Abstract
This study aims to establish an imitation task of multi-finger haptics in the context of regular grasping and regrasping processes during activities of daily living. A video guided the 26 healthy, right-handed volunteers through the three phases of the task: (1) fixation of a hand holding a cuboid, (2) observation of the sensori-motor manipulation, (3) imitation of that motor action. fMRI recorded the task; graph analysis of the acquisitions revealed the associated functional cerebral connectivity patterns. Inferred from four 60 ROI weighted graphs, the functional connectivities are consistent with a motor plan for observation and manipulation in the left hemisphere and with a network in the right hemisphere involving the inferior frontal gyrus, the site of intentional control of imitation. The networks exhibit (1) rich clubs which include sensori-motor hand, dorsal attention and cingulo-opercular communities for observation and motor execution in both hemispheres and (2) diversity clubs, significant only for manipulation and observation of the left hand, which include the dorsal visual association cortex, suggesting enhanced visual perception required for guiding the movement-limited left fingers. Short pathway analyses are consistent with these findings, confirming preferential involvement of ventral premotor cortices in the mirror network.
Similar content being viewed by others
Introduction
Classifying hand manipulations is required for understanding specific dexterity and impaired functionality. Here we studied human manipulation involving prehensile contact to a cuboid. It is prehensile as it cannot be represented by a single finger contact, but by finger movements relative to the reference frame defined by the hand base; moreover, the manipulation requires translation and rotation of the object1. The task integrated single motor acts into an action over time, leading to a dynamics of multiple independently controlled contacts to optimize within-hand object orientation, which is also important for spatial object recognition. This sequence of multifinger grasps and regrasps secures objects within the hand while executed at a frequency of routinely feasible 1 Hz. Such regularly changing grasp configurations have been explored artificially in multifinger robot hand manipulations and are termed finger gaiting. We adopted this term after we found three different finger configurations during one manipulation with a digital data glove2. The manipulation task mentioned above has been integrated into an imitation paradigm with an observation phase prior to each motor execution phase to convey the formal and temporal conditions.
Our previous explorations of the active touch in grasping and regrasping, using fMRI and principal component analysis (PCA), have indicated distinct sensori-motor patterns reflecting hierarchically ordered networks3,4. In a recent lesion study post-stroke, we delineated a prediction map for persistently impaired subjects relying on multi-voxel pattern analysis. The regions of interest (ROIs) derived from the areas involved in these three studies are detailed in Table ST1 and serve as basis for the exploration of networks underlying the specific phases of this study. In essence, the included areas form a sensorimotor network and contain neurons with mirror properties according to studies in macaques5. It should be noted that mirror neurons apparently acquire their mirror properties through sensorimotor learning processes and change them by the same route; furthermore, observed motor actions are processed rather on a low, i.e. most likely implicit level6.
The constituents of the ventral premotor cortex (vPMC), the inferior parietal lobule subarea PFt and the inferior frontal gyrus (IFG) Broca area are thought to represent core areas of the mirror neuron system7,8. In primates motor hand actions and reaching have been shown to involve areas outside of the classical mirror neuron system. These areas are endowed with mirror properties and therefore part of the extended mirror system. These include the specific premotor F5 and inferior parietal areas of the classical IPL-F5 circuit9,10, as well as the primary motor area (M1), dPMC and SMA11, IPL and intraparietal sulcus (IPS)10, and parietal-occipital cortices (POC)12.
Consistent with these patterns, functional neuromaging in humans showed overlapping areas across conditions of motor observation, execution and imagery in dPMC, SMA, M1, primary somatosensory cortex (S1), superior parietal lobule (SPL), IPL and precuneus in addition to constituents of the classical mirror neuron system13. Specifically, the action observation network (AON) appears similar to the motor execution network of which inferior frontal gyrus (IFG), ventral premotor cortex (vPMC) and IPL are subcomponents14. Subtle differences in activation patterns may reflect somatotopic organization, e.g. reaching may activate preferentially dPMC and SPL15, whereas intrinsic hand movements activate rather ventral areas such as IFG, vPMC and IPL for conditions of observation and execution16. In fact, it is unclear whether the subtle differences are action or effector specific17, i.e. dependent on action planning or automatically evoked by multiple parieto-frontal networks18.
It should be noted that subcortical structures, e.g. thalamus, putamen and cerebellum, are involved in the mirror system. In particular, the cerebellum is also involved during the observation phase19,20,21. The latter may be triggered by cognitive components associated with the task; in difficult tasks, for example, the cerebellum can contribute to visuomotor matching, since the human cerebellum is actively involved in perceiving the kinematics of the hand actions of others. Thus, the most intriguing and direct evidence for the mirror mechanism is the bilateral recruitment of cerebellar lobes when observing purposeful hand movements of others19.
Using graph theoretical analysis we intended to describe the cortical hemispheric nodes in terms of node strength and participation coefficient, and the modular structure of the specific networks. We have prioritized our aims. The primary aim was to elucidate the functional structure of the cortical network underlying the phases of the sensori-motor hand imitation task from the perspective of high local clustering into communities and short pathways linking them. A specific aim was the expression of the classical mirror neuron network associated with the exclusive within hand motor action of grasping and regrasping, independent of an additional upper limb reaching component.
Results
Behavioral data
Sensori-motor behavioral data
The results of the sensori-motor assessments were normal for power and precision grip22, two point discrimination23, picking small objects (PSO) and tactile object recognition (TOR)24. They are shown in Table ST2 of Supplementary Material and not commented further.
Digital glove data. Median finger movement frequencies of the 4 sensors in the thumb joints were 1.03 and 0.97 s− 1 for the right and left hands, respectively. Fingers 2 to 5 exhibited median frequencies in the four metacarpo-phalangeal (MCP) joints of 1.02 and 0.95 s− 1, in the four proximal interphalangeal (PIP) joints of 1.03 and 0.95 s− 1 in the right and left hands, respectively. These differences between right and left hands are significant (Table 1). Graph theory analysis of the time series of the 12 bend sensors, associated with the MCP and PIP joints of the fingers as well as interphalangeal joint (IP) of the thumb together with thumb cross and palm arc sensors, yielded a weighted, undirected graph incorporating all subjects with 12 nodes and 66 edges for each hand. These revealed three identical modules in the right and left hand. The edges of the subnetworks within the same joints, i.e. that of the PIP, MCP, and thumb joints, had positive weights; whereas those between the groups of PIP and MCP joints to the thumb joints had negative weights. The temporally asynchronous movements of PIP and MCP joint of fingers 2 to 3 as related to the thumb joints appear reduced and partly extinguished in the left hand (Fig. 1). The median global efficiencies of the unthresholded networks were 0.839 (IQR 0.806–0.869) for the left and 0.840 (IQR 0.805–0.874) for the right hand, which is not a significant difference (Table 1). Behavioral data acquired by close-up video recording during fMRI determined the manipulation frequency of the thumb, which guided the rotation of the cuboid, to be 0.95 s− 1 for the right and the left hand.
Representations of active manipulation based on graph analysis of digital glove data. These representations for left and right hand manipulation show three symmetrical modules: (i) module 1 (green) encompasses PIP joints of finger 2 to 5 and IP joint of thumb, (ii) module 2 (blue) encompasses MCP joints of fingers 2 to 5 and (iii) module 3 (red) encompasses MCP and CMC joints of thumb and palm arch. The upper row shows the interrelation of finger joints exhibiting positive weights, suggesting synchrony of joint movements within the modules. The lower row shows the interrelation of finger joints exhibiting negative weights, suggesting asynchrony among the modules. The thickness of connections between joints reflects the relative weight of the interrelations. Note asynchrony between PIP finger joints and IP thumb joint, determining extent of dynamic finger movements, is reduced in the left hand. (Networks were illustrated using Gephi (Release 0.10.1). Hands of the first author are depicted.)
Neuroimaging data: task based fMRI
Statistical parametric mapping
First level analysis of all conditions for each hand and subject yielded statistical parametric maps of two contrasts: observation and manipulation contrasted with the reference state, fixation. A 2nd level group analysis using a two-tailed t-test for each contrast was then carried out. This second level analysis showed significant results at p < 0.05 (corrected for family-wise error (FWE)) and an extension threshold level of 20 voxels. For a detailed listing of activations see Fig. 2 and Tables ST3 and ST4 (Supplementary Material).
SPM analysis of the four main conditions. (A) Manipulation right and left versus baseline and (B) observation of manipulation by the right and left hand versus baseline at significance level p of delineated clusters < 0.05 (corrected for FWE), and extension threshold 20 voxels. Activations are integrated in axial slices of mean T1-images related to the study cohort. Ventral PMC/posterior Broca area BA 44, a node of the classical mirror network, appears preferentially in the right hemisphere during motor execution of both hands in the transversal slice of MNI z-level 30 (A) but not during action observation of hands (B). For list of activation areas, including specific T-values of the involved brain areas, see Supplementary Material, Tables ST4 and ST5.
Graph theory analysis
The first level analysis of the thresholded 113-ROIs weighted undirected (WU) graphs and the second level analysis of the thresholded 60 ROI WU graphs (parameters in brackets) exhibited the following global parameters: for (1) object manipulation with the left hand: 2725 (1060) edges, a density of 0.43 (0.6) and a small world propensity ϕ of 0.47 (0.57); for (2) object manipulation with the right hand: 2936 (1143) edges, a density of 0.46 (0.65) and a small world propensity ϕ of 0.42 (0.51) ; for (3) observation of object manipulation with the left: 2901 (1159) edges, a density of 0.46 (0.66) and a small world propensity ϕ of 0.39 (0.54); and for (4) object manipulation with the right hand: 2523 (965) nodes, a density of 0.40 (0.55) and a small world propensity ϕ of 0.46 (0.55). ROI strengths and participation coefficients of the 113-ROI graphs are listed in the repository. The surface area of the 99 cortical ROIs ranged from 16.0 to 1935.3 mm2 with median 816.8 mm2. No significant correlation was determined between the strength and the area of ROIs.
In keeping with the greater small world propensities ϕ across conditions, we treat in the following only the 60-ROI graph. Each of the activation conditions and hand yielded a specific graph. Strengths and participation coefficients of selected ROIs are listed for both the unthresholded and the thresholded graph in Table ST5 and ST6 of Supplementary Material. The former served as basis for the assessment of the rich and diversity club, the latter is associated with higher optimal global and local efficiencies. Assigned to communities defined by Gordon et al.25 and the Oxford thalamic atlas26, the ROIs belong to sensori-motor hand community (c.) (n = 16), dorsal-attention c. (n = 10), cingulo-opercular c. (n = 12), cingulo-parietal c. (n = 2), salience c. (n = 2), fronto-parietal c. (n = 3), visual c. (n = 4), default network c. (n = 8), auditory c. (n = 1), and thalamic c. (n = 2).
The evaluation of the ROI matrices showed an increasingly better trade-off between global and local efficiency up to a threshold value of 0.45, after which the network becomes unstable (see Fig. 3). At this threshold, the mean (standard deviation) of the normalized path lengths across conditions was 0.858 (0.101). For the mean cluster coefficient, these measures were 0.831 (0.083). Areas exhibiting exclusively high local strength25 were all core subareas of sensori-motor c., core areas of the dorsal attention c., and core areas of the cingulo-opercular network, including the SMA compartment together with paralimbic motor anterior cingulate cortex (ACC) and dorsal ACC. The areas exhibiting exclusively high participation coefficients26 were: core areas of default mode network (DMN) module, e.g. IPL PGa, IPL PGp, forebrain subareas Fp1, Fp2; dorsal visual association cortices, e.g. subareas hOc4d and hOc3d; the cingulo-parietal cortex of subarea SPL 7P; the salience cortex of subarea ventral ACC; the parietal operculum of OP1 and OP4; the frontal parietal operculum of subareas dorsal-inferior Broca area BA44 together with the vPMC and Broca area BA 45, and the medio-dorsal thalamic nucleus. Subareas of the dorsal ACC and of SPL 7A,7P, and the paralimbic motor ACC exhibited both high strength and participation coefficients. They constituted about 4% of all nodes of the 60 ROI graphs for all conditions.
Display of three parameters related to graph efficiency in the thresholded 60 ROI WU graph. (A) The upper subgraph shows global efficiency which is related to the average inverse shortest path. (B) The middle subgraph shows local efficiency, which is computed from the neighbourhoods of all nodes and related to the clustering coefficient. (C) The small world propensity indicates the extent to which a network is characterized by high local clustering and low average path length. Note (1) a threshold of 0.45 is the optimal trade-off between high global and local efficiencies and associated high small world propensities, of which the latter are superior to those of the 113-ROI graphs (for comparison see the corresponding figure in the repository), (2) the stable evolution of small world propensity until a threshold of 0.45 for all conditions, beyond which it becomes unstable (dotted lines).
Clubs and club coefficients within 60-ROI graph
Rich and diversity clubs were determined for the unthresholded graphs. The unthresholded 60 ROI graphs computed for the four conditions are represented in Figure SF1 of Supplementary Material. As shown in Fig. 4, rich and diversity clubs were computed for all conditions; all unnormalized club coefficients (turquoise) were negative. The high strength nodes for all conditions yielded unnormalized rich club coefficients decreasing to -0.9 or less with increasing percentile. The normalized coefficients, measuring deviation from randomness and essential to evaluate significance, increased to approximately 1.1 before the deviations produced by the random graphs inflated at about 80 percentile. The maximum amplitude for which the lower confidence interval (Cl) exceeds 1 is greatest for observation of right hand manipulation and least for manipulation left, and occurs at about 70 percentile for all conditions. The decreases of the unnormalized diversity club coefficients were much larger than those of the unnormalized rich club coefficients, decreasing to less than − 0.6 for percentiles exceeding 70. Moreover, the normalized club coefficients exceeded 1 before inflation only for manipulation and observation of manipulation with the left hand, where the maximum amplitudes for which the lower Cl exceeds 1 occurred at the 90th percentile.
Rich and diversity club determination across conditions. The upper row shows club coefficients of the rich clubs for the four task conditions as a function of percentile, the lower row the diversity club coefficients. The unnormalized club coefficients are indicated in turquoise. The solid black line denotes the normalized club coefficients based on 1000 null model networks; the broken blue lines enclose the 95% CI. The percentiles are derived from the nodes’ strength for the rich club and from the participation coefficients for the diversity club. Note: Club membership requires a coefficient of > 1. The red symbols denote the maximal club coefficient together with the associated percentile of parameter.
As shown in the upper row of Fig. 5, the rich clubs for all four conditions, symmetrically distributed in both hemispheres, include areas of sensori-motor hand community (c.) such as subarea 3b and precuneus, and subarea SPL 7A,5M of dorsal attention c. Additional members are areas of cingulo-opercular c. such as SMA and paralimbic motor ACC bilaterally. Furthermore, members associated only with manipulation of both hands, and appearing symmetrically in both hemispheres, are subareas SPL 7A,7P and IPS hIp1 of dorsal attention c. These areas and subareas form the core of the prominent high strength areas also of the thresholded 60-ROI graph. As shown in the lower row of Fig. 5, the diversity clubs for observation of manipulation and execution with the left hand includes visual subareas hOc3d and hOc4d in both hemispheres. In observation of manipulation with the left hand only emerge connections of the subareas IPS hIp3, IPL PGp and medio-dorsal thalamic compartment of the left hemisphere, the latter connected to prefrontal cortex. In manipulation with the left hand only emerge connections of IPS hIp3 bilaterally and Broca area BA 45 of the right hemisphere. In the thresholded 60-ROI graph, the prominent high participation coefficient nodes prevailed in subareas hOc3d and hOc4d during observation and in subareas hOc3d, hOcd4, IPL PGp, PGa, SPL 7P and IPS hIP3 during executive control of motor imitation. These areas emerge for both hands and show rather symmetrical expression.
Neuroanatomical representations of rich and diversity clubs for all conditions. The glass brain representation of the unthresholded 60 ROI WU graphs show the club members as yellow centroids related to their areas. Communities of nodes are color-coded as specified in the legend annotation. Edges exceeding a weight of 0.6 are color coded; green denotes weights between 0.6 and 0.7, yellow weights between 0.7 and 0.8, blue weights between 0.8 and 0.9 and red weights greater than 0.9. Rich club membership required that percentile of node strength exceed 70 percentile, and diversity club membership, that percentile of participation coefficient exceed 90 percentile. Some critical nodes are designated in black script to facilitate orientation. Th prefr denotes the functional thalamic area connected to prefrontal cortex.
Functional connectivity and delineation of the shortest paths
For observation of manipulation by the left and right hand, a two-step path proceeds bilaterally via SPL 7A,7P and 7A,5M subareas to the dorsal, but not to the ventral premotor cortex; a second two-step path emerged in the right hemisphere proceeding via the SPL hIP1 subarea exclusively to the dorsal premotor cortex. For manipulation of left and right hand, a three-step path proceeds via IPL PGa and SPL hIP1 to the dorsal premotor cortex and a two- or three-step path to the ventral premotor cortex. The IPL PGa subarea constituted the second node of the latter path. From there a two-step path proceeded directly in the left hemisphere for manipulation left and in the right hemisphere for manipulation right, whereas a three-step path proceeded via the parietal operculum OP1 subarea of the right hemisphere for manipulation left and via Broca area BA 45 of the left hemisphere for manipulation right.
Important points to note in Table 2 and the left column of Fig. 6 are the fewer functional edges required in the right hemisphere than in the left for manipulation by both hands. For all conditions, the anterior precuneus, the sensory hand subarea 3b and the lateral premotor cortices were strongly interconnected with Pearson correlation coefficients exceeding 0.8. A subnetwork appears in the right hemisphere encompassing subareas visual hOc3d and hOc4d, IPL PGa and PGp, cingulo-parietal 7P, intraparietal hIp3 and SPL 7A,7P, which necessitates a minimal number of high weight edges to become functional (Table 2). The visual subarea hOc3d and the parietal subareas described above are mutually interconnected (Fig. 6). Table 2 indicates in detail functional connectivities in the context of shortest paths as described above.
Short pathways exemplified by left hand execution and action observation. Glass brain representation of the 60 ROI WU graphs thresholded at 0.45 show in the upper row short pathways for manipulation left and in the lower row for action observation of the left hand (marked in bold). A detailed summary is given in Table 2. Nodes are color-coded as specified in the legend annotation in Fig. 5. The left column (A) depicts the neuroanatomical context. The middle column (B) shows the associated internal structures of local high functional connectivity converging to the SMA, which is practically identical for both conditions. The right column (C) shows a suggested recursive loop originating at hOc3d in the occipito-parietal transition zone of the right hemisphere and strongly interconnected with subareas IPS hIP3, IPL PGa and PGp, which is functional during motor execution but not during action observation. The size of the nodes reflects the size of participation coefficients in (A) and of strength in the subfields of the (B) and (C). Edges exceeding a weight of 0.6 are color coded; green denotes weights between 0.6 and 0.7, yellow weights between 0.7 and 0.8, blue weights between 0.8 and 0.9, and red weights greater than 0.9.
Discussion
In healthy subjects, matched in terms of age to a stroke population, we investigated a prototypical somatosensory task underlying object shape exploration using two methods: the VMG 30™ data glove and fMRI. The task represents one component of daily motor skills based on somatosensory exploration and grasping, which are sensitive measures of residual post-stroke capabilities and powerful indicators of recovery potential27,28. Our analysis of the finger movements recorded by the VMG 30™ data glove confirmed and refined the findings of a previous study of a slightly smaller subject cohort2. From the time series of the main component found in the current study, we established in a graph theoretical analysis a completely symmetrical interaction in three groups of MCP, PIP, and thumb joints of both hands. These groups constitute separate modules in which the joints moved synchronously within and asynchronously among groups. The networks of both hands exhibited small world characteristics with no significant indication of laterality. They also showed high global efficiency suggesting a form of information exchange between modules. The frequency of left fingers and the temporal asynchrony across finger joints of the left hand, determining the extent of movements, was reduced compared to the right hand side (Fig. 1; Table 1).
In the SPM analysis of fMRI acquisitions of the task, statistical parametric maps of action observation and motor execution compared to fixation supported our choice of ROIs to serve as nodes in a graph theoretical analysis. Our graph theoretical analysis of these ROI nodes representing observation and manipulation with both hands showed as expected functional reconfigurations of the neural networks differentiating between observation and execution29,30. Recently, Hahn et al.31 found convergence between metabolic and neural processes specific to such feedforward connections linking visual and dorsal attention networks in accordance with the task requirements of visuo-spatial reasoning. Analysis of the 113 node with regard to node strength and participation coefficient - crucial parameters reflecting network functionality32 - indicated that an analysis of a reduced set of 60 ROIs would be appropriate.
In accordance with the proposals of Heuvel et al.33 and Bertolero et al.34 we used node strength and participation coefficient to calculate the normalized club coefficients of rich and diversity clubs, respectively, for the unthresholded 60-ROI graph. High node strength characterizes the weights of the nodes local interconnection, whereas high participation coefficient indicates diverse distribution of connections among different communities34. The rich club presents a distinct symmetric pattern especially for midline nodes, significant for manipulation and observation of both hands, whereas the diversity club showed functional significance only for manipulation by and observation of the left hand. This may be associated with the specialization of the left hand as regards stability while holding, necessitating greater effort during observation to prepare for the performance of dynamic spatial finger movements with the same manual efficiency as with the right hand35,36.
Constituents of rich club included for both conditions and both hemispheres were: subarea 3b and dorsal-anterior precuneus of the sensori-motor hand community (c.), subarea 7A,7P and 7A,5M of the dorsal attention c., and SMA together with the associated paralimbic motor ACC and dorsal ACC of the cingulo-opercular c., where the latter community participates in the willed control of motor activity37. The SMA corresponds exactly to area 6mp: SMA and mesial superior frontal gyrus, according to recent work using probabilistic cytoarchitectonic mapping38,39. Furthermore, the subareas SPL 7A,7P as well as IPS hIP1 appear in manipulation of both hands, suggesting short-term learning by observation of complex action sequences40, or an effect of short term working memory involved in the reorganization of information at a critical site mediating sensory and haptic processing41,42. As evidenced by high nodal strength, the anterior-dorsal part of precuneus plays a prominent role in both action observation and motor manipulation, which is consistent with the finding of Sakai et al.43 that the dorsal anterior precuneus was specifically involved in the transition from declarative learning of a visuomotor sequence to the effective execution. Together with other areas of the medial aspect of hemispheres, i.e. the dorsal ACC and paralimbic motor ACC, the precuneus represents an intentional component of self, the consciousness of self-agency, involved also in self-centered imagery44,45. Most important, the cingulo-opercular c. may integrate the effector specific M1 hand area into a somato-cognitive action network critical for action planning and bodily control46. The distinct role of the cingulo-opercular network in manipulation is the maintenance of increased tonic alertness and sustained task-related control47,48.
Constituents of the diversity club observed in manipulation left and observation of manipulation with the left hand included the subareas hOc3d and hOc4d in the medial dorsal visual cortex of both hemispheres, the former corresponding to the functional area V3ab and the latter to the functional area V3d49. Subarea hOc3d is especially sensitive to high frequency kinematics and functional in visuomotor feedback, visually guided pointing and spatial relocation50. Subarea hOc4d, located rostrally to subarea hOc3d, is active in higher level processing of 3D object structures and 3D finger positions49. In observation of manipulation further constituents were the IPS subarea hIP3 and the associated subarea PGp of the angular gyrus of left hemisphere51; in addition, the left medio-dorsal thalamus was connected to the prefrontal cortex26,52. This pattern of constituents within the left hemisphere suggests common processing of visual 3D information and working memory processes in fronto-parietal regions activated by prefrontal thalamic support53,54,55. During task performance, additional nodes such as the IPS subareas hIP3 bilaterally were connected with the superior occipital areas55 and the right BA 45, indicating a switch to the right hemisphere, an indication of implicit recognition of the type of movement to be imitated as evoked by the video6,41,56.
At a threshold level of 0.45, the 60-ROI graph shows an optimal trade-off between global and local efficiency, with an enhanced and, thus, cost-efficient small world propensity57. There the high strength areas correspond to those of the rich clubs. However, the nodes with highest participation coefficients include also subarea IPL PGp in the right as also subareas IPL PGa and cingulo-parietal SPL 7P in both hemispheres which are not members of the diversity clubs, whereas the dorsal attention subarea IPS hIP3 and the only visual subarea hOc3d in the right hemisphere are. This apparent discrepancy may indicate more haptic information transfer and less visual control at that level, reflecting most likely motor routine. These areas are involved in spatial attention and working memory processes but characterized by diverse connectivities58: subarea PGp is highly connected to disparate nodes of DMN, subarea PGa to the ventral premotor and ventrolateral prefrontal cortex, and hIP3 preferentially to the superior occipital cortex. In a recent microstructural update51, a dense connectivity between IPL PGa and IPS hIP1 subareas has been described. In the context of transmodal information processing, a previous study59 showed a distinct activation of the left angular gyrus while subjects imagined manual construction of object shapes with the right hand. In another study60, visual to haptic matching of 3D-dot arrays caused differential activation of this site. Thus, IPL appears in a position to integrate visual information with visuo-motor transformation processes61,62. The SPL 7P subarea belongs to the small cingulo-parietal community, which may indicate perception of agency since it is connected with the dorsal-posterior compartment of precuneus45,63. Regarding the visuo-spatial characteristics, the marked appearance of the right extrastriate subarea hOc3d, corresponding to the functional area V3d, suggests a lateralization to the right hemisphere of dynamic task performance50.
The networks described above represent the specific transfer of visual information from occipital to parietal cortices - a matter of debate for many years. Studies in macaques and humans revealed in the anterior wall of the medial parieto-occipital sulcus (POS) an additional retinotopically organized extrastriate area, i.e. area V6. This borders (1) caudally the occipital hOc3d and hOcd4d subareas64,65,66 and (2) dorsally a non-retinotopically organized and cytoarchtectonically mixed area V6a, belonging rather to the SPL67. Area V6 has been shown to be engaged in visuo-motor information of relevance to egocentric activity. The dorsomedial stream, including anterior IPS and area V6A, is involved in diverse aspects of skilled motor hand activity like grasping, necessitating fast control of prehension68.
The analysis of shortest paths underlying the functional connectivity during motor execution from visual subarea hOc4d, the posited site for processing 3D finger position and motion49, to dorsal and ventral premotor cortices is summarized in detail in Table 2. Subarea IPL PGa assumes a prominent role in both routes after transmodular information processing of the visual percept, which occurs most likely in the heteromodal association compartment of the angular gyrus at subarea PGp69,70. Here, the anterior part of the angular gyrus accesses directly or indirectly the ventral PMC and motor part of Broca area BA 4471,72, whereas connection to the dorsal PMC proceeds via the anterior IPS, which is directly connected to subarea PGa51. Indirect connection to Broca area BA 44/vPMC proceeds via subarea OP1 of the parietal operculum in the right hemisphere; in the left hemisphere, the connection includes BA 45. OP1 in the right hemisphere is densely interconnected with subareas hIp1 and hIp3 of the IPS14. Diffusion MRI studies have shown the superior longitudinal fasciculus III to project directly in varying degrees from the angular gyrus to the ventral premotor cortex, Broca areas BA 44 and 4573,74,75.
The shortest paths for task observation do not include the ventral PMC. However, shortest paths to the dorsal PMC do include SPL subareas 7A,7P and 7A,5M in both hemispheres and IPS hIP1 in the right. These areas are engaged most likely in the control of information during transient working memory phases42. The lack of shortest paths to the ventral premotor cortices during observation suggests a dorsal-ventral dissociation of representations of movement in manipulation from observation of manipulation76, i.e. in contrast to observation, motor execution seems to involve reliably the ventral PMC and Broca area BA 44 in this task, manifesting the classical mirror neuron system13. The dissociation of the dorsal and ventral PMC has also been observed in studies of virtual lesions induced through transcranial magnetic stimulation; the coupling of grasping and lifting phases of objects mediated by the dorsal PMC was dissociated from the positioning of fingers on the objects and sequential recruitment of intrinsic hand muscles regulated by the ventral PMC77. A loop involving relatively few functional, i.e. high weight edges, appears in the right hemisphere. This loop connects visual subareas hOc3d and hOc4d involved in perceiving information about visual kinematics and 3D finger positions (see above); IPL subareas PGp and PGa that generate the haptic percept and transfer kinematic information (see above); subarea IPS hIP3 implicated in visuo-motor control51; SPL 7A,7P in selective attention78; and SPL 7P in perception of agency45. The dense interconnections between visual and parietal cortex solely during motor execution suggests a recursive loop (Fig. 6) from and to the visual subarea hOc3d that complements the motor plan to foster and adapt adequate motor performance79.
In terms of limitations, it should be noted that the WU graphs presented do not provide any information about the directionality of the connections between the nodes. However, we were able to delineate systemic connectivities transmitted by the dorsal visual stream expressing specific patterns of observation, i.e. visual guidance, followed by motor execution. Not including the cerebellum and basal ganglia in the graph-theoretical analysis is a limitation, as these structures are connected to the mirror system19,20,21. It should be noted that the study focuses on the ROIs with the highest strength and participation coefficients to delineate the predominant nodes and pathways of the neural networks.
Subjects and methods
The study has been pre-registered in (https://clinicaltrials.gov/ct2/show/NCT02865642).
Subjects
Twenty-six healthy right-handed subjects, 13 males and 13 females ranging in age between 42 and 84 years (median 61), were included in the study. Their handedness scores ranged between 80 and 100 (median 100) according to the Edinburgh Handedness Questionnaire80. The subjects had no prior history of neurological and psychological disorders, achieved normal Mini-Mental State Examination (MMSE) scores81 with a range between 26 and 29 (median 29), and showed no abnormalities in MRI brain scans. The study received ethical approval from the Kantonale Ethikkommission Bern (KEK), 3010 Bern, Switzerland. Prior to the study all participants gave written informed consent before enrollment, according to the Declaration of Helsinki82. All methods were performed in accordance with the relevant guidelines and regulations.
Methods
Behavioral data
Sensori-motor assessment
All described sensori-motor functions were measured in both hands, starting with the non-dominant left hand. Fine motor skill was measured using the Jebsen-Taylor Hand Function test, using subtest picking small objects (PSO)83. We quantified two-point discrimination (2PD) of the index finger tip using a graded caliper (2-point Discriminator, Medwork Instruments, Vancouver, Canada23). Tactile object recognition (TOR) was tested using a standardized protocol with 30 everyday objects as described by Bohlhalter et al.84. Power grip was measured using a Jamar hydraulic hand dynamometer and precision grip with a Jamar hydraulic pinch gauge22.
Experimental task
We derived the experimental task from the tactual macroscopic object exploration described by Roland and Mortensen85. The purpose was to minimize cognitive processing during the task, leaving exclusively the sensori-motor task involving dynamic finger manipulations of a cuboid. The cuboid to be manipulated was made of granite with a density of 2.6 g/cm3 and side lengths of 2.254 × 2.254 × 2.257 cm resulting in a total volume of 11.5 cm3 and weight of 29.9 g, comparable with those of the aluminum cube used in one of our previous studies4. The manipulation consists of regular single motor acts performed at a frequency of 1 Hz. It is characterized by dynamic movements of the opposed thumb and fingers of one hand around the cuboid in a continuous and regular action. Each phase begins with a transaxial movement of the thumb versus the ring finger. During the concerted action of thumb and fingers, the thumb exerts tangential forces that produce a marked rotation of the object, anticlockwise in the right hand, and clockwise in the left2.
A video was recorded to instruct the subjects how to perform the task. This instruction video consisted of three, 20 s long, consecutive segments: (1) fixation, (2) observation, (3) active manipulation, each announced by written instruction on a blank white screen for 4 s. “Fixation” showed a hand holding the cube providing a baseline. “Observation” showed the same hand manipulating the cuboid at the prescribed 1 Hz frequency. Upon “active manipulation” the subjects were given the cube by the study physician and requested to manipulate the cube as shown in the video sequence displayed on the screen during the segment “Observation”. During this last segment, the same static image was shown to the subjects as in the section “fixation”. A right hand was shown for the right hand sensori-motor task and a left hand for the left hand task. The 3 segments were repeated six times and in the same order, showing in three instances male and in three female hands.
For pre-scan examination we employed the VMG 30™ data glove of Virtual Motion Labs [Virtual Motion Labs, LLC., 3010 LBJ Freeway, Dallas, Texas 75234 (www.virtualmotionlabs.com)]. The glove is equipped with 29 sensors of which 16 are bend sensors. Two finger bend sensors per finger II to V measure movement extent in the metacarpo-phalangeal (MCP) and proximal interphalangeal (PIP) joints, and two finger bend sensors at the thumb measure movement extent in the MCP and interphalangeal (IP) joints. Four sensors between the fingers measure abduction. One palm arch sensor detects spatial configuration related to the proximal and distal transverse arch of the hand described by Hertling and Kessler86. One thumb cross sensor detects the complex movement of the thumb during finger opposition at carpo-metacarpal (CMC) joint. Five sensors situated at the finger tips measure pressure and eight sensors measure hand and wrist orientation2.
During task performance, subjects were seated at a desk on which was placed a computer screen. To ensure that the subjects understood the motor task, it was explained by the study physician and subjects were requested to manipulate the cuboid with the left and right hand without the data glove for about 10 seconds as shown by the physician. Then the calibrated data glove was put on the non-dominant left hand of the subject and checked for fit by the physician. The video was started when the subject’s hand was relaxed and supine on the desktop. When the instruction “Active manipulation” appeared on the screen, the physician placed the cuboid in the subject’s hand; it was removed after completion of the sequence of manipulation. The three segments were repeated without intermission 6 times, first by the non-dominant left and, after a short pause, by the right hand leading to a total duration of 432 s–7.2 min, i.e. (20 + 4 s)*3 segments*6 repetitions for each hand. The active manipulation sequence was recorded with custom in-house software based on the source development kit provided by the manufacturer of the VMG 30™ data glove from Virtual Motion Labs [Virtual Motion Labs, LLC., 3010 LBJ Freeway, Dallas, Texas 75,234 (http://www.virtualmotion-labs.com)].
Neuroimaging data
During the fMRI acquisitions, subjects lay supine inside the scanner with their heads immobilized. The investigator received the signal to present and remove the cuboids via light projection of the instruction from outside the scanner room. The instruction video was transmitted to a screen within the scanner. The test conditions, order of test sequences and durations were identical to the data glove acquisitions. The performance of tactual object manipulation was monitored by a close-up video recorder (Model 12 M, MRC Systems GmbH, Heidelberg, Germany). The sequence of the paradigm is shown in Figure SF2 (Supplementary Material).
All acquisitions were acquired using a 3 Tesla whole-body MR scanner (Siemens Magnetom Prisma, Siemens, Erlangen, Germany) using a standard head coil. After a localizer image was made, a structural-anatomical image was acquired in the anterior-posterior (AP) direction using a Modified Driven Equilibrium Fourier Transform (MDEFT) T1-Sequence. Covering the whole brain, the image consisted of 176 contiguous sagittal single-shot slices with a thickness of 1 mm (TR 2300 ms, TE 2.26ms, flip angle of 8°, phase encoding direction from anterior to posterior). For the activation task we used an echo planar imaging–blood oxygen level-dependent (EPI - BOLD) sequence during fMRI with 49 slices per volume and a thickness of 2 mm per slice (TR 3000 ms, TE 30 ms, flip angle of 90°, phase encoding direction from anterior to posterior). The duration of the paradigm, 432 s, required 144 volumes for each investigated hand. Detailed fMRI acquisition parameters are presented in Supplementary Material (Table ST7).
Data analysis and statistics
Behavioral data
Generated by the VMG 30™ data glove, the time series of the 12 MCP/PIP (IP) finger joints, palm arch and thumb cross sensors were first submitted to PCA and cluster analyses. These 12 bend sensors appeared dominant in these analyses. The sensors were subjected additionally to frequency and graph theoretical analysis. For the graph analysis, we used GraphVar87 (Release V2.01) as implemented in Matlab. From the bend sensor time series were calculated mean matrices over all subjects yielding a WU graph with 12 nodes and 66 edges for each hand. Negative weights, corresponding to negative correlations, were retained. To investigate subnetworks, the graphs were thresholded in steps of 0.05 for positive and negative weights. Global efficiencies for both graphs were calculated without thresholds. For details see methods by Krammer et al.2. As recorded in the video during the fMRI acquisition, the manipulation frequency of the thumb, which produces a marked rotation of the object, was determined.
Image data
The MRI data were preprocessed using the CONN-toolbox (version 19.b) default preprocessing pipeline for anatomical and functional images (https://web.conn-toolbox.org/). This toolbox is based on the SPM12 toolbox as implemented in Matlab (Version 2019b). Extracerebral voxels were excluded from the analysis using a mask of the gray matter component yielded by segmentation of the anatomical image volume into gray matter, white matter, and cerebrospinal fluid. Anatomical and functional images were normalized into standard MNI space. The dimensions of the resampled functional images were 91 × 109 × 91 voxels with voxel sizes 2 × 2 × 2 mm3. Additionally, the mean BOLD-signal was extracted using the CONN-toolbox for each run and subject for each of the 333 cortical ROIs provided in the atlas of cortical area parcellation derived from resting-state correlations by Gordon et al.25 and the 14 ROIs in the Oxford thalamic connectivity atlas26. Consistent with the voxel size of the resampled images, the BOLD signals were extracted from the atlas file. Since the Oxford thalamic atlas26 does not provide the surface area for each ROI, the surface area for each thalamic ROI was calculated using the multi-image analysis GUI Mango (Version 4.1) (http://www.mangoviewer.com).
SPM analysis
The 144 fMRI normalized image volumes for each investigated hand, corresponding to 432 s, were decomposed according to activation conditions: instruction (1), fixation (2), observation of object manipulation (3), and active manipulation (4). The temporal onsets of the conditions in seconds, i.e. instruction, observation and manipulation, are specified in Figure SF2 of Supplementary Material. With the 20 s duration of the BOLD activations, these onsets formed the regressor yielding two 1st level voxel-wise image contrasts of conditions, observation (3), and manipulation (4), respectively, with the reference state, fixation (2), for each hand and subject. Finally, a 2nd level group analysis using a two-tailed t-test for each contrast was carried out. No covariates were used. Voxels showing an alpha-level below 0.05 corrected for multiple comparison using FWE were considered statistical significant. The extension threshold was set at 20 voxels, which corresponded to 1 resel.
Graph theory analysis
First, we defined ROIs for inclusion in the graph matrix based on our previous findings in normal volunteers and post-stroke patients regarding significant network nodes underlying tactile object exploration and its pure motor execution without cognitive load. These ROIs associated with grasping and regrasping formed the basis for definition of the graph matrix and are listed in Table ST1 in the supplementary material. They included core areas of sensori-motor activity, SPL attention areas and areas serving integrative fronto-parietal control. Furthermore, we included occipital and parieto-occipital cortices engaged in visual processing and visuo-motor information transfer as well as midline structures involved in salience recognition, attention, motivation and self-agency44,88; thalamic ROIs were added as detailed below.
The graph analysis of the fMRI acquisitions required brain parcellations to define the ROIs serving as nodes. We used the data-driven brain atlas of Gordon et al.25, which is capable of capturing both RSFC pattern changes and localization of these changes within borders of architectonic divisions89. The ROIs were secondarily assigned to the cytoarchitectonically defined sub-areas of the Julich brain atlas, which verified the underlying structures90,91. The landmarks of the structurally defined ROIs within the probabilistic Julich atlas were transferred into MNI brain space to overlap with the RSFC parcels of the Gordon brain atlas25. Thalamic subareas involved in various thalamic-cortical loops were delineated using the atlas of Behrens et al.26. The centers of gravity of each ROI were calculated. Each ROI, distinguished by its center of gravity, corresponded to one graph node. There resulted graphs with 113 nodes: 49 cortical nodes in the left and 50 cortical nodes in the right hemisphere, and 7 thalamic nodes on each side.
Graphs were computed for each subject, activation condition and hand. For each condition, i.e. observation of manipulation and active object manipulation, we averaged for each ROI the BOLD time series consisting of 36 time points with respect to the six repetitions performed. Computing the temporal correlations among ROIs using the function corrcoef.m implemented in Matlab yielded four graphs of dimension 1132 for each of the 26 subjects. The correlation coefficients between two graph nodes provided the weights of the edges. Finally, the mean of the correlation coefficients with respect to subjects for each condition yielded four weighted, undirected (WU) graphs for each condition. The density, small world propensity, strength and participation coefficient as well as the average degree of all nodes were calculated for each of the four graphs. These graph measures were calculated using the Brain Connectivity Toolbox92 with a connection density threshold of 0.45. This was done in the search of the optimal trade-off between global and local efficiency, which relate to normalized path length, Λ, and normalized cluster coefficient, Γ. Global efficiency and local efficiency were calculated using efficiency.wei.m program of the toolbox mentioned above93,94. Depending on the ranks of the ROI parameter strength, i.e. the nodal metric of the sum of a node’s edge weights, and the ranks of the participation coefficient, i.e. a nodal metric of diversity characterizing a node’s connection across communities34, the 113-ROI graph was pared down to a 60-ROI WU graph. To calculate small world propensity, the graphs were normalized and compared with 1000 weighted undirected null-model networks using the function null_model_and_sign.m of the Brain Connectivity Toolbox. This function preserves the degree-, weight- and strength- distributions in networks with positive and negative weights. Summarizing, the main parameters for further exploration of the graphs were (1) strength and (2) participation coefficient of the 60 ROI WU graphs. This provided the option to address the questions of rich and diversity clubs. Short pathways were determined on the basis of the edge weights.
Clubs and club coefficients
We determined the members of rich and diversity clubs for the completely connected, i.e. unthresholded, 60-ROI graph. This analysis followed the procedure of van den Heuvel and Sporns33 using modified versions of the algorithm, rich_club_wu.m, found in the Brain Connectivity Toolbox, which is described in Eq. 1 of Opsahl et al.95. The modifications included the replacement of node degree by node strength for rich club and by participation coefficient for diversity club in the selection of nodes. In both cases, percentile replaced node degree in computing the course of clubness, which was calculated in increases of 2% from 0 to 100%. For the calculation of participation coefficients, the a-priori modules, i.e. communities, were derived from the atlas of Gordon et al.25. This procedure yielded unnormalized club coefficients, θunnorm. As is conventional33,34,96, they were normalized at each percentile increase by the mean club coefficient of random networks, θrand , that preserved the degree and strength distribution of the network for which the unnormalized club coefficient was calculated97,98. Computing 1000 random graphs for each percentile increase, we determined the mean random club coefficient used to normalize the club coefficient: θnorm = θunnorm/θrand. The course of the normalized club coefficients with respect to percentile was supplemented by the 95% confidence intervals determined by the deviations of the 1000 random graphs about the mean to yield a percentile of maximum significance for each club and condition. Club membership required a coefficient greater than 1.
Shortest paths
The minimal paths were evaluated visually with the NeuroMArVL- tool (https://immersive.erc.monash.edu/neuromarvl/), see also Baker et al.99. We expected the minimum short paths for visual and haptic information transfer from the hOcd4 subarea, involved in object form perception49, to the targets of motor control in the SMA and Broca area BA 44/vPMC7,8. We determined the number of total edges in the graphs for each condition using the network analysis program Gephi according to the formula e = \(\:\frac{n\left(n-1\right)}{2}\)100; furthermore, the minimal number of functional edges based on their weights needed to generate the short pathways101.
Data availability
The datasets generated during the current study and study protocol are available in the Open Science Framework (OSF) repository at https://osf.io/e5q2h/ (DOI 10.17605/OSF.IO/E5Q2H).
References
Bullock, I. M., Ma, R. R. & Dollar, A. M. A hand-centric classification of human and robot dexterous manipulation. IEEE Trans. Haptics. 6, 129–144 (2013).
Krammer, W. et al. Sensing form - finger gaiting as key to tactile object exploration - A data g love analysis of a prototypical daily task. J. Neuroeng. Rehabil 17, (2020).
Kägi, G., Missimer, J. H., Abela, E., Seitz, R. J. & Weder, B. J. Neural networks engaged in tactile object manipulation: patterns of expression among healthy individuals. Behav. Brain Funct. 6, 6 (2010).
Weisstanner, C. et al. The effect of a single dose of escitalopram on sensorimotor networks. Brain Behav. 8, (2018).
Bonini, L., Rotunno, C., Arcuri, E. & Gallese, V. Mirror neurons 30 years later: implications and applications. Trends in Cognitive Sciences vol. 26 767–781 Preprint at (2022). https://doi.org/10.1016/j.tics.2022.06.003
Heyes, C. & Catmur, C. What happened to Mirror neurons? Perspect. Psychol. Sci. 17, 153–168 (2022).
Rizzolatti, G. & Sinigaglia, C. The functional role of the parieto-frontal mirror circuit: I nterpretations and misinterpretations. Nat. Rev. Neurosci. 11, 264–274 (2010).
Binkofski, F. & Buccino, G. The role of ventral premotor cortex in action execution and action understanding. J. Physiol. Paris. 99, 396–405 (2006).
Fogassi, L. & Ferrari, P. F. Mirror systems. Wiley Interdiscip Rev. Cogn. Sci. 2, 22–38 (2011).
Fogassi, L. & Luppino, G. Motor functions of the parietal lobe. Curr. Opin. Neurobiol. 15, 626–631 (2005).
Tkach, D., Reimer, J. & Hatsopoulos, N. G. Congruent activity during action and action observation in motor cortex. J. Neurosci. 27, 13241–13250 (2007).
Evangeliou, M. N., Raos, V., Galletti, C. & Savaki, H. E. Functional imaging of the parietal c ortex during action execution and observation. Cereb. Cortex. 19, 624–639 (2009).
Filimon, F., Rieth, C. A., Sereno, M. I. & Cottrell, G. W. Observed, executed, and imagined action representations can be decoded from ventral and dorsal areas. Cereb. Cortex. 25, 3144–3158 (2015).
Caspers, S., Zilles, K., Laird, A. R. & Eickhoff, S. B. ALE meta-analysis of action observation and imitation in the human brain. Neuroimage 50, 1148–1167 (2010).
Filimon, F. Human cortical control of hand movements: Parietofrontal networks for reaching, grasping, and pointing. Neuroscientist 16, 388–407 (2010).
Buccino, G. et al. Action observation activates premotor and parietal areas in a somatotopic manner: an fMRI study. Eur. J. Neurosci. 13, 400–404 (2001).
Heed, T., Beurze, S. M., Toni, I., Röder, B. & Medendorp, W. P. Functional rather than effector-specific organization of human posterior parietal cortex. J. Neurosci. 31, 3066–3076 (2011).
Hickok, G. & Hauser, M. Mis)understanding mirror neurons. Curr. Biol. 20, R593–R594 (2010).
Abdelgabar, A. R. et al. Action perception recruits the cerebellum and is impaired in patients with spinocerebellar ataxia. Brain 142, 3791–3805 (2019).
Antonioni, A. et al. The cerebellum and the Mirror Neuron System: A matter of inhibition? From neurophysiological evidence to neuromodulatory implications. A narrative review. Neuroscience and Biobehavioral Reviews vol. 164 Preprint at (2024). https://doi.org/10.1016/j.neubiorev.2024.105830
Errante, A., Gerbella, M., Mingolla, G. P. & Fogassi, L. Activation of Cerebellum, basal ganglia and Thalamus during Observation and execution of Mouth, hand, and foot actions. Brain Topogr. 36, 476–499 (2023).
Mathiowetz, V. et al. Grip and pinch strength: normative data for adults. Arch. Phys. Med. Rehabil 66, (1985).
Tong, J., Mao, O. & Goldreich, D. Two-point orientation discrimination versus the traditional two-point test for tactile spatial acuity assessment. Front. Hum. Neurosci. https://doi.org/10.3389/fnhum.2013.00579 (2013).
Abela, E. et al. Early prediction of long-term tactile object recognition performance after sensorimotor stroke. Cortex 115, 264–279 (2019).
Gordon, E. M. et al. Generation and evaluation of a cortical area parcellation from resting-state correlations. Cereb. Cortex. 26, 288–303 (2016).
Behrens, T. E. J. et al. Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging. Nat. Neurosci. 6, 750–757 (2003).
Abela, E. et al. Lesions to primary sensory and posterior parietal cortices impair recovery from hand paresis after stroke. PLoS One 7, (2012).
Ekstrand, E., Rylander, L., Lexell, J. & Brogårdh, C. Perceived ability to perform daily hand activities after stroke and associated factors: a cross-sectional study. BMC Neurol. 16, (2016).
Braun, U. et al. Dynamic reconfiguration of frontal brain networks during executive cognition in humans. Proc. Natl. Acad. Sci. U S A. 112, 11678–11683 (2015).
Davison, E. N. et al. Brain network adaptability across task states. PLoS Comput. Biol. 11, e1004029 (2015).
Hahn, A. et al. Reconfiguration of functional brain networks and metabolic cost converge during task performance. Elife 9, (2020).
Gläscher, J. Visualization of group inference data in functional neuroimaging. Neuroinformatics 7, 73–82 (2009).
van den Heuvel, M. P. & Sporns, O. Rich-club organization of the human connectome. J. Neurosci. 31, 15775–15786 (2011).
Bertolero, M. A., Yeo, B. T. T. & D’Esposito, M. The diverse club. Nat. Commun. 8, (2017).
Faisal Aldo, A. N. D., Stout, D. A. N. & D. A. J. A. N. D. B. B. The manipulative complexity of Lower Paleolithic Stone Toolmaking. PLoS One. 5, 1–11 (2010).
Guiard, Y. The Kinematic Chain as a Model for Human Asymmetrical Bimanual Cooperation. in Cognition and Action in Skilled Behaviour (eds Colley, A. M. & Beech, J. R.) vol. 55 205–228 (North-Holland, (1987).
Paus, T. Primate anterior cingulate cortex: where motor control, drive and cognition interface. Nat. Rev. Neurosci. 2, 417–423 (2001).
Ruan, J. et al. Probabilistic cytoarchitectonic map of area 6mp (SMA, mesial SFG) (9.1) [Data set]. Hum. Brain Project Neuroinformatics Platf. (2019).
Ruan, J. et al. Cytoarchitecture, probability maps, and functions of the human supplementary and pre-supplementary motor areas. Brain Struct. Funct. 223, 4169–4186 (2018).
Ossmy, O. & Mukamel, R. Activity in superior parietal cortex during training by observation predicts asymmetric learning levels across hands. Sci. Rep. 6, (2016).
Heyes, C. Causes and consequences of imitation. Trends Cogn. Sci. 5, 253–261 (2001).
Koenigs, M., Barbey, A. K., Postle, B. R. & Grafman, J. Superior parietal cortex is critical for the manipulation of information in working memory. J. Neurosci. 29, 14980–14986 (2009).
Sakai, K. et al. Transition of Brain activation from Frontal to Parietal areas in Visuomotor sequence learning. J. Neurosci. 18, 1827–1840 (1998).
Brugger, F., Galovic, M., Weder, B. J. & Kägi, G. Supplementary motor complex and disturbed motor control - A retrospective clinical and lesion analysis of patients after anterior cerebral artery stroke. Front. Neurol. 6, (2015).
Cavanna, A. E. & Trimble, M. R. The precuneus: A review of its functional anatomy and behavioural correlates. in Brain vol. 129 564–583Oxford University Press, (2006).
Gordon, E. M. et al. A somato-cognitive action network alternates with effector regions in motor cortex. Nature https://doi.org/10.1038/s41586-023-05964-2 (2023).
Sadaghiani, S. & D’Esposito, M. Functional characterization of the cingulo-opercular network in the maintenance of tonic alertness. Cereb. Cortex. 25, 2763–2773 (2015).
Dosenbach, N. U. F. et al. Distinct brain networks for adaptive and stable task control in humans. (2007).
Gomez, J., Zhen, Z. & Weiner, K. S. Human visual cortex is organized along two genetically opposed hierarchical gradients with unique developmental and evolutionary origins. PLoS Biol. 17, (2019).
Kujovic, M. et al. Cytoarchitectonic mapping of the human dorsal extrastriate cortex. Brain Struct. Funct. 218, 157–172 (2013).
Richter, M. et al. Cytoarchitectonic segregation of human posterior intraparietal and adjacent parieto-occipital sulcus and its relation to visuomotor and cognitive functions. Cereb. Cortex. 29, 1305–1327 (2019).
Alexander, G. E., DeLong, M. R. & Strick, P. L. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu. Rev. Neurosci. VOL. 9, 357–381 (1986).
Spaak, E., Watanabe, K., Funahashi, S. & Stokes, M. G. Stable and dynamic coding for working memory in primate prefrontal cortex. J. Neurosci. 37, 6503–6516 (2017).
Niu, M. & Palomero-Gallagher, N. Architecture and connectivity of the human angular gyrus and of its homolog region in the macaque brain. Brain Structure and Function vol. 228 47–61 Preprint at (2023). https://doi.org/10.1007/s00429-022-02509-7
Levy, R. & Goldman-Rakic, P. S. Segregation of working memory functions within the dorsolateral prefrontal cortex. Experimental Brain Research vol. 133 23–32 Preprint at (2000). https://doi.org/10.1007/s002210000397
Young, J. S., Lee, A. T. & Chang, E. F. SURGICAL MANAGEMENT OF ELOQUENT AREA TUMORS a review of cortical and subcortical stimulation mapping for Language. Neurosurgery 89, 331–342 (2021).
Peng, G. S., Tan, S. Y., Wu, J. & Holme, P. Trade-offs between robustness and small-world effect in complex networks. Sci. Rep. 6, (2016).
Uddin, L. Q. et al. Dissociable connectivity within human angular gyrus and intraparietal sulcus: evidence from functional and structural connectivity. Cereb. Cortex. 20, 2636–2646 (2010).
Jäncke, L., Kleinschmidt, A., Mirzazade, S., Shah, N. J. & Freund, H. J. The role of the inferior parietal cortex in linking the tactile perception and manual construction of object shapes. Cereb. Cortex. 11, 114–121 (2001).
Yang, J. et al. Functional heterogeneity in the left lateral posterior parietal cortex during visual and haptic crossmodal dot-surface matching. Brain Behav. 11, (2021).
Möhring, N., Shen, C. & Neuhaus, A. H. Spatiotemporal dynamics of early cortical gesture processing. Neuroimage 99, 42–49 (2014).
Kawashima, R. et al. Direction of cross-modal information transfer affects human brain activation: a PET study. Eur. J. Neurosci. 16, 137–144 (2002).
Darby, R. R., Joutsa, J., Burke, M. J. & Fox, M. D. Lesion network localization of free will. Proc. Natl. Acad. Sci. U S A. 115, 10792–10797 (2018).
Fattori, P. et al. Hand orientation during reach-to-grasp movements modulates neuronal activity in the medial posterior parietal area V6A. J. Neurosci. 29, 1928–1936 (2009).
Galletti, C. et al. The cortical connections of area V6: an occipito-parietal network processing visual information. Eur. J. Neurosci. 13, 1572–1588 (2001).
Pitzalis, S. et al. Wide-field retinotopy defines human cortical visual area V6. J. Neurosci. 26, 7962–7973 (2006).
Luppino, G., Ben Hamed, S., Gamberini, M., Matelli, M. & Galletti, C. Occipital (V6) and parietal (V6A) areas in the anterior wall of the parieto-occipital sulcus of the macaque: a cytoarchitectonic study. Eur. J. Neurosci. 21, 3056–3076 (2005).
Galletti, C. & Fattori, P. The dorsal visual stream revisited: stable circuits or dynamic pathways? Cortex 98, 203–217 (2018).
Marangon, M., Kubiak, A. & Króliczak, G. Haptically guided grasping. fMRI shows right-hemisphere parietal stimulus encoding, and bilateral dorso-ventral parietal gradients of object- and action-related processing during grasp execution. Front. Hum. Neurosci. 9, (2016).
Rizzolatti, G. & Fogassi, L. The mirror mechanism: recent findings and perspectives. Philosophical Trans. Royal Soc. B: Biol. Sci. 369, (2014).
Clos, M., Amunts, K., Laird, A. R., Fox, P. T. & Eickhoff, S. B. Tackling the multifunctional nature of Broca’s region meta-analytically: co-activation-based parcellation of area 44. Neuroimage 83, 174–188 (2013).
Cieslik, E. C. et al. Is there one DLPFC in cognitive action control? Evidence for heterogeneity from co-activation-based parcellation. Cereb. Cortex. 23, 2677–2689 (2013).
Catani, M. & de Schotten, M. Atlas of Human Brain connections (all tracts). in 75–238 (2012). https://doi.org/10.1093/med/9780199541164.003.0073
Janelle, F., Iorio-Morin, C., D’amour, S. & Fortin, D. Superior Longitudinal Fasciculus: A Review of the Anatomical Descriptions With Functional Correlates. Frontiers in Neurology vol. 13 Preprint at (2022). https://doi.org/10.3389/fneur.2022.794618
Schurr, R., Zelman, A. & Mezer, A. A. Subdividing the superior longitudinal fasciculus using local quantitative MRI. Neuroimage 208, (2020).
Erlikhman, G., Caplovitz, G. P., Gurariy, G., Medina, J. & Snow, J. C. Towards a unified perspective of object shape and motion processing in human dorsal cortex. Consciousness and Cognition vol. 64 106–120 Preprint at (2018). https://doi.org/10.1016/j.concog.2018.04.016
Davare, M., Andres, M., Cosnard, G., Thonnard, J. L. & Olivier, E. Dissociating the role of ventral and dorsal premotor cortex in precision grasping. J. Neurosci. 26, 2260–2268 (2006).
Behrmann, M., Geng, J. & Shomstein, S. Parietal cortex and attention. Curr. Opin. Neurobiol. 14, 212–217 (2004).
Thioux, M., Gazzola, V. & Keysers, C. Action understanding: how, what and why. Curr. Biol. 18, R431–R434 (2008).
Oldfield, R. C. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9, 97–113 (1971).
Folstein, M. F., Folstein, S. E. & McHugh, P. R. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J. Psychiatr Res. 12, 189–198 (1975).
World Medical Association declaration of Helsinki. Ethical principles for medical research involving human subjects. JAMA vol. 310 2191–2194 Preprint at (2013). https://doi.org/10.1001/jama.2013.281053
Jebsen, R. H., Taylor, N., Trieschmann, R., Trotter, M. J. & Howard, L. A. An objective and standardized test of hand function. Arch. Phys. Med. Rehabil. 50, 311–319 (1996).
Bohlhalter, S., Fretz, C. & Weder, B. Hierarchical versus parallel processing in tactile object recognition: a behavioural-neuroanatomical study of aperceptive tactile agnosia. Brain 125, 2537–2548 (2002).
Roland, P. E. & Mortensen, E. Somatosensory detection of microgeometry, macrogeometry and kinesthesia in man. Brain Res. Rev. 12, 1–42 (1987).
Hertling, D. & Kessler, R. M. Management of Common Musculoskeletal Disorders: Physical Therapy Principles and Methodsvol. 3 (J.B. Lippincott, 1996).
Waller, L. et al. GraphVar 2.0: a user-friendly toolbox for machine learning on functional connectivity measures. J. Neurosci. Methods. 308, 21–33 (2018).
Holroyd, C. B. & Yeung, N. Motivation of extended behaviors by anterior cingulate cortex. Trends in Cognitive Sciences vol. 16 122–128 Preprint at (2012). https://doi.org/10.1016/j.tics.2011.12.008
Wig, G. S., Laumann, T. O. & Petersen, S. E. An approach for parcellating human cortical areas using resting-state correlations. NeuroImage vol. 93 276–291 Preprint at (2014). https://doi.org/10.1016/j.neuroimage.2013.07.035
Eickhoff, S. B. et al. Assignment of functional activations to probabilistic cytoarchitectonic areas revisited. Neuroimage 36, 511–521 (2007).
Eickhoff, S. B. et al. A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. Neuroimage 25, 1325–1335 (2005).
Rubinov, M. & Sporns, O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage 52, 1059–1069 (2010).
Wang, Y., Ghumare, E., Vandenberghe, R. & Dupont, P. Comparison of different generalizations of clustering coefficient and local efficiency for weighted undirected graphs. Neural Computation vol. 29 313–331 Preprint at (2017). https://doi.org/10.1162/NECO_a_00914
Latora, V. & Marchiori, M. Efficient behavior of small-world networks. Phys. Rev. Lett. 87 (1-), 198701 (2001).
Opsahl, T., Colizza, V., Panzarasa, P. & Ramasco, J. J. Prominence and control: the weighted rich-club effect. (2008). https://doi.org/10.1103/PhysRevLett.101.168702
Alstott, J., Panzarasa, P., Rubinov, M., Bullmore, E. T. & Vértes, P. E. A unifying framework for measuring weighted rich clubs. Sci. Rep. 4, (2014).
Colizza, V., Flammini, A., Serrano, M. A. & Vespignani, A. Detecting rich-club ordering in complex networks. Nat. Phys. 2, 110–115 (2006).
McAuley, J. J., da Fontoura Costa, L. & Caetano, T. S. Rich-club phenomenon across complex network hierarchies. Appl. Phys. Lett. 91, 084103 (2007).
Baker, S. T. E. et al. Developmental changes in brain network hub connectivity in late adolescence. J. Neurosci. 35, 9078–9087 (2015).
Trudeau, R. J. Introduction to Graph Theory (Dover Pub., 1993).
Bastian, M., Heymann, S. & Jacomy, M. Gephi: an Open Source Software for Exploring and Manipulating Networks. Proc. Int. AAAI Conf. Web Social Media. 3, 361–362 (2009).
Acknowledgements
This work was supported by the Swiss National Science Foundation (SNF Grant number 160107, www.snf.ch), the Swiss Heart Foundation (https://www.swissheart.ch) and the Clinical Trials Unit at the Cantonal Hospital of St.Gallen, Switzerland. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Author information
Authors and Affiliations
Contributions
W.K.: Methodology; Data curation; Formal analysis; Software; Writing of Methodology and Results; original draft review &editingJ.H. M.: Conceptualization; Methodology; Formal analysis; Software; Writing original draft, review &editingV.V.: Formal analysis; original draft reviewM.P.W.: Methodology; Project administration; original draft review G.K.: Funding acquisition; Resources; R.W.: Funding acquisition; Resources; original draft reviewB.J. W.: Funding acquisition; Conceptualization; Methodology; Data curation; Formal analysis; Project administration; Writing original draft, review &editing.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Ethical standards
The study objectives of the pre-registration document (https://clinicaltrials.gov/study/NCT02865642) do not apply to this study, which presents a methodology for the evaluation of the intended patient populations and reference results for normal subjects.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Krammer, W., Missimer, J.H., Vallesi, V. et al. Exploring imitation of within hand prehensile object manipulation using fMRI and graph theory analysis. Sci Rep 15, 3641 (2025). https://doi.org/10.1038/s41598-025-86157-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-025-86157-x
