Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 May 25;7(1):635.
doi: 10.1038/s42003-024-06309-z.

Higher surface folding of the human premotor cortex is associated with better long-term learning capability

Marco Taubert  1   2   3 , Gabriel Ziegler  4   5   6 , Nico Lehmann  7   4   8
Affiliations

Higher surface folding of the human premotor cortex is associated with better long-term learning capability

Marco Taubert et al. Commun Biol. .

Abstract

The capacity to learn enabled the human species to adapt to various challenging environmental conditions and pass important achievements on to the next generation. A growing body of research suggests links between neocortical folding properties and numerous aspects of human behavior, but their impact on enhanced human learning capacity remains unexplored. Here we leverage three training cohorts to demonstrate that higher levels of premotor cortical folding reliably predict individual long-term learning gains in a challenging new motor task, above and beyond initial performance differences. Individual folding-related predisposition to motor learning was found to be independent of cortical thickness and intracortical microstructure, but dependent on larger cortical surface area in premotor regions. We further show that learning-relevant features of cortical folding occurred in close spatial proximity to practice-induced structural brain plasticity. Our results suggest a link between neocortical surface folding and human behavioral adaptability.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Behavioral data.
Motor learning task, performance improvements, performance stabilization and increased inter-individual differences in motor learning over 6 practice sessions. a We tested motor learning of a challenging whole-body balancing task. Participants were instructed to keep a seesaw-like moving stabilometer balance platform in a horizontal target interval (±3°) as long as possible during a trial length of 30 s. b Motor performance was measured as the time (in seconds) in which participants kept the board within the ±3° target interval in each of 15 practice trials per session (see Supplementary Video files for motor performance of participants at the beginning and end of practice). c Decrease in trial-to-trial variability (coefficient of variation, COV) of session-specific motor performance. d Increase of the interquartile range (IQR) of session-specific between-person variation in motor performance. IQR increased from 3.7 s at session 1 to 8.7 s at session 6. e From the first to the sixth session, participants tended to maintain their performance rank (Spearman correlation between initial and final performance, R2 = 0.238, p < 0.001) but there were large individual differences in learning (blue/yellow: higher/lower performance than predicted from baseline).
Fig. 2
Fig. 2. Cortical folding is related to motor learning.
Results of whole-brain correlation of vertex-wise cortical curvature and learning rate (N = 84). a Uncorrected results at p < 0.001 (left) and family-wise error-corrected results at p < 0.05 (inset) were projected onto a template showing surface variations in sulcus depth. b Association between cortical folding (in the cluster representing the FWE-corrected effect in the original analysis [A]) and learning rate (displayed for visualization of the range of individual values only and not for inference). c Subsample results in the three independent learning experiments (displayed for visualization of the range of individual values only and not for inference, detailed information on sub-samples in Supplementary Table 1). d Structural equation model depicting relationships between cortical folding in pre-SMA/SMA (cluster from 2 A, unadjusted for a), learning rate (adjusted for a) and final performance on session 6 (unadjusted for a). Results of a separate analysis of final performance are depicted in Supplementary Fig. 8. Standardized coefficients with 95% bootstrapped confidence intervals (CI) are represented on paths. e Pearson correlations between cortical folding and motor performance. Gray bars represent session-specific performance controlled for initial performance in session 1 (i.e., residual gain) and black bars represent correlations with actual session-specific performance. * indicate significant paths at p < 0.05 (with CIs not including zero).
Fig. 3
Fig. 3. Cortical surface area, but not cortical thickness, is related to the effect of cortical folding on learning.
Interrelationship between folding, thickness and surface area. a SEM depicting the relationships between cortical folding (‘folding’), cortical surface area (‘surface’), cortical thickness (‘thickness’), and learning rate n (‘learning’) in the left caudal superior frontal gyrus ROI. Standardized coefficients with 95% bootstrapped CIs are represented on paths. Correlations between average folding index and surface area in the ROI with learning rate n. Folding index is either unadjusted (b) or adjusted (c) for differences in surface area and cortical thickness. Note that all variables used in the model and for correlation analyses were corrected for differences in age, gender, height, study, head coil, baseline performance, and total intracranial volume. * indicate significant paths/correlations at p < 0.05 (with CIs not including zero).
Fig. 4
Fig. 4. Cortical folding ties to learning rates independent of cortical myelination and cortical neurite density.
Analysis of microstructural tissue properties of the premotor cortex. Distribution of myelin-sensitive magnetization transfer saturation (MT) values (a) and the neurite density index NDI (b) across the left hemisphere. Color bars show regions of high MT or NDI in red (e.g., primary motor and somatosensory cortices) and regions of lower MT and NDI in blue (e.g., anterior prefrontal regions). Note the MT product-sequence-specific representation of MT values with a factor of 2. c MT and NDI values were analyzed in pre-SMA/SMA, the cluster in which cortical folding positively correlated with learning rate n (Fig. 2a). d Pearson correlations between MT in superficial gray matter (GM), deep GM, and cortex-adjacent white matter with chronological age. e Pearson correlations between MT in superficial GM, deep GM, and cortex-adjacent white matter with learning rate n. f Partial correlations between cortical folding and learning rate adjusted for MT in superficial GM, deep GM, and cortex-adjacent white matter. * indicate significant correlations at p < 0.05, while ns indicates no significant correlation.
Fig. 5
Fig. 5. Cortical folding is associated with learning in regions undergoing practice-induced structural plasticity.
Relationship between cortical folding and plasticity in the premotor cortex. a The clusters of significant learning-induced gray matter changes (left) that overlapped with positive correlation of cortical curvature in pre-SMA/SMA and learning rate (right). b, c Whole-sample and sub-sample correlations between learning rate and cortical curvature in the pre-SMA/SMA cluster in A. d SEM depicting the relationship between cortical folding in pre-SMA/SMA, learning rate (adjusted for a) and final performance on session 6. Standardized coefficients with 95% bootstrapped confidence intervals (CI) are represented on paths. e Pearson correlation coefficients between residualized cortical folding and motor performance. Gray bars represent session-specific performance controlled for initial performance in session 1 (i.e., residual gain) and black bars represent correlations with actual session-specific performance. * indicate significant correlations/paths at p < 0.05 (with CIs not including zero).
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Grasby, K. L. et al. The genetic architecture of the human cerebral cortex. Science36710.1126/science.aay6690 (2020).
    1. Fornito A. Individual differences in anterior cingulate/paracingulate morphology are related to executive functions in healthy males. Cereb. Cortex. 2004;14:424–431. doi: 10.1093/cercor/bhh004. - DOI - PubMed
    1. Hopkins WD, Cantalupo C, Taglialatela J. Handedness is associated with asymmetries in gyrification of the cerebral cortex of chimpanzees. Cereb. Cortex. 2007;17:1750–1756. doi: 10.1093/cercor/bhl085. - DOI - PMC - PubMed
    1. Llinares-Benadero C, Borrell V. Deconstructing cortical folding: genetic, cellular and mechanical determinants. Nat. Rev. Neurosci. 2019;20:161–176. doi: 10.1038/s41583-018-0112-2. - DOI - PubMed
    1. Miller JA, Weiner KS. Unfolding the evolution of human cognition. Trends Cogn. Sci. 2022;26:735–737. doi: 10.1016/j.tics.2022.06.008. - DOI - PubMed