in childhood planned and conducted in collaboration with Yana Fandakova and Silvia Bunge (Dissertation and postdoctoral work by Sina Schwarze). Second, following up on the results of the FLEX study, we are currently investigating the ontogeny of hierarchical control in childhood. Third, we investigated how brain maturation shapes learning in the context of motor sequence learning.
Plasticity of task switching in middle childhood
Childhood is characterized by increased proficiency in a variety of skills and cognitive abilities supported by maturational changes in brain structure and function. During middle and late childhood, such changes are particularly pronounced for cognitive control processes, including the ability to flexibly switch between different tasks (Schwarze et al., 2024). To explore how these maturational changes interact with training, we conducted the FLEX study, an extensive training study of task-switching performance in children aged 8 to 11 years (Figure 1). The study has been planned by Silvia Bunge (University of California at Berkeley), Yana Fandakova (now University of Trier), and Ulman Lindenberger, and has been conducted at the MPIB by Yana Fadankova in the context of a priority program funded by the German Research Foundation during her leadership of the Plasticity project.
We first examined age differences between children and adults before training, by looking at task-specific activation patterns and functional connectivity. With increased task-switching demands children showed less upregulation of brain activation but a larger increase in connectivity between the inferior frontal junction (IFJ), a key task-switching region, and lateral prefrontal cortex (lPFC). This increased connectivity might offer an alternative and possibly developmentally earlier mechanism to manage task-switching demands (Schwarze et al., 2023). Additionally, we examined task-related neural representations as captured by multivariate patterns of brain activation, showing that the distinctiveness of neural representations of task information did not differ between children and adults, suggesting a remarkable degree of maturity of neural representations of task-relevant information in late childhood (Schwarze, Bonati, et al., 2025).
The main aim of the FLEX study was to delineate patterns of neural and behavioral plasticity in the course of training. Children in the experimental group practiced switching between a large number of different task sets over the course of 9 weeks with a total of 27 training sessions. Children in the active control group trained the identical tasks but at a markedly lower switching frequency. Both groups showed initial decreases in the costs of performing tasks intermixed, suggesting faster processing of task demands when multiple task rules had to be maintained, monitored, and reconfigured. However, only children in the experimental group, who were exposed to higher switching demands, maintained these changes in the second half of training. These behavioral changes were accompanied by reduced activations in the dorsolateral prefrontal cortex in the experimental group (Figure 2). Results suggest that task-switching training enhances the efficiency of regions that support task switching in children, rather than moving children’s task-related activation more rapidly towards an adult-like pattern (Schwarze, Laube, et al., 2025).
Previous studies have shown that training outcomes vary greatly between individuals of all ages, pointing to the need to understand why some individuals benefit more from training than others. Thus, a further set of analyses was aimed at understanding how individual differences, specifically differences in brain development as captured by the organization of functional brain networks might contribute to individual differences in training outcomes. The organization principle of network modularity, defined as the extent to which brain regions are more strongly connected to regions within the same functional subnetwork than to regions outside of the subnetwork, was of particular interest here as it has been shown to predict training outcomes in adults. We observed a similar pattern in our group of children: Children with more modular network organization before training showed earlier and greater effects of training than children with less modular network organization across the two training groups. These results suggest that more modular networks allow for faster adaptation to training demands. Importantly, while children showed lower network modularity overall than adults, network modularity was not associated with age within the group of children, indicating that modularity might provide additional information of maturation beyond a child’s age. Finally, ongoing analyses examine structural white matter connections in the brain that might be an additional factor contributing to individual differences in training outcomes.
The neural organization of cognitive control in children.
One challenge in examining plastic changes with cognitive training is that mechanisms supporting cognitive control are implemented in a widespread network of brain regions. Hence, it is not quite clear whether and where we should expect training-related local changes in brain structure in accordance with the EESR theory when taxing children’s cognitive control abilities. To approach this problem in a systematic fashion, we are currently examining the development of hierarchical organization of the lateral prefrontal cortex (lPFC) during a cognitive control paradigm in 10–13-year-olds. Studies in adults have shown that hierarchically higher levels of control (e.g., involving integration with previous or future task requirements) recruit more anterior regions of the lPFC, while hierarchically lower levels of control (e.g., involving stimulus-response mappings) recruit more posterior lPFC. How this organization develops during childhood is currently unclear but might help to inform where a certain type of cognitive training could elicit changes in brain structure in childhood. Gaining knowledge about maturational changes in the organization of cognitive control during childhood may also inform the design of age-appropriate training programs.
How does brain maturation shape motor skill acquisition?
In daily life, we perform a variety of motor actions such as typing, playing the piano or video gaming. In this project, we asked how the brain supports the acquisition of such skills, and how this might change with brain maturation from childhood to adulthood. Children aged 7 to 10 years and young adults aged 20 to 32 years learned several motor sequences in an associative visuomotor learning task. Both age groups became faster at performing the sequences with learning, with adults improving more.
During the task, we measured brain activity using functional magnetic resonance imaging (fMRI). In both age groups, early learning was associated with heightened prefrontal cortex (PFC) activation, a region important for cognitive control. Later learning was characterized by increased activation in the left primary motor cortex (M1) and bilateral supplementary motor area, regions that are involved in integrating sensory and motor associations and storing a more unified representation of the sequence. While PFC activation and PFC-M1 connectivity decreased similarly with learning in both age groups, adults showed stronger learning-related increases in M1 activation than children. Furthermore, both age groups showed increasing similarity of M1 activation patterns for correct repetitions of the same sequence over time, but this increase was greater in adults.
Together, these findings suggest that while children and adults engage cognitive control regions (like the PFC) similarly during early learning, age‑related differences in how motor regions (like M1) strengthen and stabilize motor representations contribute to more effective motor sequence learning in adulthood.