Skip to main content
NCBI home page
Cerebral Cortex (New York, NY) logoLink to Cerebral Cortex (New York, NY)
. 2024 Jun 4;34(6):bhae230. doi: 10.1093/cercor/bhae230

Spatiotemporal dynamics of cortical somatosensory network in typically developing children

Yanlong Song 1,2,3,#, Sadra Shahdadian 4,5,#, Eryn Armstrong 6, Emily Brock 7, Shannon E Conrad 8, Stephanie Acord 9, Yvette R Johnson 10,11, Warren Marks 12, Christos Papadelis 13,14,15,✉
PMCID: PMC11151116  PMID: 38836408

Abstract

Sense of touch is essential for our interactions with external objects and fine control of hand actions. Despite extensive research on human somatosensory processing, it is still elusive how involved brain regions interact as a dynamic network in processing tactile information. Few studies probed temporal dynamics of somatosensory information flow and reported inconsistent results. Here, we examined cortical somatosensory processing through magnetic source imaging and cortico–cortical coupling dynamics. We recorded magnetoencephalography signals from typically developing children during unilateral pneumatic stimulation. Neural activities underlying somatosensory evoked fields were mapped with dynamic statistical parametric mapping, assessed with spatiotemporal activation analysis, and modeled by Granger causality. Unilateral pneumatic stimulation evoked prominent and consistent activations in the contralateral primary and secondary somatosensory areas but weaker and less consistent activations in the ipsilateral primary and secondary somatosensory areas. Activations in the contralateral primary motor cortex and supramarginal gyrus were also consistently observed. Spatiotemporal activation and Granger causality analysis revealed initial serial information flow from contralateral primary to supramarginal gyrus, contralateral primary motor cortex, and contralateral secondary and later dynamic and parallel information flows between the consistently activated contralateral cortical areas. Our study reveals the spatiotemporal dynamics of cortical somatosensory processing in the normal developing brain.

Keywords: Granger causality, information flow, magnetoencephalography, somatosensory processing

Introduction

Sense of touch is essential for dexterous manipulation of objects and fine control of hand actions. When a tactile stimulus contacts the skin, skin deformation activates mechanoreceptors that encode information on the stimulus properties, such as texture. The resultant neural signals are transmitted through peripheral nerves to the spinal cord, relayed by the thalamus, and finally processed by a distributed cortical somatosensory system (see review, Abraira  and  Ginty  2013). Several previous studies have used electrical stimulation of peripheral nerves with the aim to reveal the underlying mechanisms of the cortical somatosensory system (Forss  et al.  1994; Lin  et al.  2003; Inui  et al.  2004; Nihashi  et al.  2005; Papadelis  et al.  2011; Hautasaari  et al.  2019). Yet, electrical stimulation experimental paradigms are not suitable to assess the cortical processing of touch, since they involve unnatural stimuli, which rarely occur in real life. Moreover, they activate simultaneously a large number of superficial and deep receptors and nerve fibers (Forss  et al.  1994) and activate ascending pathways and cortical areas differently (Forss  et al.  1994; Abraira  and  Ginty  2013; Hautasaari  et al.  2019), hence, prevent us from precisely delineating spatiotemporal dynamics of cortical processing of touch. Tactile stimulation mimicking natural touch, though not widely employed, is more suitable for testing cortical processing of touch.

Past human brain research employing different methods [e.g. magnetoencephalography (MEG), electroencephalography (EEG), intracranial EEG, and functional magnetic resonance imaging (fMRI)] have revealed that different peripheral stimulation with electric, tactile, or nociceptive stimuli often activates the contralateral primary (cS1) and secondary (cS2) somatosensory cortices, as well as the ipsilateral secondary (iS2) somatosensory cortex (see review, Iwamura  1998; Hari  and  Forss  1999; Lamp  et al.  2019). Activations of other cortical areas from tactile or electric stimulation, including the ipsilateral primary (iS1) somatosensory cortex (Allison et al. 1989; Korvenoja  et al.  1999; Nihashi  et al.  2005; Hlushchuk  and  Hari  2006), posterior parietal cortex (Chen  et al.  2008; Naeije  et al.  2016), and supplementary motor area (SMA; Korvenoja  et al.  1999; Naeije  et al.  2016), have not been consistently reported. Direct comparisons of cortical activations evoked by electric and tactile stimulation in same cohorts of participants revealed differences on latency and amplitude of early cortical response at cS1 and the activation of posterior parietal cortex (Forss  et al.  1994) and late cortical activations at bilateral S1 areas (Hautasaari  et al.  2019). Although research over decades has improved our understanding of the cortical areas involved in human somatosensory processing, it is still elusive how these cortical areas are integrated into a dynamic brain network. For example, it is still unclear how information flows among the involved brain areas and whether it takes place in a serial or parallel fashion. Furthermore, most previous research focused on early cortical somatosensory responses, which may mature as early as 5 years of age (see review, Nevalainen  et al.  2014). Late cortical somatosensory responses are not well studied and may still be developing in normal developing brains.

The human somatosensory system has been well established as an anatomical hierarchy, with S2 serving as a higher-order area relative to S1 (Iwamura  1998). Yet, there is still a debate on how thalamic somatosensory inputs enter the cortex through cS1 and cS2. The serial view claims that thalamic inputs first reach cS1 and then are relayed to cS2 (Iwamura  1998), while the parallel view asserts that thalamic inputs project simultaneously to both cS1 and cS2 (Karhu  and  Tesche  1999). Several functional neuroimaging studies have tried to resolve this debate but reached discordant conclusions (Inui  et al.  2004; Liang  et al.  2011; Papadelis  et al.  2011; Kalberlah  et al.  2013; Chung  et al.  2014; Khoshnejad  et al.  2014; Hu et al. 2015; Klingner  et al.  2016; Song  et al.  2021). Methodological differences in imaging modality (e.g. fMRI, MEG, and EEG), peripheral stimulation (e.g. electric, tactile, and laser), data analysis approaches [e.g. dynamic causal modeling, Granger causality (GC), and source signal latency], and cortical areas included in the analysis, likely contribute to these discordant conclusions. Thus, there is an unmet need for further research aiming to better understand the brain network underlying human cortical somatosensory processing.

In this study, we aim to elucidate the spatiotemporal activations of cortical areas evoked by unilateral pneumatic stimulation and assess the temporal dynamics of information flow among these areas in the normal developing brain. We hypothesize that pneumatic stimulation of upper extremities evokes activations in bilateral somatosensory areas, and that cortical somatosensory processing of unilateral stimulation primarily proceeds in a serial hierarchal manner from cS1 to cS2 and to other brain areas. To test our hypothesis, we recorded somatosensory evoked fields (SEFs) with MEG from typically developing (TD) children during unilateral pneumatic stimulation of their upper extremities. MEG provides direct measures of neural activities with superior temporal resolution and sufficient spatial resolution (Lopes  da  Silva  2013), which are critical to the fine analysis of spatiotemporal cortical activations and information flow temporal dynamics in response to peripheral touch. We then calculated dynamic statistical parametric maps after pneumatic stimuli onset, assessed cortical activations with spatiotemporal activation analysis, and estimated the temporal dynamics of cortical information flow using multivariate GC.

Material and methods

Participants

Thirty-five TD children participated in this study. All participants had no known neurological impairments or disorders and demonstrated normal cognitive function. Participants’ finger touch sensitivity was assessed with Von Frey monofilaments. The Institutional Review Board at Cook Children’s Health Care System approved this study (IRB number 2019-068; PI: C. Papadelis). Before any data collection, participants gave assent and their legal guardians provided informed written consent.

MRI acquisition

T1-weighted structural MRI scans were acquired from a Siemens Skyra 3T MR scanner and a 10-channel head coil by using 3D magnetization-prepared rapid-acquisition gradient-echo sequence with parameters: echo time = 3.22 ms, inversion time = 450 ms, repetition time = 8.21 ms, flip angle = 15 degrees, field of view = 24 cm, matrix size = 256 × 256, 184 slices, acquired resolution = 1.0 × 1.0 × 1.0 mm3.

MEG acquisition

MEG signals were recorded with participants in a sitting position inside a 1-layer magnetically shielded room (MSR; Imedco, Hägendorf, Switzerland). MEG data were collected using a whole-head Neuromag® Triux 306-sensor system (MEGIN, Finland). We initially placed 5 head position indicator (HPI) coils on the child’s head at the bilateral auricular points and the forehead along the hairline. We then digitized the position of each HPI coil, head cardinal landmarks (i.e. nasion, left and right pre-auricular points), and ~500 extra scalp points using a digitizer (Fastrak Polhemus, United States). This process allows for the coregistration of the MEG sensors location with the participant’s structural MRI. Electrocardiography (ECG) was recorded with a pair of electrodes placed under the left clavicle and later used for the removal of MEG signal artifacts from heartbeats. Electrooculography (EOG) was recorded with two electrodes placed above and below one eye and later used for the removal of MEG signal artifacts from eye blinks. After initial preparation, we accompanied the participant into the MSR where the recordings were performed. We attached pneumatic stimulation clips that deliver single pulse compressed air to the tips of left and the right middle fingers. The pressure of the compressed air pulses was above participants’ perception threshold. To collect SEFs, we asynchronously stimulated the middle fingers of both hands with an inter-stimulus interval of 1.5 ± 0.5 s following a pseudorandom order. Each digit received 400 stimuli. During the recording, participants placed their arms and hands in a comfortable and still position on a flat tray. To minimize participants’ movements that severely contaminate MEG signals, we let participants watch cartoon movies projected on a screen in front of them. It has been shown that this experimental approach does not modulate properties of the early and late somatosensory-evoked potentials (Espenhahn et al. 2020). MEG signals were recorded with a sampling rate of 1 KHz.

MEG data analysis

Raw MEG recordings were filtered with MaxFilter (MEGIN, Finland) using the temporal extension of signal-space separation to reduce environmental noise and compensate for head movements. Preprocessing and source localization of MEG data were performed with Brainstorm (version: 2 November 2020; Tadel  et al.  2011). After removal of the direct current offset, the data were bandpass filtered between 1 and 100 Hz and notch filtered at 60 Hz and its harmonics to remove power line noise. Bad segments and bad channels were manually removed. Artifacts caused by heartbeats and blinks were automatically detected from ECG and EOG recordings, then removed with the application of signal-space projection (Tesche  et al.  1995) through Brainstorm and independent component analysis (ICA) (Bell  and  Sejnowski  1995) through the Infomax algorithm implemented by calling EEGLAB (Delorme  and  Makeig  2004) function RunICA in Brainstorm (Tadel  et al.  2011).

To assess sensor SEFs from stimulation of a digit, the continuous sensor data were initially epoched into event-locked trials of 700 ms (200 ms prestimulus and 500 ms poststimulus onset). Grand averages across all runs were computed for the sensor SEFs by averaging the trials of a digit in each individual run and then averaging individual run mean for each participant (Fig. 1). To localize the neuronal sources underlying sensor SEFs, we employed dynamic statistical parametric mapping (dSPM; Dale  et al.  2000), which is a noise-normalized solution based on the minimum norm estimate (Hämäläinen  and  Ilmoniemi  1994). Individual canonical surfaces consisting of the cortex, inner skull, outer skull, and scalp were initially generated from the T1-weighted brain MRI scans. For the participants who did not complete the MRI acquisition process, we used the MNI152 T1-weighted structural template (Grabner  et al.  2006). Then, MEG signals were coregistered to the canonical surfaces by projecting digitized head points to the canonical scalp surface. The forward model was estimated by using a boundary element model with the OpenMEEG toolbox (Gramfort  et al.  2010). Noise covariance matrix was computed from the prestimulus activity. The dSPM source activity was calculated for the cortex surface (~15,000 vertices) using Brainstorm default parameters with both gradiometers and magnetometers included and dipole orientations constrained to the cortex and then averaged across runs per stimulation site for each participant.

Fig. 1.

Fig. 1

Open in a new tab

SEFs and peak/trough as detected from upper and lower envelopes. The upper panel is the sensor SEFs and the lower panel is the source SEFs from a participant (Sub18, 10 year-old, male). Black curves represent SEFs and green curves the envelopes. Peaks and troughs highlighted with red diamonds were detected in the envelopes based on the thresholds marked with dashed lines. L: left; R: right.

Cortical spatiotemporal activation analysis

We performed individual- and group-level spatiotemporal activation analysis. We reconstructed source SEFs for each participant to determine individual’s spatiotemporal activation in response to pneumatic stimulation. Upper and lower envelopes of source SEFs from every cortical vertex were initially determined with the Matlab function envelop. Then, maximum of upper envelopes and minimum of lower envelopes across all cortical vertices were determined and used to detect peaks and troughs of source SEFs. The Matlab function findpeaks was applied in the detection of peaks in individual source SEFs with constraint of minimal peak height of eight times the prestimulus field amplitude (Fig. 1). Automatically-identified pairs of peak and trough were further visually evaluated. The location of the maximal activation time-locked to a pair of peak and trough in individual source SEFs was labeled based on the Human Connectome Project Multi-Modal Parcellation version 1.0 (HCP-MMP1) atlas (Glasser  et al.  2016). Figure 2 displays the timing and the maximal activation location at each pair of peak and trough in individual source SEFs for all participants for the left and right middle finger stimulation. To determine group-level spatiotemporal activation from the pneumatic stimulation of a middle finger, individual’s average of source SEFs were further averaged across participants. Components of group-level source SEFs were identified based on its morphology (Fig. 3). Group-level source activation analysis was performed in a common MNI152 template cortex source space (Grabner  et al.  2006) by projecting each individual’s cortical activations to the template. To determine cortical activations corresponding to a component at the group level, we first computed the mean of the dSPM values in a component time window for each source vertex for each participant. Correspondingly, the same computations were performed for the prestimulus phase for each participant. Then, the cortical activations corresponding to a component at the group level were further determined with paired-sample permutation t-test and partial conjunction analysis (Heller  et al.  2007), which quantifies inter-subject consistency of cortical activations within a source vertex (Fig. 4). Significant group-level activations from the t-test may be biased by a small number of participants if inter-subject variation is large. Considering the age range (5 to 18 years) and head size difference in our sample, we applied partial conjunction analysis as a complementary method. For t-test, individual mean dSPM values in a component window were tested against their mean dSPM values in the prestimulus window for each source vertex (15,000 total); activations with P < 0.001 were considered as significant. For partial conjunction analysis, we applied the procedure proposed by Heller  et al.  (2007) and tested the least number of participants who showed significant activations at P < 0.001 threshold at each vertex in a component window by pooling p values at a vertex from all independent participants included in the group analysis with the Fisher method (equation 7 in Heller  et al.  2007).

Fig. 2.

Fig. 2

Open in a new tab

Peak latency and source localization for all participants from the left and the right middle finger stimulation. For each participant, latencies and the maximal activation source localizations time locked to peaks of source somatosensory-evoked fields were displayed. Cortical area labels are from the HCP-MMP1 atlas (Glasser  et al.  2016). AIP: anterior intraparietal; AVI: anterior ventral insular; FOP: frontal operculum; LIPd: lateral intraparietal dorsal; OFC: orbitofrontal cortex; OP: operculum; PGi: inferior parietal area PG; PCV: precuneus visual area; PF: inferior parietal area PF; POS: parieto-occipital sulcus; Pres: presubiculum; pros: prostriate area; STS: superior temporal sulcus; TE: temporal area TE; TG: temporal area TG; TPOJ: temporoparietoocciptal junction. L: left; R: right; PK: peak.

Fig. 3.

Fig. 3

Open in a new tab

Group-level components in source SEFs. Components in group average source SEFs (black curves) were determined by SEFs morphology. Color patches highlight the components of SEFs. Green curves represent the envelopes of the group average source SEFs. Peaks and troughs are highlighted with diamonds.

Fig. 4.

Fig. 4

Open in a new tab

Group level spatiotemporal activations time-locked to components in source SEFs. The upper panel displays significant cortical activations locked to individual components as determined by group level paired t-tests relative to prestimulus phase (thresholded at P < 0.001). The lower panel shows the least number of participants at cortical areas where individual participants show significant activations (thresholded at P < 0.001) in a component as determined by partial conjunction analysis.

Analysis of developmental changes

We estimated Pearson correlations between individual participant’s age and individual source SEFs peak latency and correlations between age and the maximal source amplitude at one of the first three peaks that were observed in most of the participants in individual source SEFs (see Results section). We further separated participants into three size-balanced subgroups, the 5 ~ 8  years subgroup (nine participants), the 10 ~ 11 years subgroup (ten participants), and the 12 ~ 18  years subgroup (ten participants) in reference to theoretical developmental stages (Thompson  2021). For each subgroup, we performed cortical spatiotemporal activation analysis with the same methods described above at the group level analysis, except that the significance threshold was reduced to P < 0.01 considering the small subgroup sample size. Furthermore, we performed one-way analysis of variance (ANOVA) to detect differences among subgroups on cortical spatiotemporal activations in the components of the group-level source SEFs.

GC estimation

We estimated broadband multivariate GC between commonly activated cortical areas from stimulation of left and right middle fingers by using the MVGC multivariate GC toolbox (Barnett  and  Seth  2014); this tool eclipses the traditional bivariate GC that is prone to spurious effects through joint dependencies of two measured variables to another set of unobserved variables (Barnett  and  Seth  2014). Moreover, the toolbox’s default linear state space model has greater statistical power and smaller bias than linear autoregressive model (Barnett  and  Seth  2015). To estimate the GC among common active brain regions, we firstly defined regions of interest (ROIs) based on the results from the group-level spatiotemporal activation analysis (Fig. 5) and subgroup-levelspatiotemporal activation analysis (Fig. 7). Specifically, we combined results from the t-test and partial conjunction analysis in each component window and considered only activations in those source vertices with P < 0.001 from both the t-test and partial conjunction analysis. We then overlaid the combined significant activations from all group-level components from stimulation of a middle finger. In subgroup analysis, due to the small sample size in each subgroup, the significance threshold was relaxed to P < 0.01. This procedure revealed four common contralateral cortical areas that were consistently activated in the whole group and in each of the three subgroups. These four areas were selected as ROIs for the GC analysis. The group-level ROIs were projected back to individual source space. Source waveforms for each ROI were then estimated from individual trials for each participant. The estimated source waveforms for the ROIs were used for GC analysis. To estimate the dynamics of information flow among the ROIs, we chose a sliding (by 20 ms) window length of 60 ms considering the trade-off between signal stationarity and model estimation accuracy. Signal stationarity was automatically tested by the MVGC toolbox before model estimation. Model order was selected based on Bayesian information criterion by the toolbox. Across participants, signal stationarity in each sliding window passed the toolbox test and model order was in the range of 15 to 20.

Fig. 5.

Fig. 5

Open in a new tab

Common cortical areas with significant activations as revealed by combining group t-tests and partial conjunction analysis across all components. The upper panel displays cortical areas with significant activations (P < 0.001 from both group t-tests and partial conjunction analysis, same below) in the contralateral hemisphere from stimulation of the left or the right middle finger. Common areas from both the left and right finger stimulation are marked with lines and labels. The lower panel shows cortical areas with significant activations in the ipsilateral hemisphere from stimulation of the left or the right middle finger. Color scale represents the least number of participants with significant activations (P < 0.001) at each vertex. The anatomical boundaries of commonly activated cortical areas are marked with BA parcellations based on the PALS-B12 Brodmann atlas (Van  Essen  2005).

Fig. 7.

Fig. 7

Open in a new tab

Cortical areas with significant activations as revealed by combining subgroup t-tests and partial conjunction analysis across all components in three subgroups. The significance threshold is P < 0.01.

Statistical analysis

We used paired permutation t-tests (5,000 permutations) that do not have any data distribution assumptions in the statistical tests. Individual mean cortical activations in a component window were tested against prestimulus mean activations. Estimations of GC were tested against prestimulus estimations. Source activation permutation t-tests were performed in Brainstorm (Tadel  et al.  2011). GC permutation tests were performed with the statistical software R (R  Core  Team  2018; version 4.1.3) by using the package RVAideMemoire (Hervé 2022). Significance level was set at 0.001 (uncorrected for multiple comparisons) to control the rate of type I errors from multiple comparisons. To further limit the rate of false discoveries on GC estimations, for the analysis in a sliding time window, we considered a GC as significant only when the estimations are significant in both the left and right middle finger stimulation.

Results

Three participants did not complete the MEG task. Five participants did not complete the MRI session. Touch sensitivity was assessed with Von Frey monofilaments in 18 participants (0.10 ± 0.085 g for the left middle finger and 0.095 ± 0.089 g for the right middle finger). Review of raw data lead to exclusion of MEG data from three children due to the presence of excessive myogenic artifacts. In the final analysis, MEG recordings from 29 TD children (10.6 ± 3.6 years, range 5–18 years; 14 females) were included.

Individual cortical evoked responses and source localizations

Figure 1 displays SEFs on sensor level and source level and corresponding peaks/troughs as detected from the upper and lower envelopes in a participant (sub 18, male, 10 years), who shows five peaks/troughs from the left middle finger stimulation and four peaks/troughs from the right middle finger stimulation, with the first peak/trough occurring at 60 ms in both left and right middle finger stimulation. The number and latency of peaks/troughs in source SEFs varied between the left and right middle finger stimulation and across participants, as shown in Fig. 2. The first two peaks/troughs were seen in all participants for both the left as well as the right middle finger stimulation. The latency of the first peak/trough was 63 ± 7 ms (range: 51 ~ 77 ms) for the left and 64 ± 6 ms (range: 52 ~ 78 ms) for the right middle finger stimulation, respectively. Source localizations of the maximal activations at the first peak/trough revealed consistent activations in the contralateral Brodmann area (BA) 3, a sub-area of S1 (20 out of 29 for the left and 16 out of 29 for the right finger stimulation). The latency of the second peak/trough was 104 ± 19 ms (range: 72 ~ 151 ms) for the left and 102 ± 20 ms (range 71 ~ 153 ms) for the right finger stimulation. Source localizations of the second peak/trough showed activations mainly at the contralateral S1 (14 out of 29 for the left and 15 out of 29 for the right finger stimulation) and S2 areas (6 out of 29 for the left and 7 out of 29 for the right finger stimulation). The latency of the third peak/trough was 159 ± 40 ms (range: 88 ~ 267 ms) for the left and 157 ± 42 ms (range 96 ~ 286 ms) for the right finger stimulation. Source localizations of the third peak/trough showed consistent activations at the contralateral S2 area (8 out of 29 for the left and 11 out of 29 for the right finger stimulation). The latency of the fourth peak/trough was 224 ± 54 ms (range: 115 ~ 313 ms) for the left and 229 ± 71 ms (range: 116 ~ 418 ms) for the right finger stimulation. Source localizations of the fourth peak/trough varied from contralateral S1, S2, and posterior parietal areas to ipsilateral S1 and S2 areas. The later peaks/troughs were not commonly displayed across all the participants. The supplementary file provides the maximal activation source localizations at individual source SEFs peaks/troughs for each participant.

Group-level cortical activations from unilateral pneumatic stimulation

Group-level analysis on source SEFs morphology revealed four distinct components for the left and the right middle finger stimulation (Fig. 3). The timing of each of the four components was comparable between the left and the right finger stimulation, with the first component occurring within 48 ~ 90 ms and 48 ~ 86 ms separately, the second component within 90 ~ 135 ms and 86 ~ 120 ms, the third component within 135 ~ 260 ms and 120 ~ 255 ms, and the fourth component within 260 ~ 385 ms and 255 ~ 385 ms. The morphology of the four components shows similarity between the left and the right finger stimulation, except the third component.

Group-level source activation analysis with paired t-tests and partial conjunction analysis (Fig. 4) revealed prominent and consistent activations in the contralateral hemisphere, mostly in the cS1 and cS2 in the first and second components for both the left and right finger stimulation. In contrast, the ipsilateral hemisphere presented weaker and less consistent activations, with significant activations observed within iS1 and iS2 in the second and third components from the left and right finger stimulation. In the fourth component, consistent significant activations returned to the contralateral hemisphere at the cS1 and the supramarginal gyrus (cSG).

Figure 5 displays common vertices with significant activations (P < 0.001) from both paired t-tests and partial conjunction analysis. The results were overlaid across all the components from the left or right finger stimulation. In the contralateral hemisphere, there were common cortical areas showing significant activations within 400 ms after the stimulus onset at the group level from both the left and right finger stimulation; these areas were the contralateral primary motor cortex (cM1; BA 4), cS1 (BA 3/1/2), cSG (BA 40), parietal–temporal–occipital junction (PTO; BA 39), cS2 (BA 43), SMA (BA 6), and frontal operculum (BA 44). Differently, in the ipsilateral hemisphere, there were common cortical areas with significant activations at the group level at the iS1 (BA 3/1/2), iS2 (BA 43), and posterior transverse temporal area (BA 42). Moreover, the consistency of significant activations at the iS1 and iS2 areas were less than that at the cS1 and cS2 areas.

Developmental changes

Figure 6 shows bivariate correlations between participant’s age and the maximal source activation amplitude at an individual source SEFs peak time in the left and right finger stimulation. At the first peak, the correlations were significant in both the left (r = 0.42, P = 0.02) and right (r = 0.36, P = 0.05) finger stimulation. At the second peak, the correlations were not significant in both the left and right finger stimulation (P > 0.19). At the third peak, the correlation was significant in the right (r = 0.53, P = 0.003) finger stimulation while the correlation was not significant in the left (P = 0.45) finger stimulation. Besides, correlations between participant’s age and latency at an individual source SEFs peak were not significant across the first, second, and third peaks in both the left and right finger stimulation (P > 0.19). Figure 7 presents cortical source activations across the four components in the group-level source SEFs in three developmental subgroups. Across the three subgroups, consistent significant activations were present in the cS1, cM1, cS2, and cSG areas in the contralateral hemisphere while significant but inconsistent activations existed mainly at iS1 and iS2 in the ipsilateral hemisphere. Subgroup cortical activation difference tested through one-way ANOVA revealed no significant differences in both hemispheres among the three subgroups at the thresholds of P < 0.001 and P < 0.01.

Fig. 6.

Fig. 6

Open in a new tab

Scatter plots of participant’s age and the maximal source activation amplitude at individual source SEFs peaks. The results for the first three peaks that were observed in most participants are provided.

Temporal dynamics of contralateral cortical somatosensory information flow

The four common contralateral cortical areas, cM1 (BA 4), cS1 (BA 3/1/2), cSG (BA 40), and cS2 (BA 43) with consistent activations from both the left and right middle finger stimulation, in the whole group and across the three developmental subgroups, were selected as ROIs for the multivariate GC analysis. Figure 8 displays diagrams of common significant GC estimations (P < 0.001), compared to the prestimulus estimations, between the ROIs from both the left and right finger stimulation in sliding time windows. The converging results showed that significant information flow as reflected by GC estimations firstly occurred from cS1 to cSG after the stimulation onset to 60 ms. After this time window, cS1 exerted significant information flow to cM1 (from 20 to 180 ms), cSG (from 20 to 280 ms), and cS2 (from 40 to 300 ms); cS1 received significant information flow from cSG (from 60 to 220 ms), cS2 (from 80 to 160 ms and from 220 to 320 ms), and cM1 (from 100 to 160 ms and 240 to 320 ms). Contralateral primary motor cortex and cS2 showed significant mutual information flow (from 20 to 300 ms). Supramarginal gyrus, besides significant information flow from and to cS1, had significant information flow to cS2 (from 20 to 320 ms) and cM1 (from 40 ms to 100 ms); it also received significant information flow from cS2 (from 40 to 100 ms and from 180 to 300 ms) and cM1 (from 60 to 180 ms).

Fig. 8.

Fig. 8

Open in a new tab

Information flow temporal dynamics among contralateral somatosensory areas. A) The 4 ROIs in Granger causality analysis for both the left and right middle finger stimulation are marked in the corresponding contralateral hemisphere; source waveforms of the four ROIs are presented in the color-coded curves. B) Converging significant granger causations in each sliding time window (P < 0.001 from t-tests relative to the prestimulus phase) between ROIs in both the left and right middle finger stimulation are marked by direction arrows. c: Contralateral; M1: primary motor cortex; S1: primary somatosensory cortex; S2: secondary somatosensory cortex; SG: supramarginal gyrus; L: left; R: right.

Discussion

In the present study, we assessed the spatiotemporal cortical activations evoked by unilateral pneumatic stimulation of a digit (middle finger) with MEG and the temporal dynamics of cortical somatosensory information flow through GC analysis. Our results show prominent and consistent activations in the contralateral hemisphere including cS1 (BA 3/1/2), cS2 (BA 43), cM1 (BA 4), and cSG (BA 40) in both the left and right middle finger stimulation at the group level and across the three developmental subgroups. Differently, in the ipsilateral hemisphere, weaker and less consistent activations were observed at the iS1 and iS2 at the group level. At the individual level, in contrast to the consistent activation time-locked to the first peak/trough in source SEFs, activations time-locked to the second and later peaks/troughs showed variations across participants. Cortical somatosensory information flowed initially in a serial manner from cS1 to cSG, cM1, and cS2 and later parallelly and dynamically between the consistently activated contralateral cortical areas.

Contralateral cortical activations by unilateral pneumatic stimulation

At the individual level, the latency and source localization of the first peak/trough of SEFs show remarkable consistency across participants and between the left and right finger stimulation, with a mean latency of 60 ms and source localization at cS1 primarily its subarea BA3, which are consistent with previous findings (see review, Iwamura  1998; Hari  and  Forss  1999; Nevalainen  et al.  2014; Lamp  et al.  2019). In contrast to the consistencies on the latency and localization of the first peak/trough, numbers, latencies, and localizations of late peaks/troughs of SEFs show vast inter-subject variations (Fig. 2). The exact causes for these inter-subject variations remain unknown, but anatomical differences (e.g. variation in the size and orientation of cortical generators; Woods  and  Courchesne  1987), transcriptomic and macroscopic variability shaped by genetic factors (Li  et al.  2021), prestimulus spontaneous fluctuations of neural activity such as alpha oscillations (Iscan  et al.  2016; Iemi  et al.  2019), and general and task-specific psychological factors like variability in attention (Woods  and  Courchesne  1987) might be factors that result in varying endogenous cortical responses across repeated stimuli presentation. Despite so, at the individual level, we observed activations in cS2 or adjacent areas time-locked to some later peaks/troughs in most participants.

At the group level, we demonstrated, from the stimulus onset up to 400 ms in both the left and right middle finger stimulation, prominent and consistent cortical activations in cS1 and cS2 (Fig. 5 and 7). These findings are consistent with previous studies showing that somatosensory stimulation with electric, pneumatic, or vibrotactile stimuli evokes robust activations in cS1 and cS2 (Inui  et al.  2004; Liang  et al.  2011; Papadelis  et al.  2011, 2012, 2014; 2017; Klingner  et al.  2016). Besides, we also showed significant cortical activations in the contralateral hemisphere at cM1 (BA 4), PTO (BA 39), cSG (BA 40), and SMA (BA 6), which were not consistently reported in previous studies. M1 has dense reciprocal connections with S1 and receives sensory inputs from S1 (Gómez  et al.  2021). PTO (BA 39) and SG (BA 40) are within the parietal–temporal–occipital associative cortical area and assemble somatosensory, auditory, visual, and motor information from multiple cortical and subcortical areas (Whitlock  2017). SMA also has interconnections with the area BA 3b of S1 in mammals (Krubitzer  and  Kaas  1990). Our results of significant activations at cM1 (BA4), PTO (BA39), cSG (BA 40), and SMA (BA 6) are aligned with these findings (Whitlock  2017; Gómez  et al.  2021).

Ipsilateral cortical activations by unilateral pneumatic stimulation

In contrast to the contralateral hemisphere, the ipsilateral hemisphere shows weaker and less consistent activations at the group level. Activations at iS2 from unilateral stimulation was reported in previous studies (see review, Lamp  et al.  2019); however, activation at iS1 was not consistently reported but was indeed documented in previous human studies (Allison  et al.  1991; Tan et al. 2004; Hlushchuk  and  Hari  2006). Here, we observed significant activations at iS1 and iS2 from the left and right middle finger stimulation at the group level.

How iS1 and iS2 are activated and involved in cortical somatosensory processing of unilateral stimulation still needs further investigation. Neurons of S2 have receptive fields including bilateral inputs (Iwamura  2000; Taoka  et al.  2016). Ipsilateral secondary could be activated via different cortical projections including callosal connections from cS2, callosal projections from cS1, or reciprocal connections from iS1 (Jones  and  Powell  1969). Neurons of S1 at subarea BA 3 representing distal body parts, such as hand only, have unilateral receptive fields, but neurons of S1 at subareas BA 2/1 have bilateral receptive fields (Iwamura  2000). iS1 could be activated via callosal projections between bilateral BA 2/1 areas or from posterior parietal cortex or cS2 (Jones  and  Powell  1969; Iwamura  et al.  1994; Krubitzer  et al.  1998). Activations at iS2 and iS1 are likely transmitted from callosal connections from their contralateral homologous parts as revealed by cortical lesions (Iwamura  2000). These two ipsilateral areas may function to integrate bilateral somatosensory representations (Iwamura  2000; Tamè  et al.  2016). The potential integration of ipsilateral somatosensory inputs in iS1 might be soon after the occurrence of the initial cortical response, as indicated by previous findings showing that when two somatosensory stimuli are separated at a short interval as 20 ~ 25 ms, early cortical response evoked by the second stimulus is suppressed compared to when no stimulus precedes it (Ragert  et al.  2011; Tamè et al. 2015).

Developmental changes

We did not find any correlations between the participant’s age and individual peak latency at the first, second, and third peaks that commonly existed in most individual source SEFs. This may indicate the likely maturation of early somatosensory information conduction from peripheral to the cortex in our cohort of participants whose age range is 5 ~ 18 years. These results are consistent with a conclusion from Nevalainen  et al.  (2014), who reviewed MEG somatosensory studies in children and reported that the latency of the early cortical somatosensory response decreases from infancy to about 3 ~ 5  years and then slightly increases or stabilizes. We did find correlations between the participant’s age and individual maximal cortical activation amplitude at the first peak in both the left and right middle finger stimulation and at the third peak only in the right middle finger stimulation. The maximal cortical activation at the first individual source SEFs peak is the cS1 in most of our participants. These results hint that the development of S1, its local neural circuit excitability as reflected by activation amplitude, is still immature in the age range of 5 ~ 18 years. The refinement of developing neural circuits involves both structural and function maturation and is regulated by a serial of neurochemical factors (Clarke  and  Barres  2013; Park  and  Poo  2013) and dependent on activity and experience (Holtmaat  and  Svoboda  2009; Ganguly  and  Poo  2013).

Converging evidence of GC favors initial serial somatosensory processing

From pneumatic stimulation onset, cS1 had significant Granger causal influences initially to cSG (0 ~ 60 ms), then to cM1 (20 ~ 80 ms) and cS2 (40 ~ 100 ms). This favors the possibility of early serial somatosensory processing, which claims that thalamic somatosensory inputs firstly reach cS1 and are then relayed to cS2 (Iwamura  1998; Inui  et al.  2004). Contralateral primary had no significant Granger causation to cS1 until from 80 to 160 ms, disfavoring the possibility of parallel thalamic inputs to cS2. Our results are consistent with findings in previous studies that applied GC in MEG and EEG source waveforms and reported the earliest significant Granger causation occurred from cS1 to cS2 (Hu  et al.  2012; Gao  et al.  2015). The results are also consistent with findings from human and nonhuman primate intracranial recordings (Allison  et al.  1991). Our results suggest that early somatosensory processing progresses in a serial hierarchical way, which may also exist in the cS1 with information serially transmitted from subarea 3b to subarea 1 as evidenced by fine-grained analysis of cS1 subdivision activations evoked by median nerve stimulation (Papadelis  et al.  2011). Differently, Klingner  et al.  (2016) applied dynamical causal modeling (DCM) to model MEG signals and reported evidence supporting parallel thalamic inputs to cS1 and cS2. Comparison of results from DCM and GC could be difficult due to factors such as nodes and connections included in the modeling and the use of prior knowledge of structural connectivity.

Different from previous studies, we included cM1 and cSG in the GC analysis. cM1 received significant information flow from cS1 from 20 to 180 ms, likely through the dense reciprocal connections between M1 and S1 (Gómez  et al.  2021). Contralateral primary motor cortex also received significant information flow from cSG and cS2 before 100 ms. We speculate that information outflow from cS1, cSG, and cS2 to cM1 may be relevant to functions, such as preparation or planning of potential movements triggered by somatosensory cues. cSG is a part of inferior parietal cortex, consists of subregions of BA 40 or parietal area F, has structural connections with cS1, cS2, and frontal cortices, and involves in a serial of functions from sensation, intention, attention, short-term memory, to language (Miquée  et al.  2008; Baker  et al.  2018). Our GC results revealed the earliest significant information flow from cS1 to cSG after stimulus onset and later dynamic mutual information flow between cSG and cM1, cS2, and cS1, likely because of cSG widespread structural connections with these areas. The initial serial information outflow from cS1 and later dynamic parallel information flow between cS1, cM1, cS2, and cSG might be essential for the tight link between somatosensory processing and motor control. Somatosensory information provides cue for planning of a movement and feedback on the sensory state of the body for execution of a planned movement, but due to the noisy and delayed nature of somatosensory feedback, sensory prediction is thought to be needed to update the estimation of the body sensory state and modify motor commands to achieve precise execution of a movement (Shadmehr  et al.  2010). Dynamic interactions between somatosensory cortices, association areas, and motor cortex might be relevant to dynamic update of the sensory state of the body and modification of motor commands.

Limitations

Some limitations should be noted in the current study. First, just as in previous MEG and EEG studies [with the exception of Klingner et al. (2016) that simulated thalamic inputs with a Gaussian function], we did not include the thalamus as a node in the GC analysis. Thalamus is a well-recognized relay for ascending sensory information. Reliable estimation of thalamic neural activity and including the thalamus as a node in the modeling of effective connectivity might be decisive for the elucidation of the serial vs. parallel processing debate. However, it is challenging to obtain accurate estimations or direct recordings of thalamic neural activities in healthy brains. MEG and EEG are generally thought to be incapable of localizing subcortical neural activities (Andersen  et al.  2020). Although fMRI has the capacity to measure subcortical hemodynamic activities, it does not directly measure neural activity. Future investigations may attempt to gain direct measures of thalamic neural activities. Some studies have claimed that neural activity of deep brain sources, such as the thalamus, is detectable with MEG (Papadelis  et al.  2012) or EEG (Seeber  et al.  2019). It may be possible to estimate thalamic activations with MEG or EEG by employing more powerful but challenging experimental manipulations, such as an abundant number of trials or more potent peripheral stimulation. Another caveat is that we did not clarify relationships between Granger causations and somatosensory behavior. This limited our capacity in the interpretation of functions of the significant cortical activations and Granger causations. Although we measured individual touch sensitivity with Von Frey monofilaments, the results had limited range in our cohort as it was expected. Hence, we could not get reliable estimations of correlations between touch sensitivity and cortical activations or Granger causations. In future studies, we plan to employ more delicate experimental tasks to examine these relationships.

Conclusions

The present study employed MEG to elucidate spatiotemporal activations evoked by unilateral pneumatic stimulation in TD children. The contralateral hemisphere showed evident and consistent activations primarily in cS1, cM1, cSG, and cS2 areas, while the ipsilateral hemisphere had weaker and less consistent activations mostly in iS1 and iS2 areas. These results support a distributed cortical somatosensory network. At the individual level, different from the consistency of activations time-locked to the first peak/trough, activations time-locked to the second and later peaks/troughs show variations across participants. Spatiotemporal activation analysis and multivariate GC analysis of MEG source waveforms revealed initial serial cortical somatosensory information outflow from cS1 to cSG, cM1, and cS2 and later dynamic parallel information flow between the consistently activated contralateral cortical areas. Findings from the current study may provide references for the design of future brain stimulation studies and applications for pediatric patients with somatosensory deficits.

Supplementary Material

Supplementary_File_bhae230
supplementary_file_bhae230.pdf (3.9MB, pdf)

Acknowledgments

We gratefully thank Sabrina Shandley, PhD, for help in participant recruitment.

Contributor Information

Yanlong Song, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States; Department of Bioengineering, University of Texas at Arlington, 500 UTA Blvd., Arlington, TX 76010, United States; Departments of Physical Medicine and Rehabilitation and Radiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75390, United States.

Sadra Shahdadian, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States; Department of Bioengineering, University of Texas at Arlington, 500 UTA Blvd., Arlington, TX 76010, United States.

Eryn Armstrong, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States.

Emily Brock, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States.

Shannon E Conrad, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States.

Stephanie Acord, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States.

Yvette R Johnson, NEST Developmental Follow-up Center, Neonatology, Cook Children’s Health Care System, 1521 Cooper St., Fort Worth, TX 76104, United States; Department of Pediatrics, Burnett School of Medicine, Texas Christian University, TCU Box 297085, Fort Worth, TX 76129, United States.

Warren Marks, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States.

Christos Papadelis, Neuroscience Research Center, Jane and John Justin Institute for Mind Health, Cook Children’s Health Care System, 1500 Cooper St., Fort Worth, TX 76104, United States; Department of Bioengineering, University of Texas at Arlington, 500 UTA Blvd., Arlington, TX 76010, United States; Department of Pediatrics, Burnett School of Medicine, Texas Christian University, TCU Box 297085, Fort Worth, TX 76129, United States.

Author contributions

Yanlong Song (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing—original draft, Writing—review & editing) Sadra Shahdadian (Data curation, Formal analysis, Investigation, Methodology, Writing—review & editing), Eryn Armstrong (Data curation, Investigation, Project administration), Emily Brock (Data curation, Investigation, Methodology, Project administration), Shannon E. Conrad (Data curation, Investigation, Methodology, Project administration), Stephanie Acord (Investigation, Methodology, Writing—review & editing), Yvette R. Johnson (Investigation, Methodology, Writing—review & editing), Warren Marks (Investigation, Methodology, Writing—review & editing), and Christos Papadelis (Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing—review & editing).

Funding

This study was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (7R21HD090549 to C.P.).

 

Conflict of interest statement: None declared.

References

  1. Abraira VE, Ginty DD. The sensory neurons of touch. Neuron. 2013:79(4):618–639. 10.1016/j.neuron.2013.07.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allison T, McCarthy G, Wood CC, Williamson PD, Spencer DD. Human cortical potentials evoked by stimulation of the median nerve. II. Cytoarchitectonic areas generating long-latency activity. J Neurophysiol. 1989:62(3):711–722. 10.1152/jn.1989.62.3.711. [DOI] [PubMed] [Google Scholar]
  3. Allison T, Wood CC, McCarthy G, Spencer DD. Cortical somatosensory evoked potentials. II. Effects of excision of somatosensory or motor cortex in humans and monkeys. J Neurophysiol. 1991:66(1):64–82. 10.1152/jn.1991.66.1.64. [DOI] [PubMed] [Google Scholar]
  4. Andersen LM, Jerbi K, Dalal SS. Can EEG and MEG detect signals from the human cerebellum?  NeuroImage. 2020:215:116817. 10.1016/j.neuroimage.2020.116817. Epub 2020 Apr 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Baker CM, Burks JD, Briggs RG, Conner AK, Glenn CA, Taylor KN, Sali G, McCoy TM, Battiste JD, O’Donoghue DL, et al.  A connectomic atlas of the human cerebrum—chapter 7: the lateral parietal lobe. Oper Neurosurg (Hagerstown). 2018:15(suppl_1):S295–S349. 10.1093/ons/opy261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Barnett L, Seth AK. The MVGC multivariate Granger causality toolbox: a new approach to granger-causal inference. J Neurosci Methods. 2014:223:50–68. 10.1016/j.jneumeth.2013.10.018. [DOI] [PubMed] [Google Scholar]
  7. Barnett L, Seth AK. Granger causality for state-space models. Phys Rev E Stat Nonlinear Soft Matter Phys. 2015:91(4):040101. 10.1103/PhysRevE.91.040101. [DOI] [PubMed] [Google Scholar]
  8. Bell AJ, Sejnowski TJ. An information-maximization approach to blind separation and blind deconvolution. Neural Comput. 1995:7(6):1129–1159. 10.1162/neco.1995.7.6.1129. [DOI] [PubMed] [Google Scholar]
  9. Chen TL, Babiloni C, Ferretti A, Perrucci MG, Romani GL, Rossini PM, Tartaro A, Del Gratta C. Human secondary somatosensory cortex is involved in the processing of somatosensory rare stimuli: an fMRI study. NeuroImage. 2008:40(4):1765–1771. 10.1016/j.neuroimage.2008.01.020. [DOI] [PubMed] [Google Scholar]
  10. Chung YG, Han SW, Kim HS, Chung SC, Park JY, Wallraven C, Kim SP. Intra- and inter-hemispheric effective connectivity in the human somatosensory cortex during pressure stimulation. BMC Neurosci. 2014:15:43. 10.1186/1471-2202-15-43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Clarke LE, Barres BA. Emerging roles of astrocytes in neural circuit development. Nat Rev Neurosci. 2013:14(5):311–321. 10.1038/nrn3484. Epub 2013 Apr 18. Erratum in: Nat Rev Neurosci. 2013;14(6):451. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Dale AM, Liu AK, Fischl BR, Buckner RL, Belliveau JW, Lewine JD, Halgren E. Dynamic statistical parametric mapping: combining fMRI and MEG for high-resolution imaging of cortical activity. Neuron. 2000:26(1):55–67. 10.1016/s0896-6273(00)81138-1. PMID: 10798392. [DOI] [PubMed] [Google Scholar]
  13. Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods. 2004:134(1):9–21. 10.1016/j.jneumeth.2003.10.009. PMID: 15102499. [DOI] [PubMed] [Google Scholar]
  14. Espenhahn S, Rossiter HE, van Wijk BCM, Redman N, Rondina JM, Diedrichsen J, Ward NS. Sensorimotor cortex beta oscillations reflect motor skill learning ability after stroke. Brain Commun. 2020:2(2):fcaa161. 10.1093/braincomms/fcaa161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Forss N, Salmelin R, Hari R. Comparison of somatosensory evoked fields to airpuff and electric stimuli. Electroencephalogr Clin Neurophysiol. 1994:92(6):510–517. 10.1016/0168-5597(94)90135-X. [DOI] [PubMed] [Google Scholar]
  16. Ganguly K, Poo MM. Activity-dependent neural plasticity from bench to bedside. Neuron. 2013:80(3):729–741. 10.1016/j.neuron.2013.10.028. PMID: 24183023. [DOI] [PubMed] [Google Scholar]
  17. Gao L, Sommerlade L, Coffman B, Zhang T, Stephen JM, Li D, Wang J, Grebogi C, Schelter B. Granger causal time-dependent source connectivity in the somatosensory network. Sci Rep. 2015:5:10399. 10.1038/srep10399. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Glasser MF, Coalson TS, Robinson EC, Hacker CD, Harwell J, Yacoub E, Ugurbil K, Andersson J, Beckmann CF, Jenkinson M, et al.  A multi-modal parcellation of human cerebral cortex. Nature. 2016:536(7615):171–178. 10.1038/nature18933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gómez LJ, Dooley JC, Sokoloff G, Blumberg MS. Parallel and serial sensory processing in developing primary somatosensory and motor cortex. J Neurosci. 2021:41(15):3418–3431. 10.1523/JNEUROSCI.2614-20.2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Grabner G, Janke AL, Budge MM, Smith D, Pruessner J, Collins DL. Symmetric atlasing and model based segmentation: an application to the hippocampus in older adults. Med Image Comput Assist Interv. 2006:9(Pt 2):58–66. [DOI] [PubMed] [Google Scholar]
  21. Gramfort A, Papadopoulo T, Olivi E, Clerc M. OpenMEEG: opensource software for quasistatic bioelectromagnetics. Biomed Eng Online. 2010:9(1):45. 10.1186/1475-925X-9-45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hämäläinen MS, Ilmoniemi RJ. Interpreting magnetic fields of the brain: minimum norm estimates. Med Biol Eng Comput. 1994:32(1):35–42. 10.1007/BF02512476. PMID: 8182960. [DOI] [PubMed] [Google Scholar]
  23. Hari R, Forss N. Magnetoencephalography in the study of human somatosensory cortical processing. Philos Trans R Soc Lond Ser B Biol Sci. 1999:354(1387):1145–1154. 10.1098/rstb.1999.0470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hautasaari P, Kujala UM, Tarkka IM. Detecting differences with magnetoencephalography of somatosensory processing after tactile and electrical stimuli. J Neurosci Methods. 2019:311:331–337. 10.1016/j.jneumeth.2018.09.014. [DOI] [PubMed] [Google Scholar]
  25. Heller R, Golland Y, Malach R, Benjamini Y. Conjunction group analysis: an alternative to mixed/random effect analysis. NeuroImage. 2007:37(4):1178–1185. 10.1016/j.neuroimage.2007.05.051. [DOI] [PubMed] [Google Scholar]
  26. Hervé M  Package “RVAideMemoire”. 2022. Available at: https://cran.rproject.org/web/packages/RVAideMemoire/index.html (accessed 17 June 2022).
  27. Hlushchuk Y, Hari R. Transient suppression of ipsilateral primary somatosensory cortex during tactile finger stimulation. J Neurosci. 2006:26(21):5819–5824. 10.1523/JNEUROSCI.5536-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Holtmaat A, Svoboda K. Experience-dependent structural synaptic plasticity in the mammalian brain. Nat Rev Neurosci. 2009:10(9):647–658. 10.1038/nrn2699. Erratum in: Nat Rev Neurosci. 2009;10(10):759. PMID: 19693029. [DOI] [PubMed] [Google Scholar]
  29. Hu L, Zhang ZG, Hu Y. A time-varying source connectivity approach to reveal human somatosensory information processing. NeuroImage. 2012:62(1):217–228. 10.1016/j.neuroimage.2012.03.094. [DOI] [PubMed] [Google Scholar]
  30. Iemi L, Busch NA, Laudini A, Haegens S, Samaha J, Villringer A, Nikulin VV. Multiple mechanisms link prestimulus neural oscillations to sensory responses. elife. 2019:8:e43620. 10.7554/eLife.43620. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Inui K, Wang X, Tamura Y, Kaneoke Y, Kakigi R. Serial processing in the human somatosensory system. Cereb Cortex. 2004:14(8):851–857. 10.1093/cercor/bhh043. [DOI] [PubMed] [Google Scholar]
  32. Iscan Z, Nazarova M, Fedele T, Blagovechtchenski E, Nikulin VV. Pre-stimulus alpha oscillations and inter-subject variability of motor evoked potentials in single- and paired-pulse TMS paradigms. Front Hum Neurosci. 2016:10:504. 10.3389/fnhum.2016.00504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Iwamura Y. Hierarchical somatosensory processing. Curr Opin Neurobiol. 1998:8(4):522–528. 10.1016/S0959-4388(98)80041-X. [DOI] [PubMed] [Google Scholar]
  34. Iwamura Y. Bilateral receptive field neurons and callosal connections in the somatosensory cortex. Philos Trans R Soc Lond Ser B Biol Sci. 2000:355(1394):267–273. 10.1098/rstb.2000.0563. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Iwamura Y, Iriki A, Tanaka M. Bilateral hand representation in the postcentral somatosensory cortex. Nature. 1994:369(6481):554–556. 10.1038/369554a0. [DOI] [PubMed] [Google Scholar]
  36. Jones EG, Powell TP. Connexions of the somatic sensory cortex of the rhesus monkey. I. Ipsilateral cortical connexions. Brain. 1969:92(3):477–502. [DOI] [PubMed] [Google Scholar]
  37. Kalberlah C, Villringer A, Pleger B. Dynamic causal modeling suggests serial processing of tactile vibratory stimuli in the human somatosensory cortex—an fMRI study. NeuroImage. 2013:74:164–171. 10.1016/j.neuroimage.2013.02.018. [DOI] [PubMed] [Google Scholar]
  38. Karhu J, Tesche CD. Simultaneous early processing of sensory input in human primary (SI) and secondary (SII) somatosensory cortices. J Neurophysiol. 1999:81(5):2017–2025. 10.1152/jn.1999.81.5.2017. [DOI] [PubMed] [Google Scholar]
  39. Khoshnejad M, Piché M, Saleh S, Duncan G, Rainville P. Serial processing in primary and secondary somatosensory cortex: a DCM analysis of human fMRI data in response to innocuous and noxious electrical stimulation. Neurosci Lett. 2014:577:83–88. 10.1016/j.neulet.2014.06.013. [DOI] [PubMed] [Google Scholar]
  40. Klingner CM, Brodoehl S, Huonker R, Witte OW. The processing of somatosensory information shifts from an early parallel into a serial processing mode: a combined fMRI/MEG study. Front Syst Neurosci. 2016:10:103. 10.3389/fnsys.2016.00103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Korvenoja A, Huttunen J, Salli E, Pohjonen H, Martinkauppi S, Palva JM, Lauronen L, Virtanen J, Ilmoniemi RJ, Aronen HJ. Activation of multiple cortical areas in response to somatosensory stimulation: combined magnetoencephalographic and functional magnetic resonance imaging. Hum Brain Mapp. 1999:8(1):13–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Krubitzer LA, Kaas JH. The organization and connections of somatosensory cortex in marmosets. J Neurosci. 1990:10(3):952–974. 10.1523/JNEUROSCI.10-03-00952.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Krubitzer L, Clarey JC, Tweedale R, Calford MB. Interhemispheric connections of somatosensory cortex in the flying fox. J Comp Neurol. 1998:402(4):538–559. [PubMed] [Google Scholar]
  44. Lamp G, Goodin P, Palmer S, Low E, Barutchu A, Carey LM. Activation of bilateral secondary somatosensory cortex with right hand touch stimulation: a meta-analysis of functional neuroimaging studies. Front Neurol. 2019:9:1129. 10.3389/fneur.2018.01129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Li J, Wu GR, Li B, Fan F, Zhao X, Meng Y, Zhong P, Yang S, Biswal BB, Chen H, et al.  Transcriptomic and macroscopic architectures of intersubject functional variability in human brain white-matter. Commun Biol. 2021:4(1):1417. 10.1038/s42003-021-02952-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Liang M, Mouraux A, Iannetti GD. Parallel processing of nociceptive and non-nociceptive somatosensory information in the human primary and secondary somatosensory cortices: evidence from dynamic causal modeling of functional magnetic resonance imaging data. J Neurosci. 2011:31(24):8976–8985. 10.1523/JNEUROSCI.6207-10.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Lin YY, Shih YH, Chen JT, Hsieh JC, Yeh TC, Liao KK, Kao CD, Lin KP, Wu ZA, Ho LT. Differential effects of stimulus intensity on peripheral and neuromagnetic cortical responses to median nerve stimulation. NeuroImage. 2003:20(2):909–917. 10.1016/S1053-8119(03)00387-2. [DOI] [PubMed] [Google Scholar]
  48. Lopes da Silva F. EEG and MEG: relevance to neuroscience. Neuron. 2013:80(5):1112–1128. 10.1016/j.neuron.2013.10.017. [DOI] [PubMed] [Google Scholar]
  49. Miquée A, Xerri C, Rainville C, Anton JL, Nazarian B, Roth M, Zennou-Azogui Y. Neuronal substrates of haptic shape encoding and matching: a functional magnetic resonance imaging study. Neuroscience. 2008:152(1):29–39. 10.1016/j.neuroscience.2007.12.021. PMID: 18255234. [DOI] [PubMed] [Google Scholar]
  50. Naeije G, Vaulet T, Wens V, Marty B, Goldman S, De Tiège X. Multilevel cortical processing of somatosensory novelty: a magnetoencephalography study. Front Hum Neurosci. 2016:10:259. 10.3389/fnhum.2016.00259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Nevalainen P, Lauronen L, Pihko E. Development of human somatosensory cortical functions—what have we learned from magnetoencephalography: a review. Front Hum Neurosci. 2014:8:158. 10.3389/fnhum.2014.00158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Nihashi T, Naganawa S, Sato C, Kawai H, Nakamura T, Fukatsu H, Ishigaki T, Aoki I. Contralateral and ipsilateral responses in primary somatosensory cortex following electrical median nerve stimulation—an fMRI study. Clin Neurophysiol. 2005:116(4):842–848. 10.1016/j.clinph.2004.10.011. [DOI] [PubMed] [Google Scholar]
  53. Papadelis C, Eickhoff SB, Zilles K, Ioannides AA. BA3b and BA1 activate in a serial fashion after median nerve stimulation: direct evidence from combining source analysis of evoked fields and cytoarchitectonic probabilistic maps. NeuroImage. 2011:54(1):60–73. 10.1016/j.neuroimage.2010.07.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Papadelis C, Leonardelli E, Staudt M, Braun C. Can magnetoencephalography track the afferent information flow along white matter thalamo-cortical fibers?  NeuroImage. 2012:60(2):1092–1105. 10.1016/j.neuroimage.2012.01.054. [DOI] [PubMed] [Google Scholar]
  55. Papadelis C, Ahtam B, Nazarova M, Nimec D, Snyder B, Grant PE, Okada Y. Cortical somatosensory reorganization in children with spastic cerebral palsy: a multimodal neuroimaging study. Front Hum Neurosci. 2014:8:725. 10.3389/fnhum.2014.00725. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Papadelis C, Butler EE, Rubenstein M, Sun L, Zollei L, Nimec D, Snyder B, Grant PE. Reorganization of the somatosensory cortex in hemiplegic cerebral palsy associated with impaired sensory tracts. Neuroimage Clin. 2017:17:198–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Park H, Poo MM. Neurotrophin regulation of neural circuit development and function. Nat Rev Neurosci. 2013:14(1):7–23. 10.1038/nrn3379. PMID: 23254191. [DOI] [PubMed] [Google Scholar]
  58. R Core Team: The R Project for Statistical Computing . 2018. Retrieved 23 May 2022, from  https://www.R-project.org/.
  59. Ragert P, Nierhaus T, Cohen LG, Villringer A. Interhemispheric interactions between the human primary somatosensory cortices. PLoS One. 2011:6(2):e16150. 10.1371/journal.pone.0016150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Seeber M, Cantonas LM, Hoevels M, Sesia T, Visser-Vandewalle V, Michel CM. Subcortical electrophysiological activity is detectable with high-density EEG source imaging. Nat Commun. 2019:10(1):753. 10.1038/s41467-019-08725-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Shadmehr R, Smith MA, Krakauer JW. Error correction, sensory prediction, and adaptation in motor control. Annu Rev Neurosci. 2010:33:89–108. 10.1146/annurev-neuro-060909-153135. PMID: 20367317. [DOI] [PubMed] [Google Scholar]
  62. Song Y, Su Q, Yang Q, Zhao R, Yin G, Qin W, Iannetti GD, Yu C, Liang M. Feedforward and feedback pathways of nociceptive and tactile processing in human somatosensory system: a study of dynamic causal modeling of fMRI data. NeuroImage. 2021:234:117957. 10.1016/j.neuroimage.2021.117957. [DOI] [PubMed] [Google Scholar]
  63. Tadel F, Baillet S, Mosher JC, Pantazis D, Leahy RM. Brainstorm: a user-friendly application for MEG/EEG analysis. Comput Intell Neurosci. 2011:2011:1–13. 10.1155/2011/879716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Tamè L, Braun C, Holmes NP, Farnè A, Pavani F. Bilateral representations of touch in the primary somatosensory cortex. Cogn Neuropsychol. 2016:33(1–2):48–66. 10.1080/02643294.2016.1159547. [DOI] [PubMed] [Google Scholar]
  65. Tamè L, Pavani F, Papadelis C, Farnè A, Braun C. Early integration of bilateral touch in the primary somatosensory cortex. Hum Brain Mapp. 2015:36(4):1506–1523. 10.1002/hbm.22719. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Tan H-RM, Wühle A, Braun C.  Unilaterally applied stimuli in a frequency discrimination task are represented bilaterally in primary somatosensory cortex. Neurol Clin Neurophysiol. 2004:2004:83. [PubMed] [Google Scholar]
  67. Taoka M, Toda T, Hihara S, Tanaka M, Iriki A, Iwamura Y. A systematic analysis of neurons with large somatosensory receptive fields covering multiple body regions in the secondary somatosensory area of macaque monkeys. J Neurophysiol. 2016:116(5):2152–2162. 10.1152/jn.00241.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Tesche CD, Uusitalo MA, Ilmoniemi RJ, Huotilainen M, Kajola M, Salonen O. Signal-space projections of MEG data characterize both distributed and well-localized neuronal sources. Electroencephalogr Clin Neurophysiol. 1995:95(3):189–200. 10.1016/0013-4694(95)00064-6. [DOI] [PubMed] [Google Scholar]
  69. Thompson MJ. Piaget’s stages of cognitive development and Erikson’s stages of psychosocial development. In: Laver-Bradbury C, Thompson MJ, Gale C, Hooper CM, editors. Child and adolescent mental health: theory and practice. 3rd ed. Boca Raton: CRC Press; 2021. [Google Scholar]
  70. Van Essen DC. A population-average, landmark- and surface-based (PALS) atlas of human cerebral cortex. NeuroImage. 2005:28(3):635–662. 10.1016/j.neuroimage.2005.06.058. Epub 2005 Sep 19. [DOI] [PubMed] [Google Scholar]
  71. Whitlock JR. Posterior parietal cortex. Curr Biol. 2017:27(14):R691–R695. 10.1016/j.cub.2017.06.007. [DOI] [PubMed] [Google Scholar]
  72. Woods DL, Courchesne E. Intersubject variability elucidates the cerebral generators and psychological correlates of ERPs. Electroencephalogr Clin Neurophysiol Suppl. 1987:40:293–299. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary_File_bhae230
supplementary_file_bhae230.pdf (3.9MB, pdf)

Articles from Cerebral Cortex (New York, NY) are provided here courtesy of Oxford University Press

RESOURCES