Networks in the parietal and premotor cortices enable essential human abilities regarding motor processing, including attention and tool use. Even though our knowledge on its topography has steadily increased, a detailed picture of hemisphere-specific integrating pathways is still lacking. With the help of multishell diffusion magnetic resonance imaging, probabilistic tractography, and the Graph Theory Analysis, we investigated connectivity patterns between frontal premotor and posterior parietal brain areas in healthy individuals. With a two-stage node characterization approach, we defined the network role of precisely mapped cortical regions from the Julich-Brain atlas. We found evidence for a third, left-sided, medio-dorsal subpathway in a successively graded dorsal stream, referencing more specialized motor processing on the left. Supplementary motor areas had a strongly lateralized connectivity to either left dorsal or right ventral parietal domains, representing an action-attention dichotomy between hemispheres. The left sulcal parietal regions primarily coupled with areas 44 and 45, mirrored by the inferior frontal junction (IFJ) on the right, a structural lateralization we termed as “Broca’s-IFJ switch.” We were able to deepen knowledge on gyral and sulcal pathways as well as domain-specific contributions in parieto-premotor networks. Our study sheds new light on the complex lateralization of cortical routes for motor activity in the human brain. Author Summary Human motor abilities are processed via specialized yet intertwined pathways in the parietal and premotor cortex. These can be parcellated into networks of brain areas, sharing connection patterns that differ between hemispheres. We differentiated a set of graded pathways, connecting definable network hubs. The well-known Dorsal Stream for visuomotor transformation appears to be left-dominantly divisible into three distinct substreams, providing more anatomical detail about its specialization into visual and semantic segments. Supplementary motor areas show lateralized couplings with left dorsal and right ventral parietal areas, respectively, while left-sided AIPS connectivity to area 44/45 is mirrored by right-sided intersulcal links to the inferior frontal sulcus, both deepening our understanding of incorporated multi-task faculties like attention and speech.

Author Summary In recent years, new insights into brain network communication related to cognitive performance differences in older age have been gained. Simultaneously, an increasing number of studies has turned to machine learning (ML) approaches for the development of biomarkers in health and disease. Given the increasing aging population and the impact cognition has on the quality of life of older adults, automated markers for cognitive aging gain importance. This study addressed the classification and prediction power of resting-state functional connectivity (RSFC) strength measures for cognitive performance in healthy older adults using a battery of standard ML approaches. Classifiability and predictability of cognitive abilities was found to be low across analytic choices. Results emphasize limited potential of these metrics as sole biomarker for cognitive aging. Abstract Age-related cognitive decline varies greatly in healthy older adults, which may partly be explained by differences in the functional architecture of brain networks. Resting-state functional connectivity (RSFC) derived network parameters as widely used markers describing this architecture have even been successfully used to support diagnosis of neurodegenerative diseases. The current study aimed at examining whether these parameters may also be useful in classifying and predicting cognitive performance differences in the normally aging brain by using machine learning (ML). Classifiability and predictability of global and domain-specific cognitive performance differences from nodal and network-level RSFC strength measures were examined in healthy older adults from the 1000BRAINS study (age range: 55–85 years). ML performance was systematically evaluated across different analytic choices in a robust cross-validation scheme. Across these analyses, classification performance did not exceed 60% accuracy for global and domain-specific cognition. Prediction performance was equally low with high mean absolute errors ( MAE s ≥ 0.75) and low to none explained variance ( R 2 ≤ 0.07) for different cognitive targets, feature sets, and pipeline configurations. Current results highlight limited potential of functional network parameters to serve as sole biomarker for cognitive aging and emphasize that predicting cognition from functional network patterns may be challenging.