Organizing our behaviors so that they match the demands of a given situation depends on establishing contingencies between specific features of an action and whether they lead to the desired outcome. In many real-world tasks, this is a challenging endeavor as the brain must learn how to modify its actions based on a single measure of performance feedback that reflects overall task success (Berniker and Kording, 2008; Dhawale et al., 2017; Houk et al., 1996; Wolpert et al., 2001). Moreover, this feedback must be communicated to multiple discrete neural systems distributed across the cortex and striatum, many of which are topographically segregated from one another (Averbeck and O’Doherty, 2022). Numerous studies have shown that neural systems anchored within the medial prefrontal cortex (MPFC) and striatum are important in evaluating whether the results of behavior are in line with expectations (Averbeck and O’Doherty, 2022; Klein-Flügge et al., 2022; Lee et al., 2012; O’Doherty et al., 2017). Specifically, when discrepancies arise between the expected versus actual results of an action—termed a ‘prediction error’—this information serves as the teaching signal that can be used to update behavior directly (Bayer and Glimcher, 2005; O’Doherty et al., 2003; Samejima et al., 2005; Schultz et al., 1997). How exactly this information is communicated in a coherent manner across the multiple, distributed neural systems that guide behavior remains poorly understood. Our study addresses this gap in our knowledge using state-of-the-art manifold learning techniques to describe how the landscape of brain activity changes during reward-guided motor learning.

Contemporary systems and cognitive neuroscience have identified many large-scale neural systems that have each been linked to different components of effective behavior. For example, areas in sensory cortex provide representations of the external environment, whereas areas in motor cortex are involved in generating the final motor commands required for action. At higher levels in the cortical hierarchy, regions within the frontoparietal system, along with those that make up the brain’s attention-orienting systems (dorsal and ventral attention systems), are important in the selection of sensory inputs and the guidance of rule-driven behavior (Corbetta et al., 2008; Corbetta and Shulman, 2002). Yet, how the activity of these various brain systems is coordinated during the learning process is unclear. An emerging literature suggests that this coordination may depend, in part, on functional activity in several key regions of higher-order association cortex, known collectively as the default mode network (DMN) (Buckner et al., 2008; Fox and Raichle, 2007; Margulies et al., 2016; Raichle, 2015; Smallwood et al., 2021).

Initially identified through its tendency to deactivate during cognitively demanding tasks, the DMN has traditionally been implicated in largely introspective, abstract cognitive functions such as autobiographical memory and internal mentation (Andrews‐Hanna et al., 2014; Binder et al., 2009; Christoff et al., 2016; Schacter et al., 2012; Spreng et al., 2009). In recent years, however, this characterization of DMN activity has been difficult to reconcile, with an emerging body of functional MRI (fMRI) and neurophysiological evidence showing that areas of this network are activated during demanding decision-making and working-memory tasks (Foster et al., 2023; Hayden et al., 2009; Heilbronner and Platt, 2013; Murphy et al., 2019; Murphy et al., 2018; Pearson et al., 2009; Vatansever et al., 2017). One hypothesis concerning this system’s function pertains to its unique topographic positioning on the cortical mantle (Smallwood et al., 2021): Each core region of the DMN is located in regions of association cortex that are equidistant between different primary systems; for example, posteromedial cortex, a key node of the DMN, is located precisely at the midpoint between the calcarine (visual) and central (motor) sulci (Margulies et al., 2016). This unique topographic location is hypothesized to allow DMN regions broad oversight over distributed brain functions, enabling them to play a role in the coordination of activity across cortex (Smallwood et al., 2021).

Consistent with this contemporary perspective, recent work implicates several regions of the DMN in organizing different modes of behavior over time. For instance, DMN areas such as medial frontal cortex and posteromedial cortex appear to play an important role in shifting between information gathering versus information exploitation during reward-guided decision-making tasks (Barack et al., 2017; Foster et al., 2023; Pearson et al., 2011; Pearson et al., 2009; Schuck et al., 2015; Trudel et al., 2021)—the so-called explore/exploit trade-off (Frank et al., 2009; Sutton and Barto, 2018). Consistent with this, recent studies have argued that broad features of the DMN’s activity can be explained under the auspices that it supports behavior under conditions in which performance depends on knowledge accrued across several trials rather than by immediate sensory inputs (Hayden et al., 2008; Murphy et al., 2019; Murphy et al., 2018; Vatansever et al., 2017). Extending these ideas, we recently showed that the DMN plays a role in motor adaptation, showing that connectivity between DMN regions and sensorimotor cortex is altered when normal visual-motor contingencies governing behavior are interrupted over the course of several trials (Gale et al., 2022). Given that the DMN is hypothesized to exert influence on functional brain activity via its topographic positioning on cortex (Smallwood et al., 2021), understanding how the DMN supports task behavior likely requires analytical techniques that allow for a characterization of whole-brain changes in functional architecture.

In the current study, we explored changes in the landscape of cortical and striatal activity during a reward-based motor task in which human participants learned to produce, through purely reinforcement feedback, a specific movement trajectory that was initially unknown to them. To characterize learning-related changes at the neural level, we leveraged advanced manifold learning approaches that provide a low-dimensional description of cortical activity (Huntenburg et al., 2018; Margulies et al., 2016; Vos de Wael et al., 2020). This approach builds on recent electrophysiological studies in macaques, demonstrating that high-dimensional neural population activity can be described along a low-dimensional subspace or manifold (Cunningham and Yu, 2014; Gallego et al., 2017; Shenoy et al., 2013; Vyas et al., 2020), reflecting covariance patterns across the entire population. This same organizational structure also appears to govern the macroscale activity of cortex, with this manifold approach having recently provided key insights into the overarching structural and functional architecture of the human brain (Huntenburg et al., 2018; Paquola et al., 2019; Shine et al., 2019a; Shine et al., 2019b; Vázquez-Rodríguez et al., 2019). Here, we applied this manifold approach to explore how brain activity across widely distributed cortical and striatal systems is coordinated during reward-based motor learning. We were particularly interested in characterizing how connectivity between regions within the DMN and the rest of the brain changes as participants shift from learning the relationship between motor commands and reward feedback, during early learning, to subsequently using this information, during late learning. We were also interested in exploring whether learning-dependent changes in manifold structure relate to variation in subject motor performance.

Prior studies examining the neural processes underlying reward-based learning have typically used tasks requiring simple motor responses, such as button presses or lever movements (Averbeck and O’Doherty, 2022; Daw et al., 2006; Klein-Flügge et al., 2022; Lee et al., 2012; O’Doherty et al., 2017; Rushworth et al., 2011). The simplicity of these motor responses is intended to isolate participants’ choice behavior by eliminating any variability related to movement execution (i.e., motor implementation of the choice) as a potential confounding factor in the learning process (McDougle et al., 2019; McDougle et al., 2016). However, recent theories on learning, supported by both human and animal studies (Dhawale et al., 2019; Dhawale et al., 2017; Wu et al., 2014), have highlighted the crucial role of movement variability—and motor exploration in particular—as a key ingredient for effective learning (Dam et al., 2013; Wilson et al., 2021; Wu et al., 2014). In order to incorporate this aspect to learning, and inspired by recent work in the field (Dam et al., 2013; Wu et al., 2014), we developed an MRI-compatible reward-based motor task in which human participants (N = 36) learned to shape their hand trajectories purely through reinforcement feedback.

In this task, subjects used their right finger on an MRI-compatible touchpad to trace, without visual feedback of their finger, a rightward-curved path displayed on a screen (Figure 1A and B). Participants began the MRI study by performing a Baseline block of 70 trials, wherein they did not receive any feedback about their performance. Following this, subjects began a separate Learning block of 200 trials in which they were told that they would now receive score feedback (from 0 to 100 points), presented at the end of each trial, based on how accurately they traced the visual path displayed on the screen. However, unbeknownst to subjects, the score they actually received was based on how well they traced a hidden mirror-image path (the ‘reward’ path, which was reflected across the vertical axis; Figure 1C). Importantly, because subjects received no visual feedback about their actual finger trajectory and could not see their own hand, they could only use the score feedback—and thus only reward-based learning mechanisms—to modify their movements from one trial to the next (Dam et al., 2013; Wu et al., 2014). That is, subjects could not use error-based learning mechanisms to achieve learning in our study as this form of learning requires sensory errors that convey both the change in direction and magnitude needed to correct the movement.

Figure 1 with 2 supplements see all

Download asset Open asset

Figure 1C shows an example of a single subject’s finger trajectories across trials. Initially, the subject begins by tracing the visual path displayed on the screen (as instructed), albeit with some expected motor noise due to the absence of any visual feedback about their finger paths (see cyan trajectories). However, over time, the subject learns to gradually trace a path more similar to the rewarded, mirror-image path (dark pink trajectories). As can be seen in Figure 1D, subjects on average were able to use the reward-based feedback to increase their score, and thus produce a trajectory more similar to the hidden path, over the 200 learning trials (see Figure 1—figure supplement 1 for other measures of changes in subject motor behavior throughout learning, and Figure 1—figure supplement 2 for examples of all subjects’ finger trajectories across trials).

In order to study the changes in functional cortical and striatal organization during the learning task, we used three distinct, equal-length epochs over the time course of the study. Specifically, in addition to the task baseline epoch (70 trials), we defined early and late learning epochs as the subsequent initial and final 70 trials, respectively, following the presentation onset of reward feedback. For each participant, we extracted mean blood oxygenation level-dependent (BOLD) time-series data for each cortical region defined by the Shaefer 1000 cortical parcellation (Schaefer et al., 2018) and for striatal regions defined by the Harvard-Oxford parcellation (Avants et al., 2008; Frazier et al., 2005), and then estimated covariance (functional connectivity) matrices for each epoch (baseline, early, and late; Figure 1E; for a similar approach, see Gale et al., 2022).

Because prior work (Gordon et al., 2017; Gratton et al., 2018), including our own (Areshenkoff et al., 2022; Areshenkoff et al., 2021; Gale et al., 2022), suggests that individual differences in functional connectivity can obscure any task-related effects, we centered the connectivity matrices using the Riemmanian manifold approach (Areshenkoff et al., 2022; Areshenkoff et al., 2021; Gale et al., 2022; Zhao et al., 2018; see Figure 2—figure supplement 1 for an overview of the approach). To illustrate the effects of this centering procedure, and why it is important for elucidating task-related effects in the data, we projected participants’ covariance matrices both prior to, and after the centering procedure, using uniform manifold approximation (UMAP; McInnes et al., 2018). As shown in Figure 2A, prior to the centering procedure the covariance matrices mainly cluster according to subject identity, consistent with prior findings showing that this subject-level structure explains the majority of the variance in functional connectivity data (Gratton et al., 2018). Clearly, this subject-level clustering could impact the ability to detect task-related effects in the data. However, after applying the centering procedure (Figure 2B), this subject-level clustering is abolished, potentially allowing for the differentiation of the three task-related epochs.

Figure 2 with 1 supplement see all

Download asset Open asset

To examine the changes in cortical and striatal connectivity during the reward-based motor learning task, we used the centered matrices from Figure 2B to estimate separate cortical-striatal connectivity manifolds for each participant’s baseline, early, and late covariance matrices (see also Gale et al., 2022). Following from prior work (Hong et al., 2020; Paquola et al., 2019; Vos de Wael et al., 2020), we transformed each matrix into an affinity matrix and then applied principal components analysis (PCA) to obtain a set of principal components (PCs) that provides a low-dimensional representation of cortical-striatal functional organization (i.e., a cortical-striatal manifold). Next, we aligned the manifolds from each participant to a template baseline manifold, which we constructed using the mean of all baseline connectivity matrices across participants (Figure 1E). We did this for two reasons: (1) the baseline manifold provided a common target for manifold alignment (Vos de Wael et al., 2020) so that all subjects could be directly compared in a common task-based neural space, and (2) it allowed us to selectively detect deviations from this baseline manifold architecture; that is, observe the changes to this manifold structure that occur as a function of learning during the task (when subjects begin receiving reward feedback about their performance).

Cortical-striatal manifold structure during baseline trials

The top three PCs of the template baseline manifold (Figure 3A) describe the cortical-striatal functional organization during baseline trials. As can be seen in Figure 3A, PC1 distinguishes visual regions (positive loadings in red) from somatomotor regions (negative loadings in blue). Meanwhile, PC2 distinguishes visual and somatomotor regions (in red) from the remaining cortical areas (in blue), most prominently high-order association regions within the DMN. Finally, PC3 mainly constitutes a gradient of frontoparietal areas of the dorsal attention network (DAN) and frontoparietal control network (FCN) versus DMN regions. Collectively, these top three PCs explain ~70% of the total variance (Figure 3B).

Figure 3 with 2 supplements see all

Download asset Open asset

When we mapped the brain regions onto their assigned intrinsic functional network architecture (Thomas Yeo et al., 2011), we confirmed that PCs 1 and 2 jointly differentiate visual, DMN, and somatomotor regions, replicating the tripartite structure of the brain’s intrinsic functional architecture (Huntenburg et al., 2018; Margulies et al., 2016; Figure 3D). Others have argued that this tripartite structure is a fundamental feature of functional brain organization, whereby the transition from unimodal cortex (visual and somatomotor networks) to transmodal cortex (the DMN) reflects a global processing hierarchy from lower- to higher-order brain systems (Huntenburg et al., 2018; Margulies et al., 2016; Smallwood et al., 2021).

We next sought to characterize the relative positions of cortical and striatal brain regions along the baseline connectivity-derived manifold space, thus providing a basis to examine future changes in the positioning of these regions during early and late learning. To this aim, and following from previous methods (Gale et al., 2022; Park et al., 2021a), we computed the manifold eccentricity of each region by taking its Euclidean distance from manifold centroid (coordinates (0,0,0); see Figure 3E). This eccentricity measure provides a multivariate index of each brain region’s embedding in the three-dimensional manifold space, whereby distal regions located at the extremes of the manifold have greater eccentricity than proximal regions located near the manifold center (Figure 3E). Under this framework, regions with higher eccentricity are interpreted as having higher functional segregation from other networks in the rest of the brain, whereas regions with lower eccentricity are interpreted as having higher integration (lower segregation) with other networks in the rest of the brain (Park et al., 2021a; Park et al., 2021b; Valk et al., 2023). Consistent with this interpretation, we find that our eccentricity measure strongly correlates with various graph theoretical measures of integration and segregation. For instance, we find that baseline eccentricity is positively related to cortical node strength (r = 0.88, two-tailed p<0.001) and within-manifold degree z-score (r = 0.45, two-tailed p<0.001), consistent with the notion that more eccentric regions are more strongly functionally coupled with other members of the same functional network (i.e., higher segregation; see Figure 3—figure supplement 1). Likewise, we find that baseline eccentricity is negatively related to a region’s participation coefficient (r = −0.74, two-tailed p<0.001), which is a measure of a region’s degree of cross-network integration. Thus, taken together, changes in a brain region’s eccentricity can provide us with a multivariate measure of changes in that region’s functional segregation versus integration during early and late learning.

Changes in cortical-striatal manifold structure during learning

To examine which regions exhibited significant changes in manifold eccentricity from (1) baseline to early learning and then from (2) early to late learning, we performed two sets of paired t-tests and corrected for multiple comparisons using a false discovery rate correction (FDR; q < 0.05). To directly test how regional eccentricity changes at the onset of learning (when subjects begin receiving reward feedback), we performed a contrast of early > baseline (Figure 4A). This contrast primarily revealed a pattern of increasing eccentricity, that is, manifold expansion, across several brain regions, indicating that these regions became segregated from the rest of the brain (red areas in Figure 4A). This included areas located throughout the cortical sensorimotor system and DAN, including bilateral superior-parietal, somatomotor, supplementary motor, and premotor cortex, as well as regions in lateral visual cortex. In addition, this contrast identified many key areas of the DMN, including bilateral medial frontal gyrus (MFG), MPFC, inferior frontal gyrus (IFG), and middle temporal cortex (MTC; for a network-level summary of these general effects, see the spider plot in Figure 4A). In contrast to this general pattern of expansion-related effects, we also found that a small subset of areas in the posterior medial cortex (PMC) and posterior angular gyrus (AG) instead exhibited a decrease in eccentricity, that is, manifold contraction (regions in blue in Figure 4A), indicating that these areas increased their integration with other areas of the brain. Notably, we did not observe any significant changes in striatal regions from baseline to early learning (however, for interested readers, Figure 4—figure supplement 1 shows the unthresholded data from both cortex and striatum to indicate any trends).

Figure 4 with 3 supplements see all

Download asset Open asset

Next, to examine how regional eccentricity changes over the course of learning, we performed a direct contrast of late > early learning (Figure 4B). This contrast mainly revealed a reversal in the general pattern of the effects observed in the DMN during early learning. Specifically, during late learning, several regions in bilateral MFG, MPFC, IFG, and MTC now exhibited contraction along the manifold, indicating an increased integration of these areas with other regions of the brain (this reversal can be easily observed by comparing the red areas in Figure 4A to the blue areas in Figure 4B). By contrast, areas in PMC and posterior AG now exhibited expansion, indicating an increased segregation of these areas from the brain. In addition, during late learning, we observed manifold expansion in several areas of the salience/ventral attention network (SalVentAttn), including the dorsal anterior cingulate cortex (dACC) and the anterior insula (AI), as well as higher-order lateral occipital cortical areas, and areas in retrosplenial cortex and medial ventro-temporal cortex (a network-level summary of these effects can be found in the Figure 4A/B spider plot). Again, as in the early > baseline contrast, we did not observe any significant changes in striatal regions from early to late learning. The only region that came close to reaching statistical significance in the striatum was the right pallidum (p=0.01), but this region did not pass whole-brain FDR correction (corrected alpha = 0.086; note that Figure 4—figure supplement 1 shows the unthresholded maps for this contrast to demonstrate the strong reversal in the pattern of effects during late learning, as well as indicate any trends).

Finally, for completeness, we also examined the contrast of late > baseline, which solely revealed a pattern of cortical expansion across several regions—in particular in areas of the SalVentAttn and visual network (see Figure 4—figure supplement 2). This indicates a continuing expansion (and segregation) of these regions as learning progresses.

Taken together, the above pattern of results suggests that, during early learning, transmodal areas of the DMN, as well as several areas of the sensorimotor system (including areas of the DAN), begin to segregate from other brain networks, whereas a subset of areas—the PMC and posterior AG in particular—begin to integrate with regions outside of their respective networks. By contrast, during late learning, there is a clear reversal in these patterns, with regions within the DMN and DAN beginning to integrate with areas belonging to other brain networks. In the next section, we directly examine these interpretations of manifold expansion and contraction during early and late learning.

Changes in connectivity that underlie patterns of manifold reconfiguration

Given that eccentricity provides a multivariate index of a region’s overall profile of connectivity (i.e., its relative positioning on the manifold), we next performed seed connectivity analyses to further characterize the patterns of effects that underlie the expansions and contractions of manifold structure during learning. To this aim, we selected several representative regions, distributed throughout the cortex, that epitomize the main changes in eccentricity that we observed during early learning (shown in Figure 4A). These regions included the left (contralateral) MPFC, PMd, and PMC, allowing us to characterize the patterns of connectivity changes across prefrontal, premotor, and parietal cortex, respectively (for seed-connectivity analyses of their right hemisphere homologs, see Figure 5—figure supplement 1). For each region, we contrasted seed connectivity maps between both the early learning vs. baseline epochs (early > baseline) and the late vs. early learning epochs (late > early) by computing region-wise paired t-tests, thus producing contrast maps associated with the connectivity change of each representative seed region (Figure 5). Note that in Figure 5 we display the unthresholded voxel-wise contrast maps (two leftmost panels), the region’s corresponding change in eccentricity across epochs (second from rightmost panel), and the corresponding spider plots depicting network-level changes (rightmost panel), thus allowing for a complete visualization of the collective changes in connectivity that contribute to the changes in regional eccentricity. (Note that, for the spider plots, we used the 17-network mapping in order to capitalize on the improved spatial precision compared to the 7-network mapping; Schaefer et al., 2018; Thomas Yeo et al., 2011.)

Figure 5 with 1 supplement see all

Download asset Open asset

During early learning, we found that the left MPFC seed region, associated with the DMN network, exhibited increased connectivity with other DMN subregions and reduced connectivity with superior parietal and premotor areas in the DAN (Figure 5A). By contrast, during late learning, we observed a reversal in this pattern of connectivity changes, with the MPFC now exhibiting increased connectivity with the same regions of the DAN but reduced connectivity with other DMN areas. Together, these results suggest that the manifold expansion of the MPFC during early learning arises from its increased connectivity with other DMN areas (i.e., segregation of the DMN), whereas the manifold contraction of this region during late learning arises from its increased connectivity with areas outside of the DMN, such as sensorimotor areas of the DAN.

Notably, for the left PMd seed region (Figure 5B), associated with the DAN, we observed an inverse pattern of results from that observed for the MPFC region above. Specifically, during early learning, we observed increased connectivity of the left PMd with other areas of the DAN, as well as areas belonging to the SalVentAttn network, such as the anterior insula/IFG, dACC, and inferior parietal cortex. Notably, this was coupled with its decreased connectivity to DMN areas and the hippocampus (Figure 5B). By contrast, during late learning, we observed a reversal in this pattern of effects, whereby connectivity with DMN areas and the hippocampus now increased, whereas connectivity with the DAN and SalVentAttn areas decreased. In this case, the pattern of manifold expansion and contraction of PMd during early and late learning, respectively, likely arises from its increased connectivity with attention networks in brain (the DAN and SalVentAttn) during early learning and an increase in between-network connectivity (i.e., integration) with DMN areas during late learning.

Finally, for the left PMC seed region, located at the border of the DMN-A and Control-C networks—and that was one of the few regions that exhibited contraction during early learning—we found that this region exhibited decreased connectivity during early learning with other DMN-A and Control-C subregions in bilateral PMC, as well as decreased connectivity with bilateral hippocampus and MPFC (Figure 5C). Instead, this PMC region exhibited increased connectivity with DAN areas in superior parietal cortex and premotor cortex, and most prominently, with areas in the anterior insula/IFG, dACC, and inferior parietal cortex (belonging to the SalVentAttn networks). By contrast, during late learning, we again observed a reversal in this pattern of connectivity changes, with the PMC seed region now exhibiting increased connectivity with other bilateral PMC areas, MPFC and the hippocampus, as well as reduced connectivity with the same DAN and SalVentAttn areas. Together, these results suggest that manifold contractions of the PMC during early learning arise from its increased integration with regions outside of the DMN-A and Control-C networks, such as the DAN and SalVentAttn, whereas the manifold expansions of this region during late learning arise from its increased within-network connectivity with other DMN and Control areas (i.e., segregation).

Taken together, the results of our seed connectivity analyses above are broadly consistent with our interpretation of the patterns of manifold expansion as reflecting increases in within-network connectivity (segregation) and the patterns of manifold contraction as reflecting increases in between-network connectivity (integration). More generally, however, these findings point to changes in the landscape of communication between regions of the DAN, SalVentAttn, and DMN in particular, as being associated with reward-based motor learning. Specifically, we find that, during early learning, there is increased functional coupling between several sensorimotor areas of the DAN with areas of the SalVentAttn network, whereas during late learning, these DAN sensorimotor areas switch their connectivity to DMN areas.

Changes in eccentricity relate to learning performance

In the previous sections, we characterized the patterns of mean changes in manifold structure during learning across all participants. However, it is well established that subjects exhibit significant variation in motor learning ability (de Brouwer et al., 2018; de Brouwer et al., 2022; Standage et al., 2023; Wu et al., 2014). Indeed, while the learning curve in Figure 1D shows that subjects, on average, improved their scores during the task, this group-level result obscures the fact that individuals differed greatly in their rates of learning (see individual subject learning curves in Figure 1—figure supplement 2). To emphasize this fact, Figure 6A highlights the learning curves for two example subjects: an individual who learned the hidden shape quite rapidly (a ‘fast learner’ in green) and a second individual who only gradually learned to trace the hidden shape (a ‘slow learner’ in red). To quantify this variation in subject performance in a manner that accounted the auto-correlation in learning performance over time (i.e., subjects who learned more quickly tend to exhibit better performance by the end of learning), we opted for a pure data-driven approach and performed functional principal component analysis (fPCA; Shang, 2014) on subjects’ learning curves. This approach allowed us to isolate the dominant patterns of variability in subject’s learning curves over time (see ‘Materials and methods’ for further details; see also Areshenkoff et al., 2022). Using this fPCA approach, we found that a single component—encoding overall learning—captured the majority (~75%) of the variability in subjects’ learning curves (Figure 6C). We thus used each subjects’ loading on this dominant component as a single scalar measure of subjects’ overall learning performance: Individuals who tended to learn the task more quickly had higher values on this ‘learning score’ component, whereas individuals who tended to learn the task more slowly had lower values on this component (see single data points in Figure 6C at the right).

Figure 6

Download asset Open asset

Next, to examine the neural correlates of intersubject differences in learning scores, we calculated, for each region, the association between participants’ scores and the change in eccentricity between the baseline and early learning epochs (early > baseline; Figure 7A). This analysis did not reveal any brain regions that survived FDR corrections for multiple comparisons (q < 0.05). However, the FDR approach is completely agnostic to any topographical patterns of effects across brain areas, which may correspond with known functional networks. Indeed, the full cortex and striatal correlation map in Figure 7A indicates that many region-level correlations exhibit a high degree of spatial contiguity, with many statistically significant regions (at p<0.05, bordered in black) being situated within much larger clusters of regions that exhibit a similar pattern of effects (i.e., areas in blue, denoting a negative correlation between learning score versus the change in eccentricity from baseline to early learning, tend to lie adjacent to other regions exhibiting a similar negative correlation). This is because topographically adjacent regions are likely to have similar connectivity profiles, and thus project onto similar locations along the manifold, resulting in similar brain–behavior relationships. This spatial topography suggests that the association between the eccentricity of certain brain regions and learning performance are likely to be better characterized at the level of distributed functional networks.

Figure 7

Download asset Open asset

To examine this, we mapped each region onto its respective functional network using the Yeo et al. 17-network parcellation (Thomas Yeo et al., 2011) and, for each participant, computed the mean manifold eccentricity for each network (i.e., network eccentricity). We then correlated the change in each brain network’s eccentricity across epochs with subject learning scores. We tested the statistical significance of these network-level correlations by building null models that account for the spatial autocorrelation in the brain maps (Markello et al., 2022; Váša et al., 2018; see ‘Materials and methods’) and corrected for multiple comparisons (across all networks) using an FDR correction (q < 0.05). Using this permutation testing approach, we found that it was only the change in eccentricity of the DAN-A network that correlated with learning score (see Figure 7C), such that the more the DAN-A network decreased in eccentricity from baseline to early learning (i.e., contracted along the manifold), the better subjects performed at the task (see Figure 7C, scatterplot at the right). Consistent with the notion that changes in the eccentricity of the DAN-A network are linked to learning performance, we also found the inverse pattern of effects during late learning, whereby the more that this same network increased in eccentricity from early to late learning (i.e., expanded along the manifold), the better subjects performed at the task (Figure 7D). We should note that this pattern of performance effects for the DAN-A—that is, greater contraction during early learning and greater expansion during late learning being associated with better learning—appears at odds with the group-level effects described in Figure 4A and B, where we generally find the opposite pattern for the entire DAN network (composed of the DAN-A and DAN-B subnetworks). However, this potential discrepancy can be explained when examining the changes in eccentricity using the 17-network parcellation (see Figure 4—figure supplement 3). At this higher resolution level, we find that these group-level effects for the entire DAN network are being largely driven by eccentricity changes in the DAN-B network (areas in anterior superior parietal cortex and premotor cortex), and not by mean changes in the DAN-A network. By contrast, our present results suggest that it is the contraction and expansion of areas of the DAN-A network (and not DAN-B network) that are selectively associated with subject learning performance.

To understand the global changes in connectivity that underlie these network eccentricity effects, we performed a network-level seed connectivity analyses (analogous to our single ROI seed connectivity analyses in the previous section) wherein constructed contrast maps, using the DAN-A as the seed network, for both the early vs. baseline epochs (early > baseline) and the late vs. early epochs (late > early). As before, we display the unthresholded voxel-wise contrast maps, along with corresponding spider plots depicting the network-level changes (Figure 7E), to allow for a complete visualization of the collective changes in network-level connectivity that underlie the changes in eccentricity of the DAN-A. As can be observed in Figure 7E, we find that, during early learning, DAN-A regions exhibited the largest increases in connectivity with one of the SalVentAttn subnetworks (SalVentAttn-B), whereas, during late learning, the DAN-A regions exhibited the largest increases in connectivity with one of the DMN subnetworks (DMN-B). These findings not only re-constitute the group-level effects reported above at the ROI level (in Figure 5), but they also suggest that this general transition in functional coupling—between the DAN and SalVentAttn areas during early learning, to DAN and DMN areas during late learning—is associated with differences in subject performance.

Complex behavior necessitates the coordinated activity of multiple specialized neural systems distributed across cortex and striatum. During motor learning, these systems must adapt their functional interactions to ensure appropriate behavior in response to changes in sensory feedback. While much research in motor learning has focused on understanding the role of sensorimotor cortex in isolation, our understanding of the contribution of higher-order brain systems, which play a role in the organization of behavior over time, remains incomplete. In the current study, we utilized state-of-the-art analytical methods that reconcile topographic and functional brain organization, enabling us to describe the changes in the landscape of cortical and striatal activity during learning.

During early learning, when subjects were establishing the relationship between motor commands and reward feedback, we found that regions within both the DAN (e.g., premotor cortex) and DMN (e.g., MPFC) exhibited expansion along the manifold. Our connectivity analyses revealed that this expansion was largely driven by an increase in within-network communication in both the DAN and DMN networks. There were, however, two notable exceptions to this general pattern. First, we found that connectivity between regions of the DAN increased with regions in the SalVentAttn network (e.g., anterior insula/IFG and anterior cingulate cortex). Second, areas within the PMC, part of the posterior core of the DMN, showed a pattern of manifold contraction that was primarily driven by a decrease in covariance with other DMN regions and an increase in covariance with regions of the superior DAN and also with several regions within the SalVentAttn network. Together, these results suggest that functional interactions of the sensorimotor system with the SalVentAttn network are important during initial learning. In other work, areas within the SalVentAttn network have been implicated in several aspects of cognitive control and motivation (Botvinick et al., 2004; Holroyd and Yeung, 2012; Pearson et al., 2009; Shenhav et al., 2014), and in studies on reward-based decision-making, these regions are thought to support exploratory behavior. For instance, several neuroimaging studies have shown that SalVentAttn areas are activated in response to novel or salient stimuli in the environment, presumably reflecting the engagement of attentional resources for sampling new information (Corbetta et al., 2008). In addition, recent work (Badre et al., 2012; Blanchard and Gershman, 2018; Boorman et al., 2009; Daw et al., 2006; Hogeveen et al., 2022; Kolling et al., 2016) implicates several SalVentAttn areas in information gathering functions, so as to optimize reward outcomes. This neural perspective is consistent with both others’ (Dam et al., 2013) and our own behavioral findings (de Brouwer et al., 2022) that, during the early phases of reward-guided learning, performance is more variable, presumably as individuals explore the relationship between motor commands and associated sensory feedback.

During late learning, we observed that many of the changes in manifold architecture observed during early learning reversed. For instance, areas within both the DAN and DMN now exhibited contraction along the manifold, whereas the PMC now exhibited expansion. Connectivity analyses showed that the contraction within the DAN and DMN was driven by increases in connectivity between these two networks—specifically, increases between premotor and superior parietal areas of the DAN with areas of the DMN. This suggests that once the mapping between motor commands and reward feedback have been learned, regions within the DAN and DMN become more integrated with one another. In the context of the current task, the shift in DAN connectivity from the SalVentAttn network during early learning to the DMN during late learning may reflect the hypothesized role of the DMN in supporting behavior using information from memory (Buckner et al., 2008; Schacter et al., 2012; Smallwood et al., 2021). This interpretation is consistent with prior work showing that connectivity between the DMN and premotor cortex tends to increase once rules have been learned (Shamloo and Helie, 2016), as well as studies from other task domains showing that the DMN contributes to behavior when actions must be guided by information from memory and a knowledge of task structure (Murphy et al., 2019; Murphy et al., 2018; Vatansever et al., 2017). Our analysis, therefore, provides additional evidence, albeit from the domain of human motor learning, that functional interactions between the DMN and brain regions involved in sensorimotor processes support a mode of action in which behavior must be guided by memory processes (Smallwood et al., 2021)—in this case, the history of reward information accrued across previous movement trajectories. Finally, it is important to note that the reversal pattern of effects noted above suggests that our findings during learning cannot be simply attributed to the introduction of reward feedback and/or the perturbation during early learning, as both of these task-related features are also present during late learning. In addition, these results cannot be simply explained due to the passage of time or increasing subject fatigue as this would predict a consistent directional change in eccentricity across the baseline, early, and Late learning epochs.

We also observed a relationship between changes in the manifold eccentricity of a subnetwork of the DAN (DAN-A) with subject learning performance. We found that the more this subnetwork contracted, and then subsequently expanded, along the manifold during early and late learning, respectively, the better subjects performed at the task. Our connectivity analysis revealed that this change in DAN-A activity was mainly driven by increases in connectivity with SalVentAttn subnetworks during early learning and by increases in connectivity with DMN subnetworks during late learning. Notably, areas of the DAN are not thought to generate top-down signals for response selection but instead transform the input signals they receive (e.g., related to reward, memory) into motor commands (Luo et al., 2010). Our analysis thus suggests that this system’s contribution to motor behavior may be facilitated through changes in its functional coupling to both the SalVentAttn and DMN over time. To speculate, this shift in functional coupling may reflect a shift from more exploratory to more exploitative modes of behavior across early to late periods of motor learning, respectively. It is also possible that some of these task-related shifts in connectivity relate to shifts in task-general processes, such as changes in the allocation of attentional resources (Bédard and Song, 2013; Rosenberg et al., 2016) or overall cognitive engagement (Aben et al., 2020), which themselves play critical roles in shaping learning (Codol et al., 2018; Holland et al., 2018; Song, 2019; Taylor and Thoroughman, 2008; Taylor and Thoroughman, 2007; for a review of these topics, see Tsay et al., 2023). Such processes are particularly important during the earlier phases of learning when sensorimotor contingencies need to be established. While these remain questions for future work, our data nevertheless suggest that this shift in connectivity may be enabled through the PMC.

Although traditionally considered a member of the DMN (Thomas Yeo et al., 2011), studies have established that the PMC contains echoes of neural signals originating from across the cortex (Leech et al., 2012). In our study, this region initially became more segregated from the rest of the DMN and increased connectivity with the SalVentAttn network during early learning. However, during late learning, the PMC reduced its connectivity with the SalVentAttn network and became more integrated with other members of the DMN. This pattern of changes differed from other areas of the DMN, indicating that the PMC may serve a different function during motor learning than other areas of this system. Prior studies have linked the activity of PMC areas to reward processing (Kable and Glimcher, 2007; Knutson and Bossaerts, 2007; McCoy and Platt, 2005) and the selection of response strategies that attempt to optimize reward outcomes (Barack et al., 2017; Pearson et al., 2009; Wan et al., 2015). Consistent with this, recent research shows that areas within the PMC are able to integrate information over particularly long periods of time (Hasson et al., 2015; Heilbronner and Platt, 2013; Lerner et al., 2011). This characteristic positions the PMC as an ideal candidate region to orchestrate the neural transition from (1) exploring the relationship between motor commands and sensory feedback during early learning to (2) subsequently exploiting this learned relationship during late learning. Taken together, these distinctive functional properties of PMC activity, coupled with its diverse patterns of whole-brain connectivity (Hagmann et al., 2008; Hutchison et al., 2015; Margulies et al., 2009), suggest an important role for this region in directing long-term behavioral adaptation in accordance with higher-order task objectives (Braga et al., 2013; Pearson et al., 2011).

While we identified several changes in the cortical manifold that are associated with reward-based motor learning, it is noteworthy that we did not observe any significant changes in manifold eccentricity within the striatum. While clearly the evidence indicates that this region plays a key role in reward-guided behavior (Averbeck and O’Doherty, 2022; O’Doherty et al., 2017), there are several possible reasons why our manifold approach did not identify this collection of brain areas. First, the relatively small size of the striatum may mean that our analysis approach was too coarse to identify changes in the connectivity of this region. Though we used a 3T scanner and employed a widely used parcellation scheme that divided the striatum into its constituent anatomical regions (e.g., hippocampus, caudate), both of these approaches may have obscured important differences in connectivity that exist within each of these regions. For example, areas such the hippocampus and caudate are not homogeneous areas but themselves exhibit gradients of connectivity (e.g., head versus tail) that can only be revealed at the voxel level (Tian et al., 2020; Vos de Wael et al., 2021). Second, while our dimension reduction approach, by design, aims to identify gradients of functional connectivity that account for the largest amounts of variance, the limited number of striatal regions (compared to cortex) necessitates that their contribution to the total whole-brain variance is relatively small. Consistent with this perspective, we found that the low-dimensional manifold architecture in cortex did not strongly depend on whether or not striatal regions were included in the analysis (see Figure 3—figure supplement 2). As such, selective changes in the patterns of functional connectivity at the level of the striatum may be obscured using our cortex × striatum dimension reduction approach. Future work can help address some of these limitations by using both finer parcellations of striatal cortex (perhaps even down to the voxel level) (Tian et al., 2020) and by focusing specifically on the changes in the interactions between the striatum and cortex during learning. The latter can be accomplished by selectively performing dimension reduction on the slice of the functional connectivity matrix that corresponds to functional coupling between striatum and cortex.

Versions of the data used specifically for analysis and the generation of the manuscript figures can be found on Dryad. In addition, BIDS-formatted versions of the raw data can be obtained at OpenNeuro: https://openneuro.org/datasets/ds005230/versions/1.0.0. Imaging data were preprocessed using fmriPrep, which is open source and freely available. Operations on covariance matrices, including estimation and centering, were performed using the R package spdm, which is freely available on GitHub (Areshenkoff, 2023). The analysis code for the paper can be found on GitHub (copy archived at Nick, 2024).

1. 1. Nick Q
   2. Gale DJ
   3. Areshenkoff CN
   4. De Brouwer AJ
   5. Nashed JY
   6. Wammes J
   7. Flanagan JR
   8. Smallwood J
   9. Gallivan JP

   (2024) Dryad Digital Repository

   Reconfigurations of cortical manifold structure during reward-based motor learning.

   https://doi.org/10.5061/dryad.7sqv9s512
2. 1. Nick Q
   2. Gale DJ
   3. Areshenkoff CN
   4. De Brouwer AJ
   5. Nashed JY
   6. Wammes J
   7. Flanagan JR
   8. Smallwood J
   9. Gallivan JP

   (2024) OpenNeuro

   Reinforcement-Learning Generalization.

   https://doi.org/10.18112/openneuro.ds005230.v1.0.0

1. 1. Aben B
   2. Buc Calderon C
   3. Van den Bussche E
   4. Verguts T

   (2020) Cognitive effort modulates connectivity between dorsal anterior cingulate cortex and task-relevant cortical areas

   The Journal of Neuroscience 40:3838–3848.

   https://doi.org/10.1523/JNEUROSCI.2948-19.2020

   • PubMed
   • Google Scholar
2. 1. Abraham A
   2. Pedregosa F
   3. Eickenberg M
   4. Gervais P
   5. Mueller A
   6. Kossaifi J
   7. Gramfort A
   8. Thirion B
   9. Varoquaux G

   (2014) Machine learning for neuroimaging with scikit-learn

   Frontiers in Neuroinformatics 8:14.

   https://doi.org/10.3389/fninf.2014.00014

   • PubMed
   • Google Scholar
3. 1. Andrews‐Hanna JR
   2. Smallwood J
   3. Spreng RN

   (2014) The default network and self‐generated thought: component processes, dynamic control, and clinical relevance

   Annals of the New York Academy of Sciences 1316:29–52.

   https://doi.org/10.1111/nyas.12360

   • Google Scholar
4. 1. Areshenkoff CN
   2. Nashed JY
   3. Hutchison RM
   4. Hutchison M
   5. Levy R
   6. Cook DJ
   7. Menon RS
   8. Everling S
   9. Gallivan JP

   (2021) Muting, not fragmentation, of functional brain networks under general anesthesia

   NeuroImage 231:117830.

   https://doi.org/10.1016/j.neuroimage.2021.117830

   • PubMed
   • Google Scholar
5. 1. Areshenkoff C
   2. Gale DJ
   3. Standage D
   4. Nashed JY
   5. Flanagan JR
   6. Gallivan JP

   (2022) Neural excursions from manifold structure explain patterns of learning during human sensorimotor adaptation

   eLife 11:e74591.

   https://doi.org/10.7554/eLife.74591

   • PubMed
   • Google Scholar
6. Software

   1. Areshenkoff CN

   (2023) spdm: functions for working with symmetric positive-definite Matrices, version R package version 1.1.0

   GitHub.

   https://github.com/areshenk-rpackages/spdm
7. 1. Avants BB
   2. Epstein CL
   3. Grossman M
   4. Gee JC

   (2008) Symmetric diffeomorphic image registration with cross-correlation: evaluating automated labeling of elderly and neurodegenerative brain

   Medical Image Analysis 12:26–41.

   https://doi.org/10.1016/j.media.2007.06.004

   • PubMed
   • Google Scholar
8. 1. Averbeck B
   2. O’Doherty JP

   (2022) Reinforcement-learning in fronto-striatal circuits

   Neuropsychopharmacology 47:147–162.

   https://doi.org/10.1038/s41386-021-01108-0

   • PubMed
   • Google Scholar
9. 1. Badre D
   2. Doll BB
   3. Long NM
   4. Frank MJ

   (2012) Rostrolateral prefrontal cortex and individual differences in uncertainty-driven exploration

   Neuron 73:595–607.

   https://doi.org/10.1016/j.neuron.2011.12.025

   • PubMed
   • Google Scholar
10. 1. Barack DL
    2. Chang SWC
    3. Platt ML

    (2017) Posterior cingulate neurons dynamically signal decisions to disengage during foraging

    Neuron 96:339–347.

    https://doi.org/10.1016/j.neuron.2017.09.048

    • PubMed
    • Google Scholar
11. 1. Bayer HM
    2. Glimcher PW

    (2005) Midbrain dopamine neurons encode a quantitative reward prediction error signal

    Neuron 47:129–141.

    https://doi.org/10.1016/j.neuron.2005.05.020

    • PubMed
    • Google Scholar
12. 1. Bédard P
    2. Song JH

    (2013) Attention modulates generalization of visuomotor adaptation

    Journal of Vision 13:12.

    https://doi.org/10.1167/13.12.12

    • PubMed
    • Google Scholar
13. 1. Behzadi Y
    2. Restom K
    3. Liau J
    4. Liu TT

    (2007) A component based noise correction method (CompCor) for BOLD and perfusion based fMRI

    NeuroImage 37:90–101.

    https://doi.org/10.1016/j.neuroimage.2007.04.042

    • PubMed
    • Google Scholar
14. 1. Berniker M
    2. Kording K

    (2008) Estimating the sources of motor errors for adaptation and generalization

    Nature Neuroscience 11:1454–1461.

    https://doi.org/10.1038/nn.2229

    • PubMed
    • Google Scholar
15. 1. Bethlehem RAI
    2. Paquola C
    3. Seidlitz J
    4. Ronan L
    5. Bernhardt B
    6. Consortium C-C
    7. Tsvetanov KA

    (2020) Dispersion of functional gradients across the adult lifespan

    NeuroImage 222:117299.

    https://doi.org/10.1016/j.neuroimage.2020.117299

    • PubMed
    • Google Scholar
16. 1. Binder JR
    2. Desai RH
    3. Graves WW
    4. Conant LL

    (2009) Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies

    Cerebral Cortex 19:2767–2796.

    https://doi.org/10.1093/cercor/bhp055

    • Google Scholar
17. 1. Blanchard TC
    2. Gershman SJ

    (2018) Pure correlates of exploration and exploitation in the human brain

    Cognitive, Affective & Behavioral Neuroscience 18:117–126.

    https://doi.org/10.3758/s13415-017-0556-2

    • PubMed
    • Google Scholar
18. 1. Boorman ED
    2. Behrens TEJ
    3. Woolrich MW
    4. Rushworth MFS

    (2009) How green is the grass on the other side? Frontopolar cortex and the evidence in favor of alternative courses of action

    Neuron 62:733–743.

    https://doi.org/10.1016/j.neuron.2009.05.014

    • PubMed
    • Google Scholar
19. 1. Botvinick MM
    2. Cohen JD
    3. Carter CS

    (2004) Conflict monitoring and anterior cingulate cortex: an update

    Trends in Cognitive Sciences 8:539–546.

    https://doi.org/10.1016/j.tics.2004.10.003

    • PubMed
    • Google Scholar
20. 1. Braga RM
    2. Sharp DJ
    3. Leeson C
    4. Wise RJS
    5. Leech R

    (2013) Echoes of the brain within default mode, association, and heteromodal cortices

    The Journal of Neuroscience 33:14031–14039.

    https://doi.org/10.1523/JNEUROSCI.0570-13.2013

    • PubMed
    • Google Scholar
21. Book

    1. Buckner RL
    2. Andrews-Hanna JR
    3. Schacter DL

    (2008) The Brain’s Default Network

    Annals of the New York Academy of Sciences.

    https://doi.org/10.1196/annals.1440.011

    • Google Scholar
22. 1. Christoff K
    2. Irving ZC
    3. Fox KCR
    4. Spreng RN
    5. Andrews-Hanna JR

    (2016) Mind-wandering as spontaneous thought: a dynamic framework

    Nature Reviews. Neuroscience 17:718–731.

    https://doi.org/10.1038/nrn.2016.113

    • PubMed
    • Google Scholar
23. 1. Codol O
    2. Holland PJ
    3. Galea JM

    (2018) The relationship between reinforcement and explicit control during visuomotor adaptation

    Scientific Reports 8:9121.

    https://doi.org/10.1038/s41598-018-27378-1

    • PubMed
    • Google Scholar
24. 1. Corbetta M
    2. Shulman GL

    (2002) Control of goal-directed and stimulus-driven attention in the brain

    Nature Reviews. Neuroscience 3:201–215.

    https://doi.org/10.1038/nrn755

    • PubMed
    • Google Scholar
25. 1. Corbetta M
    2. Patel G
    3. Shulman GL

    (2008) The reorienting system of the human brain: from environment to theory of mind

    Neuron 58:306–324.

    https://doi.org/10.1016/j.neuron.2008.04.017

    • PubMed
    • Google Scholar
26. 1. Cox RW
    2. Hyde JS

    (1997) Software tools for analysis and visualization of fMRI data

    NMR in Biomedicine 10:171–178.

    https://doi.org/10.1002/(sici)1099-1492(199706/08)10:4/5<171::aid-nbm453>3.0.co;2-l

    • PubMed
    • Google Scholar
27. 1. Cunningham JP
    2. Yu BM

    (2014) Dimensionality reduction for large-scale neural recordings

    Nature Neuroscience 17:1500–1509.

    https://doi.org/10.1038/nn.3776

    • PubMed
    • Google Scholar
28. 1. Dale AM
    2. Fischl B
    3. Sereno MI

    (1999) Cortical surface-based analysis

    NeuroImage 9:179–194.

    https://doi.org/10.1006/nimg.1998.0395

    • Google Scholar
29. 1. Dam G
    2. Kording K
    3. Wei K

    (2013) Credit assignment during movement reinforcement learning

    PLOS ONE 8:e55352.

    https://doi.org/10.1371/journal.pone.0055352

    • PubMed
    • Google Scholar
30. 1. Daw ND
    2. O’Doherty JP
    3. Dayan P
    4. Seymour B
    5. Dolan RJ

    (2006) Cortical substrates for exploratory decisions in humans

    Nature 441:876–879.

    https://doi.org/10.1038/nature04766

    • PubMed
    • Google Scholar
31. 1. de Brouwer AJ
    2. Albaghdadi M
    3. Flanagan JR
    4. Gallivan JP

    (2018) Using gaze behavior to parcellate the explicit and implicit contributions to visuomotor learning

    Journal of Neurophysiology 120:1602–1615.

    https://doi.org/10.1152/jn.00113.2018

    • Google Scholar
32. 1. de Brouwer AJ
    2. Areshenkoff CN
    3. Rashid MR
    4. Flanagan JR
    5. Poppenk J
    6. Gallivan JP

    (2022) Human variation in error-based and reinforcement motor learning is associated with entorhinal volume

    Cerebral Cortex 32:3423–3440.

    https://doi.org/10.1093/cercor/bhab424

    • Google Scholar
33. 1. Dhawale AK
    2. Smith MA
    3. Ölveczky BP

    (2017) The role of variability in motor learning

    Annual Review of Neuroscience 40:479–498.

    https://doi.org/10.1146/annurev-neuro-072116-031548

    • PubMed
    • Google Scholar
34. 1. Dhawale AK
    2. Miyamoto YR
    3. Smith MA
    4. Ölveczky BP

    (2019) Adaptive regulation of motor variability

    Current Biology 29:3551–3562.

    https://doi.org/10.1016/j.cub.2019.08.052

    • PubMed
    • Google Scholar
35. 1. Esteban O
    2. Markiewicz CJ
    3. Blair RW
    4. Moodie CA
    5. Isik AI
    6. Erramuzpe A
    7. Kent JD
    8. Goncalves M
    9. DuPre E
    10. Snyder M
    11. Oya H
    12. Ghosh SS
    13. Wright J
    14. Durnez J
    15. Poldrack RA
    16. Gorgolewski KJ

    (2019) fMRIPrep: a robust preprocessing pipeline for functional MRI

    Nature Methods 16:111–116.

    https://doi.org/10.1038/s41592-018-0235-4

    • PubMed
    • Google Scholar
36. Software

    1. Esteban O
    2. Blair R
    3. Markiewicz CJ
    4. Berleant SL

    (2024) fMRIPrep

    Software. Zenodo.

    https://zenodo.org/records/3511440
37. 1. Foster BL
    2. Koslov SR
    3. Aponik-Gremillion L
    4. Monko ME
    5. Hayden BY
    6. Heilbronner SR

    (2023) A tripartite view of the posterior cingulate cortex

    Nature Reviews. Neuroscience 24:173–189.

    https://doi.org/10.1038/s41583-022-00661-x

    • PubMed
    • Google Scholar
38. 1. Fox MD
    2. Raichle ME

    (2007) Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging

    Nature Reviews Neuroscience 8:700–711.

    https://doi.org/10.1038/nrn2201

    • Google Scholar
39. 1. Frank MJ
    2. Doll BB
    3. Oas-Terpstra J
    4. Moreno F

    (2009) Prefrontal and striatal dopaminergic genes predict individual differences in exploration and exploitation

    Nature Neuroscience 12:1062–1068.

    https://doi.org/10.1038/nn.2342

    • Google Scholar
40. 1. Frazier JA
    2. Chiu S
    3. Breeze JL
    4. Makris N
    5. Lange N
    6. Kennedy DN
    7. Herbert MR
    8. Bent EK
    9. Koneru VK
    10. Dieterich ME
    11. Hodge SM
    12. Rauch SL
    13. Grant PE
    14. Cohen BM
    15. Seidman LJ
    16. Caviness VS
    17. Biederman J

    (2005) Structural brain magnetic resonance imaging of limbic and thalamic volumes in pediatric bipolar disorder

    The American Journal of Psychiatry 162:1256–1265.

    https://doi.org/10.1176/appi.ajp.162.7.1256

    • PubMed
    • Google Scholar
41. 1. Gale DJ
    2. Areshenkoff CN
    3. Standage DI
    4. Nashed JY
    5. Markello RD
    6. Flanagan JR
    7. Smallwood J
    8. Gallivan JP

    (2022) Distinct patterns of cortical manifold expansion and contraction underlie human sensorimotor adaptation

    PNAS 119:e2209960119.

    https://doi.org/10.1073/pnas.2209960119

    • PubMed
    • Google Scholar
42. 1. Gallego JA
    2. Perich MG
    3. Miller LE
    4. Solla SA

    (2017) Neural manifolds for the control of movement

    Neuron 94:978–984.

    https://doi.org/10.1016/j.neuron.2017.05.025

    • PubMed
    • Google Scholar
43. 1. Glasser MF
    2. Sotiropoulos SN
    3. Wilson JA
    4. Coalson TS
    5. Fischl B
    6. Andersson JL
    7. Xu J
    8. Jbabdi S
    9. Webster M
    10. Polimeni JR
    11. Van Essen DC
    12. Jenkinson M

    (2013) The minimal preprocessing pipelines for the human connectome project

    NeuroImage 80:105–124.

    https://doi.org/10.1016/j.neuroimage.2013.04.127

    • Google Scholar
44. 1. Gordon EM
    2. Laumann TO
    3. Gilmore AW
    4. Newbold DJ
    5. Greene DJ
    6. Berg JJ
    7. Ortega M
    8. Hoyt-Drazen C
    9. Gratton C
    10. Sun H
    11. Hampton JM
    12. Coalson RS
    13. Nguyen AL
    14. McDermott KB
    15. Shimony JS
    16. Snyder AZ
    17. Schlaggar BL
    18. Petersen SE
    19. Nelson SM
    20. Dosenbach NUF

    (2017) Precision functional mapping of individual human brains

    Neuron 95:791–807.

    https://doi.org/10.1016/j.neuron.2017.07.011

    • PubMed
    • Google Scholar
45. 1. Gorgolewski K
    2. Burns CD
    3. Madison C
    4. Clark D
    5. Halchenko YO
    6. Waskom ML
    7. Ghosh SS

    (2011) Nipype: a flexible, lightweight and extensible neuroimaging data processing framework in python

    Frontiers in Neuroinformatics 5:13.

    https://doi.org/10.3389/fninf.2011.00013

    • PubMed
    • Google Scholar
46. 1. Gorgolewski KJ
    2. Nichols T
    3. Kennedy DN
    4. Poline J-B
    5. Poldrack RA

    (2018) Making replication prestigious

    The Behavioral and Brain Sciences 41:e131.

    https://doi.org/10.1017/S0140525X18000663

    • PubMed
    • Google Scholar
47. 1. Gratton C
    2. Laumann TO
    3. Nielsen AN
    4. Greene DJ
    5. Gordon EM
    6. Gilmore AW
    7. Nelson SM
    8. Coalson RS
    9. Snyder AZ
    10. Schlaggar BL
    11. Dosenbach NUF
    12. Petersen SE

    (2018) Functional brain networks are dominated by stable group and individual factors, not cognitive or daily variation

    Neuron 98:439–452.

    https://doi.org/10.1016/j.neuron.2018.03.035

    • PubMed
    • Google Scholar
48. 1. Greve DN
    2. Fischl B

    (2009) Accurate and robust brain image alignment using boundary-based registration

    NeuroImage 48:63–72.

    https://doi.org/10.1016/j.neuroimage.2009.06.060

    • PubMed
    • Google Scholar
49. 1. Hagmann P
    2. Cammoun L
    3. Gigandet X
    4. Meuli R
    5. Honey CJ
    6. Wedeen VJ
    7. Sporns O

    (2008) Mapping the structural core of human cerebral cortex

    PLOS Biology 6:e159.

    https://doi.org/10.1371/journal.pbio.0060159

    • PubMed
    • Google Scholar
50. Book

    1. Härdle W

    (1990) Applied Nonparametric Regression

    Cambridge University Press.

    https://doi.org/10.1017/CCOL0521382483

    • Google Scholar
51. 1. Hasson U
    2. Chen J
    3. Honey CJ

    (2015) Hierarchical process memory: memory as an integral component of information processing

    Trends in Cognitive Sciences 19:304–313.

    https://doi.org/10.1016/j.tics.2015.04.006

    • PubMed
    • Google Scholar
52. 1. Hayden BY
    2. Nair AC
    3. McCoy AN
    4. Platt ML

    (2008) Posterior cingulate cortex mediates outcome-contingent allocation of behavior

    Neuron 60:19–25.

    https://doi.org/10.1016/j.neuron.2008.09.012

    • PubMed
    • Google Scholar
53. 1. Hayden BY
    2. Smith DV
    3. Platt ML

    (2009) Electrophysiological correlates of default-mode processing in macaque posterior cingulate cortex

    PNAS 106:5948–5953.

    https://doi.org/10.1073/pnas.0812035106

    • PubMed
    • Google Scholar
54. 1. Heilbronner SR
    2. Platt ML

    (2013) Causal evidence of performance monitoring by neurons in posterior cingulate cortex during learning

    Neuron 80:1384–1391.

    https://doi.org/10.1016/j.neuron.2013.09.028

    • PubMed
    • Google Scholar
55. 1. Hogeveen J
    2. Mullins TS
    3. Romero JD
    4. Eversole E
    5. Rogge-Obando K
    6. Mayer AR
    7. Costa VD

    (2022) The neurocomputational bases of explore-exploit decision-making

    Neuron 110:1869–1879.

    https://doi.org/10.1016/j.neuron.2022.03.014

    • PubMed
    • Google Scholar
56. 1. Holland P
    2. Codol O
    3. Galea JM

    (2018) Contribution of explicit processes to reinforcement-based motor learning

    Journal of Neurophysiology 119:2241–2255.

    https://doi.org/10.1152/jn.00901.2017

    • PubMed
    • Google Scholar
57. 1. Holroyd CB
    2. Yeung N

    (2012) Motivation of extended behaviors by anterior cingulate cortex

    Trends in Cognitive Sciences 16:122–128.

    https://doi.org/10.1016/j.tics.2011.12.008

    • Google Scholar
58. 1. Hong SJ
    2. Xu T
    3. Nikolaidis A
    4. Smallwood J
    5. Margulies DS
    6. Bernhardt B
    7. Vogelstein J
    8. Milham MP

    (2020) Toward a connectivity gradient-based framework for reproducible biomarker discovery

    NeuroImage 223:117322.

    https://doi.org/10.1016/j.neuroimage.2020.117322

    • Google Scholar
59. 1. Houk JC
    2. Buckingham JT
    3. Barto AG

    (1996) Models of the cerebellum and motor learning

    Behavioral and Brain Sciences 19:368–383.

    https://doi.org/10.1017/S0140525X00081474

    • Google Scholar
60. 1. Huntenburg JM
    2. Bazin PL
    3. Margulies DS

    (2018) Large-scale gradients in human cortical organization

    Trends in Cognitive Sciences 22:21–31.

    https://doi.org/10.1016/j.tics.2017.11.002

    • PubMed
    • Google Scholar
61. 1. Hutchison RM
    2. Culham JC
    3. Flanagan JR
    4. Everling S
    5. Gallivan JP

    (2015) Functional subdivisions of medial parieto-occipital cortex in humans and nonhuman primates using resting-state fMRI

    NeuroImage 116:10–29.

    https://doi.org/10.1016/j.neuroimage.2015.04.068

    • Google Scholar
62. 1. Jenkinson M
    2. Bannister P
    3. Brady M
    4. Smith S

    (2002) Improved optimization for the robust and accurate linear registration and motion correction of brain images

    NeuroImage 17:825–841.

    https://doi.org/10.1006/nimg.2002.1132

    • Google Scholar
63. 1. Kable JW
    2. Glimcher PW

    (2007) The neural correlates of subjective value during intertemporal choice

    Nature Neuroscience 10:1625–1633.

    https://doi.org/10.1038/nn2007

    • PubMed
    • Google Scholar
64. 1. Klein A
    2. Ghosh SS
    3. Bao FS
    4. Giard J
    5. Häme Y
    6. Stavsky E
    7. Lee N
    8. Rossa B
    9. Reuter M
    10. Chaibub Neto E
    11. Keshavan A

    (2017) Mindboggling morphometry of human brains

    PLOS Computational Biology 13:e1005350.

    https://doi.org/10.1371/journal.pcbi.1005350

    • Google Scholar
65. 1. Klein-Flügge MC
    2. Bongioanni A
    3. Rushworth MFS

    (2022) Medial and orbital frontal cortex in decision-making and flexible behavior

    Neuron 110:2743–2770.

    https://doi.org/10.1016/j.neuron.2022.05.022

    • PubMed
    • Google Scholar
66. 1. Knutson B
    2. Bossaerts P

    (2007) Neural antecedents of financial decisions

    The Journal of Neuroscience 27:8174–8177.

    https://doi.org/10.1523/JNEUROSCI.1564-07.2007

    • Google Scholar
67. 1. Kolling N
    2. Wittmann MK
    3. Behrens TEJ
    4. Boorman ED
    5. Mars RB
    6. Rushworth MFS

    (2016) Value, search, persistence and model updating in anterior cingulate cortex

    Nature Neuroscience 19:1280–1285.

    https://doi.org/10.1038/nn.4382

    • Google Scholar
68. 1. Lanczos C

    (1964) Evaluation of noisy data

    Journal of the Society for Industrial and Applied Mathematics Series B Numerical Analysis 1:76–85.

    https://doi.org/10.1137/0701007

    • Google Scholar
69. 1. Ledoit O
    2. Wolf M

    (2004) A well-conditioned estimator for large-dimensional covariance matrices

    Journal of Multivariate Analysis 88:365–411.

    https://doi.org/10.1016/S0047-259X(03)00096-4

    • Google Scholar
70. 1. Lee D
    2. Seo H
    3. Jung MW

    (2012) Neural basis of reinforcement learning and decision making

    Annual Review of Neuroscience 35:287–308.

    https://doi.org/10.1146/annurev-neuro-062111-150512

    • PubMed
    • Google Scholar
71. 1. Leech R
    2. Braga R
    3. Sharp DJ

    (2012) Echoes of the brain within the posterior cingulate cortex

    The Journal of Neuroscience 32:215–222.

    https://doi.org/10.1523/JNEUROSCI.3689-11.2012

    • PubMed
    • Google Scholar
72. 1. Lerner Y
    2. Honey CJ
    3. Silbert LJ
    4. Hasson U

    (2011) Topographic mapping of a hierarchy of temporal receptive windows using a narrated story

    The Journal of Neuroscience 31:2906–2915.

    https://doi.org/10.1523/JNEUROSCI.3684-10.2011

    • Google Scholar
73. 1. Luo L
    2. Rodriguez E
    3. Jerbi K
    4. Lachaux J-P
    5. Martinerie J
    6. Corbetta M
    7. Shulman GL
    8. Piomelli D
    9. Turrigiano GG
    10. Nelson SB
    11. Joëls M
    12. de Kloet ER
    13. Holsboer F
    14. Amodio DM
    15. Frith CD
    16. Block ML
    17. Zecca L
    18. Hong J-S
    19. Dantzer R
    20. Kelley KW
    21. Craig AD

    (2010) Ten years of nature reviews neuroscience: Insights from the highly cited

    Nature Reviews Neuroscience 11:718–726.

    https://doi.org/10.1038/nrn2912

    • Google Scholar
74. 1. Makris N
    2. Goldstein JM
    3. Kennedy D
    4. Hodge SM
    5. Caviness VS
    6. Faraone SV
    7. Tsuang MT
    8. Seidman LJ

    (2006) Decreased volume of left and total anterior insular lobule in schizophrenia

    Schizophrenia Research 83:155–171.

    https://doi.org/10.1016/j.schres.2005.11.020

    • Google Scholar
75. 1. Margulies DS
    2. Vincent JL
    3. Kelly C
    4. Lohmann G
    5. Uddin LQ
    6. Biswal BB
    7. Villringer A
    8. Castellanos FX
    9. Milham MP
    10. Petrides M

    (2009) Precuneus shares intrinsic functional architecture in humans and monkeys

    PNAS 106:20069–20074.

    https://doi.org/10.1073/pnas.0905314106

    • PubMed
    • Google Scholar
76. 1. Margulies DS
    2. Ghosh SS
    3. Goulas A
    4. Falkiewicz M
    5. Huntenburg JM
    6. Langs G
    7. Bezgin G
    8. Eickhoff SB
    9. Castellanos FX
    10. Petrides M
    11. Jefferies E
    12. Smallwood J

    (2016) Situating the default-mode network along a principal gradient of macroscale cortical organization

    PNAS 113:12574–12579.

    https://doi.org/10.1073/pnas.1608282113

    • Google Scholar
77. 1. Markello RD
    2. Hansen JY
    3. Liu ZQ
    4. Bazinet V
    5. Shafiei G
    6. Suárez LE
    7. Blostein N
    8. Seidlitz J
    9. Baillet S
    10. Satterthwaite TD
    11. Chakravarty MM
    12. Raznahan A
    13. Misic B

    (2022) neuromaps: structural and functional interpretation of brain maps

    Nature Methods 19:1472–1479.

    https://doi.org/10.1038/s41592-022-01625-w

    • PubMed
    • Google Scholar
78. 1. McCoy AN
    2. Platt ML

    (2005) Risk-sensitive neurons in macaque posterior cingulate cortex

    Nature Neuroscience 8:1220–1227.

    https://doi.org/10.1038/nn1523

    • Google Scholar
79. 1. McDougle SD
    2. Boggess MJ
    3. Crossley MJ
    4. Parvin D
    5. Ivry RB
    6. Taylor JA

    (2016) Credit assignment in movement-dependent reinforcement learning

    PNAS 113:6797–6802.

    https://doi.org/10.1073/pnas.1523669113

    • PubMed
    • Google Scholar
80. 1. McDougle SD
    2. Butcher PA
    3. Parvin DE
    4. Mushtaq F
    5. Niv Y
    6. Ivry RB
    7. Taylor JA

    (2019) Neural signatures of prediction errors in a decision-making task are modulated by action execution failures

    Current Biology 29:1606–1613.

    https://doi.org/10.1016/j.cub.2019.04.011

    • PubMed
    • Google Scholar
81. 1. McInnes L
    2. Healy J
    3. Saul N
    4. Großberger L

    (2018) UMAP: uniform manifold approximation and projection

    Journal of Open Source Software 3:861.

    https://doi.org/10.21105/joss.00861

    • Google Scholar
82. 1. Murphy C
    2. Jefferies E
    3. Rueschemeyer S-A
    4. Sormaz M
    5. Wang H
    6. Margulies DS
    7. Smallwood J

    (2018) Distant from input: Evidence of regions within the default mode network supporting perceptually-decoupled and conceptually-guided cognition

    NeuroImage 171:393–401.

    https://doi.org/10.1016/j.neuroimage.2018.01.017

    • Google Scholar
83. 1. Murphy C
    2. Wang HT
    3. Konu D
    4. Lowndes R
    5. Margulies DS
    6. Jefferies E
    7. Smallwood J

    (2019) Modes of operation: A topographic neural gradient supporting stimulus dependent and independent cognition

    NeuroImage 186:487–496.

    https://doi.org/10.1016/j.neuroimage.2018.11.009

    • Google Scholar
84. 1. Muschelli J
    2. Nebel MB
    3. Caffo BS
    4. Barber AD
    5. Pekar JJ
    6. Mostofsky SH

    (2014) Reduction of motion-related artifacts in resting state fMRI using aCompCor

    NeuroImage 96:22–35.

    https://doi.org/10.1016/j.neuroimage.2014.03.028

    • Google Scholar
85. Software

    1. Nick Q

    (2024) Reconfigurations of cortical manifold structure during reward-based motor learning, version swh:1:rev:c593cf2e2e0fd87fa352a470e764681bf285caf2

    Software Heritage.

    https://archive.softwareheritage.org/swh:1:dir:3b14b38bac1a235c8ccf5659e278d9f9c346695b;origin=https://github.com/qniksefat/cortical-manifolds-in-reward-based-motor-learning;visit=swh:1:snp:4fdf9b50ca4445ba47d7197cc26a9529d4810a9b;anchor=swh:1:rev:c593cf2e2e0fd87fa352a470e764681bf285caf2
86. 1. O’Doherty JP
    2. Dayan P
    3. Friston K
    4. Critchley H
    5. Dolan RJ

    (2003) Temporal difference models and reward-related learning in the human brain

    Neuron 38:329–337.

    https://doi.org/10.1016/S0896-6273(03)00169-7

    • Google Scholar
87. 1. O’Doherty JP
    2. Cockburn J
    3. Pauli WM

    (2017) Learning, reward, and decision making

    Annual Review of Psychology 68:73–100.

    https://doi.org/10.1146/annurev-psych-010416-044216

    • PubMed
    • Google Scholar
88. 1. Oldfield RC

    (1971) The assessment and analysis of handedness: the Edinburgh inventory

    Neuropsychologia 9:97–113.

    https://doi.org/10.1016/0028-3932(71)90067-4

    • PubMed
    • Google Scholar
89. 1. Paquola C
    2. Vos De Wael R
    3. Wagstyl K
    4. Bethlehem RAI
    5. Hong S-J
    6. Seidlitz J
    7. Bullmore ET
    8. Evans AC
    9. Misic B
    10. Margulies DS
    11. Smallwood J
    12. Bernhardt BC

    (2019) Microstructural and functional gradients are increasingly dissociated in transmodal cortices

    PLOS Biology 17:e3000284.

    https://doi.org/10.1371/journal.pbio.3000284

    • Google Scholar
90. 1. Park B-Y
    2. Bethlehem RA
    3. Paquola C
    4. Larivière S
    5. Rodríguez-Cruces R
    6. Vos de Wael R
    7. Bullmore ET
    8. Bernhardt BC
    9. Neuroscience in Psychiatry Network (NSPN) Consortium

    (2021a) An expanding manifold in transmodal regions characterizes adolescent reconfiguration of structural connectome organization

    eLife 10:e64694.

    https://doi.org/10.7554/eLife.64694

    • PubMed
    • Google Scholar
91. 1. Park B
    2. Park H
    3. Morys F
    4. Kim M
    5. Byeon K
    6. Lee H
    7. Kim SH
    8. Valk SL
    9. Dagher A
    10. Bernhardt BC

    (2021b) Inter-individual body mass variations relate to fractionated functional brain hierarchies

    Communications Biology 4:735.

    https://doi.org/10.1038/s42003-021-02268-x

    • Google Scholar
92. 1. Pearson JM
    2. Hayden BY
    3. Raghavachari S
    4. Platt ML

    (2009) Neurons in posterior cingulate cortex signal exploratory decisions in a dynamic multioption choice task

    Current Biology 19:1532–1537.

    https://doi.org/10.1016/j.cub.2009.07.048

    • Google Scholar
93. 1. Pearson JM
    2. Heilbronner SR
    3. Barack DL
    4. Hayden BY
    5. Platt ML

    (2011) Posterior cingulate cortex: adapting behavior to a changing world

    Trends in Cognitive Sciences 15:143–151.

    https://doi.org/10.1016/j.tics.2011.02.002

    • Google Scholar
94. 1. Pekny SE
    2. Izawa J
    3. Shadmehr R

    (2015) Reward-dependent modulation of movement variability

    The Journal of Neuroscience 35:4015–4024.

    https://doi.org/10.1523/JNEUROSCI.3244-14.2015

    • PubMed
    • Google Scholar
95. 1. Raichle ME

    (2015) The brain’s default mode network

    Annual Review of Neuroscience 38:433–447.

    https://doi.org/10.1146/annurev-neuro-071013-014030

    • PubMed
    • Google Scholar
96. Book

    1. Ramsay J
    2. Silverman BW

    (2013)

    Functional Data Analysis

    Springer Science & Business Media.

    • Google Scholar
97. Software

    1. Ramsay J
    2. Wickham H
    3. Ramsay MJ
    4. deSolve S

    (2022)

    Package “fda"

    Fda.
98. 1. Reuter M
    2. Rosas HD
    3. Fischl B

    (2010) Highly accurate inverse consistent registration: A robust approach

    NeuroImage 53:1181–1196.

    https://doi.org/10.1016/j.neuroimage.2010.07.020

    • Google Scholar
99. 1. Rosenberg MD
    2. Finn ES
    3. Scheinost D
    4. Papademetris X
    5. Shen X
    6. Constable RT
    7. Chun MM

    (2016) A neuromarker of sustained attention from whole-brain functional connectivity

    Nature Neuroscience 19:165–171.

    https://doi.org/10.1038/nn.4179

    • PubMed
    • Google Scholar
100. 1. Rubinov M
     2. Sporns O

     (2010) Complex network measures of brain connectivity: uses and interpretations

     NeuroImage 52:1059–1069.

     https://doi.org/10.1016/j.neuroimage.2009.10.003

     • PubMed
     • Google Scholar
101. 1. Rushworth MFS
     2. Noonan MP
     3. Boorman ED
     4. Walton ME
     5. Behrens TE

     (2011) Frontal cortex and reward-guided learning and decision-making

     Neuron 70:1054–1069.

     https://doi.org/10.1016/j.neuron.2011.05.014

     • Google Scholar
102. 1. Samejima K
     2. Ueda Y
     3. Doya K
     4. Kimura M

     (2005) Representation of action-specific reward values in the striatum

     Science 310:1337–1340.

     https://doi.org/10.1126/science.1115270

     • PubMed
     • Google Scholar
103. 1. Satterthwaite TD
     2. Elliott MA
     3. Gerraty RT
     4. Ruparel K
     5. Loughead J
     6. Calkins ME
     7. Eickhoff SB
     8. Hakonarson H
     9. Gur RC
     10. Gur RE
     11. Wolf DH

     (2013) An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity data

     NeuroImage 64:240–256.

     https://doi.org/10.1016/j.neuroimage.2012.08.052

     • PubMed
     • Google Scholar
104. 1. Schacter DL
     2. Addis DR
     3. Hassabis D
     4. Martin VC
     5. Spreng RN
     6. Szpunar KK

     (2012) The future of memory: remembering, imagining, and the brain

     Neuron 76:677–694.

     https://doi.org/10.1016/j.neuron.2012.11.001

     • PubMed
     • Google Scholar
105. 1. Schaefer A
     2. Kong R
     3. Gordon EM
     4. Laumann TO
     5. Zuo XN
     6. Holmes AJ
     7. Eickhoff SB
     8. Yeo BTT

     (2018) Local-global parcellation of the human cerebral cortex from intrinsic functional connectivity MRI

     Cerebral Cortex 28:3095–3114.

     https://doi.org/10.1093/cercor/bhx179

     • PubMed
     • Google Scholar
106. 1. Schuck NW
     2. Gaschler R
     3. Wenke D
     4. Heinzle J
     5. Frensch PA
     6. Haynes JD
     7. Reverberi C

     (2015) Medial prefrontal cortex predicts internally driven strategy shifts

     Neuron 86:331–340.

     https://doi.org/10.1016/j.neuron.2015.03.015

     • PubMed
     • Google Scholar
107. 1. Schultz W
     2. Dayan P
     3. Montague PR

     (1997) A neural substrate of prediction and reward

     Science 275:1593–1599.

     https://doi.org/10.1126/science.275.5306.1593

     • PubMed
     • Google Scholar
108. 1. Shamloo F
     2. Helie S

     (2016) Changes in default mode network as automaticity develops in a categorization task

     Behavioural Brain Research 313:324–333.

     https://doi.org/10.1016/j.bbr.2016.07.029

     • Google Scholar
109. 1. Shang HL

     (2014) A survey of functional principal component analysis

     AStA Advances in Statistical Analysis 98:121–142.

     https://doi.org/10.1007/s10182-013-0213-1

     • Google Scholar
110. 1. Shenhav A
     2. Straccia MA
     3. Cohen JD
     4. Botvinick MM

     (2014) Anterior cingulate engagement in a foraging context reflects choice difficulty, not foraging value

     Nature Neuroscience 17:1249–1254.

     https://doi.org/10.1038/nn.3771

     • PubMed
     • Google Scholar
111. 1. Shenoy KV
     2. Sahani M
     3. Churchland MM

     (2013) Cortical control of arm movements: a dynamical systems perspective

     Annual Review of Neuroscience 36:337–359.

     https://doi.org/10.1146/annurev-neuro-062111-150509

     • PubMed
     • Google Scholar
112. 1. Shine JM
     2. Breakspear M
     3. Bell PT
     4. Ehgoetz Martens KA
     5. Shine R
     6. Koyejo O
     7. Sporns O
     8. Poldrack RA

     (2019a) Human cognition involves the dynamic integration of neural activity and neuromodulatory systems

     Nature Neuroscience 22:289–296.

     https://doi.org/10.1038/s41593-018-0312-0

     • Google Scholar
113. 1. Shine JM
     2. Hearne LJ
     3. Breakspear M
     4. Hwang K
     5. Müller EJ
     6. Sporns O
     7. Poldrack RA
     8. Mattingley JB
     9. Cocchi L

     (2019b) The low-dimensional neural architecture of cognitive complexity is related to activity in medial thalamic nuclei

     Neuron 104:849–855.

     https://doi.org/10.1016/j.neuron.2019.09.002

     • PubMed
     • Google Scholar
114. 1. Smallwood J
     2. Bernhardt BC
     3. Leech R
     4. Bzdok D
     5. Jefferies E
     6. Margulies DS

     (2021) The default mode network in cognition: a topographical perspective

     Nature Reviews. Neuroscience 22:503–513.

     https://doi.org/10.1038/s41583-021-00474-4

     • PubMed
     • Google Scholar
115. 1. Song JH

     (2019) The role of attention in motor control and learning

     Current Opinion in Psychology 29:261–265.

     https://doi.org/10.1016/j.copsyc.2019.08.002

     • PubMed
     • Google Scholar
116. 1. Spreng RN
     2. Mar RA
     3. Kim ASN

     (2009) The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: a quantitative meta-analysis

     Journal of Cognitive Neuroscience 21:489–510.

     https://doi.org/10.1162/jocn.2008.21029

     • PubMed
     • Google Scholar
117. 1. Standage DI
     2. Areshenkoff CN
     3. Gale DJ
     4. Nashed JY
     5. Flanagan JR
     6. Gallivan JP

     (2023) Whole-brain dynamics of human sensorimotor adaptation

     Cerebral Cortex 33:4761–4778.

     https://doi.org/10.1093/cercor/bhac378

     • Google Scholar
118. Book

     1. Sutton RS
     2. Barto AG

     (2018)

     Reinforcement Learning, second edition: An Introduction

     MIT Press.

     • Google Scholar
119. 1. Taylor JA
     2. Thoroughman KA

     (2007) Divided attention impairs human motor adaptation but not feedback control

     Journal of Neurophysiology 98:317–326.

     https://doi.org/10.1152/jn.01070.2006

     • PubMed
     • Google Scholar
120. 1. Taylor JA
     2. Thoroughman KA

     (2008) Motor adaptation scaled by the difficulty of a secondary cognitive task

     PLOS ONE 3:e2485.

     https://doi.org/10.1371/journal.pone.0002485

     • PubMed
     • Google Scholar
121. 1. Thomas Yeo BT
     2. Krienen FM
     3. Sepulcre J
     4. Sabuncu MR
     5. Lashkari D
     6. Hollinshead M
     7. Roffman JL
     8. Smoller JW
     9. Zöllei L
     10. Polimeni JR
     11. Fischl B
     12. Liu H
     13. Buckner RL

     (2011) The organization of the human cerebral cortex estimated by intrinsic functional connectivity

     Journal of Neurophysiology 106:1125–1165.

     https://doi.org/10.1152/jn.00338.2011

     • Google Scholar
122. 1. Tian Y
     2. Margulies DS
     3. Breakspear M
     4. Zalesky A

     (2020) Topographic organization of the human subcortex unveiled with functional connectivity gradients

     Nature Neuroscience 23:1421–1432.

     https://doi.org/10.1038/s41593-020-00711-6

     • PubMed
     • Google Scholar
123. 1. Trudel N
     2. Scholl J
     3. Klein-Flügge MC
     4. Fouragnan E
     5. Tankelevitch L
     6. Wittmann MK
     7. Rushworth MFS

     (2021) Polarity of uncertainty representation during exploration and exploitation in ventromedial prefrontal cortex

     Nature Human Behaviour 5:83–98.

     https://doi.org/10.1038/s41562-020-0929-3

     • PubMed
     • Google Scholar
124. Preprint

     1. Tsay JS
     2. Kim HE
     3. McDougle SD
     4. Taylor JA
     5. Haith A
     6. Avraham G
     7. Krakauer JW
     8. Collins A
     9. Ivry R

     (2023) Fundamental processes in sensorimotor learning: Reasoning, refinement, and retrieval

     PsyArXiv.

     https://doi.org/10.31234/osf.io/x4652

     • Google Scholar
125. 1. Tustison NJ
     2. Avants BB
     3. Cook PA
     4. Zheng Y
     5. Egan A
     6. Yushkevich PA
     7. Gee JC

     (2010) N4ITK: improved N3 bias correction

     IEEE Transactions on Medical Imaging 29:1310–1320.

     https://doi.org/10.1109/TMI.2010.2046908

     • PubMed
     • Google Scholar
126. 1. Valk SL
     2. Kanske P
     3. Park BY
     4. Hong SJ
     5. Böckler A
     6. Trautwein FM
     7. Bernhardt BC
     8. Singer T

     (2023) Functional and microstructural plasticity following social and interoceptive mental training

     eLife 12:e85188.

     https://doi.org/10.7554/eLife.85188

     • PubMed
     • Google Scholar
127. 1. Váša F
     2. Seidlitz J
     3. Romero-Garcia R
     4. Whitaker KJ
     5. Rosenthal G
     6. Vértes PE
     7. Shinn M
     8. Alexander-Bloch A
     9. Fonagy P
     10. Dolan RJ
     11. Jones PB
     12. Goodyer IM
     13. consortium N
     14. Sporns O
     15. Bullmore ET

     (2018) Adolescent tuning of association cortex in human structural brain networks

     Cerebral Cortex 28:281–294.

     https://doi.org/10.1093/cercor/bhx249

     • PubMed
     • Google Scholar
128. 1. Vatansever D
     2. Menon DK
     3. Stamatakis EA

     (2017) Default mode contributions to automated information processing

     PNAS 114:12821–12826.

     https://doi.org/10.1073/pnas.1710521114

     • PubMed
     • Google Scholar
129. 1. Vázquez-Rodríguez B
     2. Suárez LE
     3. Markello RD
     4. Shafiei G
     5. Paquola C
     6. Hagmann P
     7. van den Heuvel MP
     8. Bernhardt BC
     9. Spreng RN
     10. Misic B

     (2019) Gradients of structure–function tethering across neocortex

     PNAS 116:21219–21227.

     https://doi.org/10.1073/pnas.1903403116

     • Google Scholar
130. 1. Vos de Wael R
     2. Benkarim O
     3. Paquola C
     4. Lariviere S
     5. Royer J
     6. Tavakol S
     7. Xu T
     8. Hong S-J
     9. Langs G
     10. Valk S
     11. Misic B
     12. Milham M
     13. Margulies D
     14. Smallwood J
     15. Bernhardt BC

     (2020) BrainSpace: a toolbox for the analysis of macroscale gradients in neuroimaging and connectomics datasets

     Communications Biology 3:103.

     https://doi.org/10.1038/s42003-020-0794-7

     • PubMed
     • Google Scholar
131. 1. Vos de Wael R
     2. Royer J
     3. Tavakol S
     4. Wang Y
     5. Paquola C
     6. Benkarim O
     7. Eichert N
     8. Larivière S
     9. Xu T
     10. Misic B
     11. Smallwood J
     12. Valk SL
     13. Bernhardt BC

     (2021) Structural connectivity gradients of the temporal lobe serve as multiscale axes of brain organization and cortical evolution

     Cerebral Cortex 31:5151–5164.

     https://doi.org/10.1093/cercor/bhab149

     • Google Scholar
132. 1. Vyas S
     2. Golub MD
     3. Sussillo D
     4. Shenoy KV

     (2020) Computation through neural population dynamics

     Annual Review of Neuroscience 43:249–275.

     https://doi.org/10.1146/annurev-neuro-092619-094115

     • Google Scholar
133. 1. Wan X
     2. Cheng K
     3. Tanaka K

     (2015) Neural encoding of opposing strategy values in anterior and posterior cingulate cortex

     Nature Neuroscience 18:752–759.

     https://doi.org/10.1038/nn.3999

     • PubMed
     • Google Scholar
134. 1. Wilson RC
     2. Bonawitz E
     3. Costa VD
     4. Ebitz RB

     (2021) Balancing exploration and exploitation with information and randomization

     Current Opinion in Behavioral Sciences 38:49–56.

     https://doi.org/10.1016/j.cobeha.2020.10.001

     • PubMed
     • Google Scholar
135. 1. Wolpert DM
     2. Ghahramani Z
     3. Flanagan JR

     (2001) Perspectives and problems in motor learning

     Trends in Cognitive Sciences 5:487–494.

     https://doi.org/10.1016/s1364-6613(00)01773-3

     • PubMed
     • Google Scholar
136. 1. Wolpert DM
     2. Diedrichsen J
     3. Flanagan JR

     (2011) Principles of sensorimotor learning

     Nature Reviews. Neuroscience 12:739–751.

     https://doi.org/10.1038/nrn3112

     • PubMed
     • Google Scholar
137. 1. Wu HG
     2. Miyamoto YR
     3. Gonzalez Castro LN
     4. Ölveczky BP
     5. Smith MA

     (2014) Temporal structure of motor variability is dynamically regulated and predicts motor learning ability

     Nature Neuroscience 17:312–321.

     https://doi.org/10.1038/nn.3616

     • PubMed
     • Google Scholar
138. 1. Zhang Y
     2. Brady M
     3. Smith S

     (2001) Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm

     IEEE Transactions on Medical Imaging 20:45–57.

     https://doi.org/10.1109/42.906424

     • PubMed
     • Google Scholar
139. 1. Zhao Q
     2. Kwon D
     3. Pohl KM

     (2018) A riemannian framework for longitudinal analysis of resting-state functional connectivity

     Medical Image Computing and Computer-Assisted Intervention 11072:145–153.

     https://doi.org/10.1007/978-3-030-00931-1_17

     • PubMed
     • Google Scholar

Author details

1. Qasem Nick

   1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
   2. Department of Psychology, Queen’s University, Kingston, Canada

   Contribution

   Resources, Software, Formal analysis, Validation, Investigation, Visualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

2. Daniel J Gale

   Centre for Neuroscience Studies, Queen’s University, Kingston, Canada

   Contribution

   Resources, Software

   Competing interests

   is an employee of Voxel AI Inc. This funding source had no role in the design, management, data analysis, presentation, or interpretation and write-up of the data

3. Corson Areshenkoff

   1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
   2. Department of Psychology, Queen’s University, Kingston, Canada

   Contribution

   Resources, Investigation, Methodology, Formal analysis, Software

   Competing interests

   No competing interests declared

4. Anouk De Brouwer

   Centre for Neuroscience Studies, Queen’s University, Kingston, Canada

   Contribution

   Formal analysis, Investigation, Methodology, Resources

   Competing interests

   No competing interests declared

5. Joseph Nashed

   1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
   2. Department of Medicine, Queen's University, Kingston, Canada

   Contribution

   Investigation, Methodology

   Competing interests

   No competing interests declared

6. Jeffrey Wammes

   1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
   2. Department of Psychology, Queen’s University, Kingston, Canada

   Contribution

   Resources

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8923-5441

7. Tianyao Zhu

   Centre for Neuroscience Studies, Queen’s University, Kingston, Canada

   Contribution

   Resources, Formal analysis

   Competing interests

   No competing interests declared

8. Randy Flanagan

   1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
   2. Department of Psychology, Queen’s University, Kingston, Canada

   Contribution

   Conceptualization, Resources, Writing – review and editing

   Competing interests

   No competing interests declared

9. Jonny Smallwood

   1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
   2. Department of Psychology, Queen’s University, Kingston, Canada

   Contribution

   Visualization, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

10. Jason Gallivan

    1. Centre for Neuroscience Studies, Queen’s University, Kingston, Canada
    2. Department of Psychology, Queen’s University, Kingston, Canada
    3. Department of Biomedical and Molecular Sciences, Queen’s University, Kingston, Canada

    Contribution

    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Writing – original draft, Project administration, Writing – review and editing, Methodology

    For correspondence

    gallivan@queensu.ca

    Competing interests

    is an employee of Voxel AI Inc. This funding source had no role in the design, management, data analysis, presentation, or interpretation and write-up of the data

    "This ORCID iD identifies the author of this article:" 0000-0002-7362-109X

• 1,058

  views

• 76

  downloads

• 2

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.