Children with Cerebral Palsy Hyper-Gate Somatosensory Stimulations of the Foot

Abstract

We currently have a substantial knowledge gap in our understanding of the neurophysiological underpinnings of the sensory perception deficits often reported in the clinic for children with cerebral palsy (CP). In this investigation, we have begun to address this knowledge gap by using magnetoencephalography (MEG) brain imaging to evaluate the sensory gating of neural oscillations in the somatosensory cortices. A cohort of children with CP (Gross Motor Function Classification System II–III) and typically developing children underwent paired-pulse electrical stimulation of the tibial nerve during MEG. Advanced beamforming methods were used to image significant oscillatory responses, and subsequently the time series of neural activity was extracted from peak voxels. Our experimental results showed that somatosensory cortical oscillations (10–75 Hz) were weaker in the children with CP for both stimulations. Despite this reduction, the children with CP actually exhibited a hyper-gating response to the second, redundant peripheral stimulation applied to the foot. These results have further established the nexus of the cortical somatosensory processing deficits that are likely responsible for the degraded sensory perceptions reported in the clinic for children with CP.

Introduction

Almost one of every 500 children incurs a prenatal or perinatal brain injury that results in cerebral palsy (CP; Graham et al. 2016). The current definition of CP recognizes that the motor impairments seen throughout development in these children are partially a product of aberrant somatosensory processing (Rosenbaum et al. 2007). This notion is supported by the numerous clinical reports of proprioception, stereognosis, and tactile discrimination deficits seen in the upper and lower extremities of these children (Cooper et al. 1995; Clayton et al. 2003; Sanger and Kukke 2007; Wingert et al. 2009, 2010; Auld et al. 2012; Maitre et al. 2012; Damiano et al. 2013; Robert et al. 2013). There has been a growing interest in identifying the neurophysiological underpinnings of these somatosensory processing deficits, and to decipher if they can be resolved through current physical therapy based treatment approaches (Kurz et al. 2012; Robert et al. 2013).

Numerous studies have used magnetoencephalography (MEG) and electroencephalography to examine the sensory processing deficits seen in children with CP following a peripheral stimulation. The overwhelming consensus from these investigations has been that the somatosensory evoked-potentials/fields for the hand, foot, and lips are diminished, and in some cases latent in children with CP (Kulak et al. 2005, 2006; Riquelme and Montoya 2010; Kurz and Wilson 2011; Teflioudi et al. 2011; Guo et al. 2012; Kurz et al. 2012; Maitre et al. 2012; Papadelis et al. 2014). Our recent MEG studies in this area have extended these findings by revealing that theta-alpha (4–14 Hz) and beta (18–34 Hz) oscillations are aberrant in somatosensory cortices following stimulation of the mechanoreceptors of the foot and hands of children with CP (Kurz et al. 2014; Kurz, Becker, et al. 2015; Kurz, Heinrichs-Graham, et al. 2015). Moreover, we have identified that these anomalous somatosensory cortical oscillations are strongly linked with the degree of motor errors, muscular weakness, and the mobility deficits seen in these children.

Sensory gating is a well-known neurophysiological process that is associated with a weakened cortical response to the second stimulus when 2 identical sensory stimuli are presented sequentially (Edgar et al. 2005; Thoma et al. 2007; Weiland et al. 2008; Cheng and Lin 2013; Hsiao et al. 2013; Cheng et al. 2015; Wiesman et al. 2017). Such weakened cortical activity is thought to reflect active filtering of incoming sensory information that is redundant (i.e., does not provide new information; Cromwell et al. 2008). Numerous investigations have shown that sensory gating is aberrant in a wide range of disorders such as schizophrenia, autism, and Alzheimer’s disease (Adler et al. 1982; Jessen et al. 2001; Cromwell et al. 2008; Matsuzaki et al. 2014). Potentially, the aberrant gating seen in these patient populations is directly related to the perceptual processing deficits that are common features across these different disorders.

Our prior experimental work has established that the perceptual deficits seen in children with CP are likely a result of anomalous oscillatory activity in somatosensory cortices (Kurz et al. 2014; Kurz, Becker, et al. 2015; Kurz, Heinrichs-Graham, et al. 2015). However, these studies all used a single stimulation paradigm and whether such aberrant somatosensory oscillations will be sustained when sequential, redundant stimulation is applied to the foot remains unknown. To address this knowledge gap, we used MEG imaging to evaluate oscillatory changes within the somatosensory cortices of typically developing (TD) children and children with CP during a paired-pulse electrical stimulation paradigm applied to the tibial nerve. Our key hypotheses were: (1) that oscillatory somatosensory responses would be diminished in children with CP following the first peripheral stimulation (Kurz et al. 2014; Kurz, Becker, et al. 2015; Kurz, Heinrichs-Graham, et al. 2015) and (2) that somatosensory gating responses would be aberrant in children with CP.

Methods

Participants

Fifteen children with a diagnosis of either spastic diplegia or hemiplegia CP (Males = 12; Age = 14.7 ± 0.7 years), and a Gross Motor Function Classification System (GMFCS) level between II and III participated in this investigation (see Table 1). An additional 19 demographically matched TD children (Males = 11; Age = 14.2 ± 0.6 years) with no neurological or musculoskeletal impairments served as a control group. The children with CP were excluded if they had an orthopedic surgery or antispasticity treatments within the last 6 months, and if they had undergone a dorsal rhizotomy. The Institutional Review Board at the University of Nebraska Medical Center reviewed and approved this investigation. Informed consent was acquired from the parents and the children assented to participate in the experiment.

Table 1.

Description of the children with CP that participated in this investigation

Subject	Leg tested	Gender	Age	Presentation	GMFCS	MRI remarks
1	R	Male	18	Spastic diplegic	II	Periventricular Leukomalacia
2	L	Male	18	Hemiplegic	II	Periventricular Leukomalacia
3	L	Male	17	Spastic diplegic	II	Periventricular Leukomalacia
4	R	Male	16	Spastic diplegic	II	Periventricular Leukomalacia
5	R	Male	17	Spastic diplegic	III	Periventricular Leukomalacia
6	R	Male	10	Hemiplegia	II	Periventricular Leukomalacia; Encephalomalacia
7	R	Male	10	Spastic diplegia	III	Periventricular Leukomalacia; volume loss in the corpus callosum
8	R	Male	13	Spastic diplegic	III	Periventricular Leukomalacia; volume loss in the corpus callosum
9	L	Female	17	Spastic diplegic	III	Periventricular Leukomalacia; volume loss in the corpus callosum
10	L	Male	15	Spastic diplegic	III	Periventricular Leukomalacia
11	L	Male	15	Spastic diplegic	II	No notable lesions or volume loss
12	L	Female	13	Spastic diplegic	II	Periventricular Leukomalacia
13	R	Male	18	Spastic diplegic	III	Periventricular Leukomalacia
14	R	Female	15	Spastic diplegic	II	Periventricular Leukomalacia
15	L	Male	13	Spastic diplegic	III	Periventricular Leukomalacia

The remarks on the magnetic resonance images were based on a neuroradiology review. R, right leg; L, left leg.

MEG Data Acquisition and Experimental Paradigm

The neuromagnetic responses were acquired with a bandwidth of 0.1–330 Hz and were sampled continuously at 1 kHz using an Elekta MEG system (Helsinki, Finland) that had 306 magnetic sensors (204 planar gradiometers and 102 magnetometers). All recordings were conducted in a one-layer magnetically shielded room with active shielding engaged for advanced environmental noise compensation. During data acquisition, the children were monitored via real-time audio–video feeds from inside the shielded room.

During the experiment, participants were seated with their eyes closed in a custom-made nonmagnetic chair with their head positioned within the helmet-shaped MEG sensor array. We chose to have the children close their eyes during the experiment to limit the possibility of eye blinks contaminating the data. Each child wore a custom-made head stabilization device, which consisted of an airbag system that was inflated with a hand pump until the void between the Dewar and the head was completely filled. This device minimized any head movements that might occur during the data collections, and thus helped maintain the head’s position relative to the magnetic sensors. Once the child was positioned for MEG, unilateral electrical stimulation was applied to the posterior tibial nerve of the most affected leg of the children with CP and the nondominant leg of the TD children using an external cutaneous stimulator that was positioned posterior to the medial malleolus (Digitimer DS7A, HW Medical Products). The more affected leg was determined through manual muscle testing and the participant’s self-report of which leg they had more difficulty controlling. The more affected leg was assumed to have more significant somatosensory impairments. For the TD children, we chose to stimulate the nondominate leg because we felt that it was a more equal comparison with the impaired leg used for the children with CP. For each child, 120 paired-pulse trials were collected using an interstimulus interval of 500 ms, and an interpair interval that randomly varied between 4.5 and 4.8 s. Each pulse was comprised of a 0.2 ms constant-current square wave that was set to 10% above the motor threshold required to elicit a subtle but visible flexor twitch in the hallux. A 500 ms interval between the pulse pairs was used to evaluate somato-gating response. A 500 ms interval between the pulse pairs is known to elicit robust somatosensory gating responses (Adler et al. 1982; Bak et al. 2011; Hsiao et al. 2013).

Structural Magnetic Resonance Imaging Processing and MEG Coregistration

Structural magnetic resonance imaging (MRI) data were acquired using a Philips Achieva 3 T scanner. High-resolution T1-weighted sagittal images were obtained with an eight-channel head coil using a 3D fast field echo sequence with the following parameters: FOV: 24 cm, 1 mm slice thickness, no gap, in-plane resolution of 1.0 × 1.0 mm and sense factor of 2.0.

Prior to the start of the MEG experiment, 4 coils were affixed to the head of the child. The location of these coils, 3 fiducial points, and the scalp surface were digitized to determine their 3D position relative to each other (Fastrak 3SF0002, Polhemus Navigator Sciences, Colchester, VT, USA). Head localization was accomplished by continuously feeding an electric current with a unique frequency label (e.g., 322 Hz) to each of the coils during the data collection. This induced a measurable magnetic field, and allowed each coil to be localized in reference to the sensors throughout the recording session. Because the coil locations were also known in head coordinates, all MEG measurements could be transformed into a common coordinate system. With this coordinate system, each child’s functional MEG data were coregistered with native space structural T1-weighted MRI data. Briefly, the structural MRI data were aligned parallel to the anterior and posterior commissures, and then coregistered with the MEG data using a least-squares fit. After beamforming (see below), the structural MRI data were transformed into standardized space (Talairach and Tournoux 1998) using a direct 3D-spline interpolation, and the resulting transformation matrix was applied to the functional output images from beamforming. Coregistering and transforming these data used the BESA MRI software (Version 2.0).

MEG Preprocessing, Time-Frequency Transformation, and Sensor-Level Statistics

Each MEG data set was individually corrected for any head motion that may have occurred during task performance, and was subjected to noise reduction using the signal space separation method with a temporal extension (Taulu et al. 2005). The continuous magnetic time series was divided into epochs of 2200 ms duration, with the baseline being defined as −700 to −300 ms before initial stimulus onset. Of note, we shifted our baseline away from the period immediately preceding stimulus onset to avoid potential contamination by any anticipatory responses. Artifact rejection was based on a fixed threshold method and supplemented with visual inspection. The mean amplitude threshold for declaring epochs as artifactual (i.e., bad) was 1.29 pT, which resulted in an epoch acceptance rate of 86% (104 epochs) in TD children and 82% (98 epochs) in children with CP. This difference was not significant (P = 0.18).

The artifact-free epochs were transformed into the time-frequency domain using complex demodulation, and the resulting spectral power estimations per sensor were averaged over trials to generate time-frequency plots of the mean spectral density. These sensor-level data were normalized by dividing the power value of each time-frequency bin by the respective bin’s baseline power, which was calculated as the mean power over the −700 ms to −300 ms time period. The specific time-frequency windows used for imaging were determined by a fully data-driven approach that began with a statistical analysis of the sensor-level spectrograms across the entire array of gradiometers. Each data point in the spectrogram was initially evaluated using a mass univariate approach based on the general linear model. To reduce the risk of false-positive results while maintaining reasonable sensitivity, a two-stage procedure was followed to control for Type 1 error. In the first stage, one-sample t-tests were conducted on each data point and the output spectrogram of t-values was thresholded at P < 0.05 to define time-frequency bins containing potentially significant oscillatory deviations across all participants. In stage 2, time-frequency bins that survived the threshold were clustered with temporally and/or spectrally neighboring bins that were also above the threshold (P < 0.05), and a cluster value was derived by summing all of the t-values of all data points in the cluster. Nonparametric permutation testing was then used to derive a distribution of cluster values, and the significance level of the observed clusters (from stage 1) was tested directly using this distribution (Ernst 2004; Maris and Oostenveld 2007). For each comparison, at least 10 000 permutations were computed to build a distribution of cluster values. Based on these analyses, the time-frequency windows that contained significant oscillatory events across all participants (described in the Results section) were subjected to a beamforming analysis.

MEG Source Imaging and Statistics

A minimum variance vector beamforming algorithm was employed to calculate the source power across the entire brain volume (Gross et al. 2001; Hillebrand et al. 2005). The single images were derived from the cross spectral densities of all combinations of MEG gradiometers within the time-frequency ranges of interest, and the solution of the forward problem for each location on a grid specified by input voxel space. Following convention, the source power in these images was normalized per subject using a separately averaged prestimulus noise period of equal duration and bandwidth (van Veen et al. 1997; Hillebrand et al. 2005). Such images are typically referred to as pseudo-t maps, with units (pseudo-t) that reflect noise-normalized power differences (i.e., active vs. baseline) per voxel. Thus, the normalized power per voxel for each response of interest was computed over the entire brain volume per participant at 4.0 × 4.0 × 4.0 mm resolution. As mentioned above, each child’s functional images, which were coregistered to native space anatomical images prior to beamforming, were transformed into standard space using the transform that was previously applied to the structural MRI volume and spatially resampled. MEG preprocessing and imaging used the BESA software (BESA v6.0; Grafelfing, Germany).

Independent of TD/CP group status, participants were first divided into separate groups based on whether the peripheral stimulation was applied to their right or left leg. To compare differences in somatosensory cortical activity between the children with CP and TD children, we first extracted a virtual sensor (i.e., voxel time series) from the peak voxel of the grand averaged beamformer images per left/right stimulation group. Essentially, we imaged the response to each peripheral stimulation per participant using a beamformer, and then averaged all output pseudo-t maps across participants within the right leg stimulation group and separately for the left leg stimulation group. We then extracted the peak coordinate from each of these 2 images, which were located in the leg region of the contralateral somatosensory cortices. To compute the virtual sensors, we applied the sensor weighting matrix derived through the forward computation to the preprocessed signal vector, which yielded a time series for the specific coordinate in source space. Note that this virtual sensor extraction was done per participant individually, once the coordinates of interest were known, and that the purpose of the right/left leg stimulation groups was only to extract peak coordinates to facilitate TD/CP group comparisons at the virtual sensor level. The maximums that occurred in these time courses after each stimulus were subsequently extracted per participant, and a mixed-model ANOVA (Group × Stimulation) was employed to evaluate differences between TD children and those with CP after the respective peripheral stimulations. Somatosensory gating was also calculated by taking the ratio of the respective local maximums. A value closer to one indicated that there was less somatosensory gating, while a value less than 0.5 indicated a strong gating response. An independent t-test was used to determine if there was a significant difference between the somatosensory gating of the respective groups. All statistical analyses were conducted at the 0.05 alpha level.

Results

Sensor-Level and Beamforming Analyses

Analysis of the sensor-level spectrograms showed 2 broadband (10–75 Hz) synchronizations, each beginning shortly after the onset of the respective stimulations and continuing for about 100 ms (P < 0.001, corrected; Fig. 1). These 2 synchronizations were observed across a cluster of gradiometers near the superior and medial fronto-parietal region. Separate beamformer images for the 10–75 Hz range were then computed for each participant using the 100 ms time window that immediately followed each stimulation onset (i.e., 0–100 and 500–600 ms), and a common baseline period of the same duration and bandwidth. The 0–100 and 500–600 ms images were subsequently grand averaged for each leg group to determine the location of the peak activity within the somatosensory cortices (Fig. 2). These leg specific maps showed a robust synchronization in the paracentral lobule of the contralateral somatosensory cortices (i.e., near the expected leg representation for somatosensation).

Figure 1.

Grand averaged and group time-frequency spectrograms from a gradiometer sensor located near the sensorimotor leg region (i.e., the same sensor was averaged across all participants). Time (in ms) is denoted on the x-axis, with 0 and 500 ms defined as the onset of the first and second stimulations, respectively. Spectral power is expressed as the percent difference from the baseline period (−700 to −300 ms). As shown, neuronal activity strongly increased across a broad frequency range (10–75 Hz) for roughly 100 ms in response to each stimulation (areas are outlined with dotted lines). A color scale bar is shown to the right. A 2D map of the sensor array is also shown to illustrate the sensors where significant 10–75 Hz responses were detected. As can be discerned, these clustered around the medial sensorimotor cortices, which is consistent with the somatotopic organization for the foot area.

Figure 2.

Beamformer images of each neural response were computed separately per participant, and then averaged across all participants who had the same leg stimulated. Specifically, images were calculated from 10 to 75 Hz for the 100 ms time windows that immediately followed each stimulation onset. The images for each stimulation were subsequently averaged across those who received stimulation on the left leg, and separately for who received right leg stimulation, to determine the location of the peak voxel within the somatosensory cortices. These peak voxels were then used as the seed for extracting voxel time series in each participant, and these time series were used for group comparisons between children with CP and TD children. Regardless of leg stimulated, a robust synchronization within the somatosensory cortices was detected in the contralateral hemisphere with peak activity being almost symmetrical between groups. Images are displayed in neurological convention.

Virtual Sensor Analysis

In each participant, virtual sensor time series data were extracted from the peak coordinate derived through the analysis of their respective leg group’s grand averaged beamformer images. These data were then averaged separately across TD children and those with CP, regardless of which leg was stimulated, to identify the temporal dynamics (Fig. 3) and evaluated for group effects using mixed-model ANOVA. Statistical analyses revealed that there was an amplitude main effect, which indicated that somatosensory activity was attenuated for the second stimulation for both the TD children and children with CP (Stimulation 1 = 21 ± 3%; Stimulation 2 = 14 ± 2%; P = 0.0001). There was also a group main effect (CP = 11 ± 2%; TD = 23 ± %; P = 0.013) indicating that somatosensory cortical activity was much weaker overall for the children with CP. We also found a significant group × amplitude interaction (P = 0.04) and post hoc analysis indicated that compared with TD children, the children with CP had more attenuated cortical responses to the first (CP = 14 ± 3%; TD = 26 ± 5%; P = 0.037) and second (CP = 8 ± 2%; TD = 18 ± 3%; P = 0.006) peripheral stimulations. Lastly, the children with CP displayed stronger somatosensory cortical gating (CP = 0.45 ± 0.08; TD = 0.75 ± 0.05; P = 0.004).

Figure 3.

Group averaged virtual sensor time courses (i.e., voxel time series) of the peak neural activity in the contralateral somatosensory cortices of the TD children (top) and children with CP (bottom). Time (in ms) is denoted on the x-axis, with 0 and 500 ms defined as the onset of the first and second stimulations, respectively (grayed areas). The amplitude of neural activity relative to the baseline is shown on the y-axis. The somatosensory activity was clearly diminished in the children with CP. In addition, the amplitude of the somatosensory cortical response to the first stimulation was greater than that to the second stimulation, and the magnitude of this gating effect was significantly stronger (i.e., hyper-gating) in the children with CP relative to the TD children.

Discussion

It is widely recognized that children with CP often have somatosensory processing deficits, but the neurophysiological nexus of these perceptual deficits is not completely understood. The experiment conducted in this investigation was aimed at addressing this knowledge gap by using MEG to evaluate somatosensory gating, based on a paired-pulse electrical stimulation paradigm applied to the tibial nerve, in children with CP and TD children. The primary outcome of our investigation was that children with CP have a hyper-gating response within the primary somatosensory cortices when repeated, redundant peripheral stimulations are applied to the foot. In addition, we found that somatosensory cortical oscillations were weaker overall. Altogether, these results clearly show that the perinatal brain injuries experienced by these children caused long-term alterations in the neurophysiology of somatosensory cortical processing networks.

Sensory gating is a ubiquitous cortical response to repetitive, redundant peripheral stimuli (Edgar et al. 2005; Thoma et al. 2007; Weiland et al. 2008; Cheng and Lin 2013; Hsiao et al. 2013; Cheng et al. 2015; Wiesman et al. 2017). The current consensus is that gating responses represent the selective filtering of sensations that do not provide new information. Our results are significant because they are the first to show that children with CP hyper-gate repeated somato-sensations of the foot, which is remarkable because it well known that children with CP have weak somatosensory cortical responses to single stimulations that are applied to the hands and feet (Kulak et al. 2005, 2006; Kurz and Wilson 2011; Teflioudi et al. 2011; Kurz et al. 2012, 2014; Maitre et al. 2012; Papadelis et al. 2014; Kurz, Becker, et al. 2015; Kurz, Heinrichs-Graham, et al. 2015). Potentially, the hyper-gating of subsequent somato-sensations may even further diminish the processing of important peripheral information. Prior diffusion tensor imaging studies have shown that damage to the thalamocortical tracts is likely related to the somatosensory processing impairments seen in children with CP (Rose et al. 2007; Trivedi et al. 2008, 2010; Hoon et al. 2009). These results suggest that the abnormal somatosensory gating reported in this investigation may be directly related to perinatal damage to the thalamocortical fiber tracts. We suspect that such fiber tract damage may alter the signal-to-noise ratio in a way that the somatosensory cortices are biased toward hyper-gating incoming sensory stimulation in order to reduce the overall neurological noise in somatosensory processing networks. Although this hypothesis is clearly plausible, there is currently a substantial gap in our understanding of the relationship between the structural damage seen in children with CP and the parallel deficits in the somatosensory cortical activity. Thus, future studies should take a multimodal approach and try to link deficits across structural and functional domains in the same children.

The children with CP also had weaker somatosensory cortical activity following both the first and second peripheral stimulation. As mentioned above, this finding is well aligned with what has been previously presented in the literature (Kulak et al. 2005, 2006, 2014; Riquelme and Montoya 2010; Kurz and Wilson 2011; Teflioudi et al. 2011; Guo et al. 2012; Kurz et al. 2012; Maitre et al. 2012; Papadelis et al. 2014; Kurz, Becker, et al. 2015; Kurz, Heinrichs-Graham, et al. 2015). However, our results significantly extend the current literature by characterizing the oscillatory aspect (vs. the time-domain response) of these somatosensory responses. Our prior MEG studies identified that the somatosensory cortical oscillations induced by stimulations applied to the foot primarily reside at the beta (18–34 Hz) frequency (Kurz et al. 2014; Kurz, Heinrichs-Graham, et al. 2015), while in this investigation we have identified that the 10–75 Hz range is aberrant after stimulation of the tibial nerve. The major difference across these results is likely attributable to the different experimental approaches used to stimulate the somatosensory system. In our prior studies, we used mechanical stimulation (e.g., air-puffs) to selectively stimulate only the foot mechanoreceptors, while in this investigation we employed an electrical stimulation paradigm that activated numerous receptors. The presence of high-frequency oscillations in the somatosensory system has been supported by several previous studies conducted in both healthy adults and children (Hirata et al. 2002; Gaetz and Cheyne 2003; Ihara et al. 2003; Bauer et al. 2006; Wiesman et al. 2017). Nevertheless, our results are the first to show that these high-frequency oscillations are also aberrant in the somatosensory cortices of children with CP.

The field’s recognition of the impact of sensory deficits on the motor performance errors seen in children with CP is rapidly growing across the clinical literature. Several studies have reported that sensory discrimination deficits are linked with the child’s motor impairments (Gordon and Duff 1999; Krumlinde-Sundholm and Eliasson 2002; Sakzewski et al. 2010), and the likelihood that the child will learn a new motor skill (Robert et al. 2013). Hence, the broader clinical question is how these pervasive somatosensory processing deficits can be overcome. Behavioral results from a previous investigation have suggested that somato-sensations are recalibrated after practicing a new motor skill (Ostry et al. 2010). Based on these results, we suspect that increased practice with the impaired limb may result in enhanced neural activity within the somatosensory processing networks of children with CP. This premise is partially support by a preliminary study we conducted that suggested that the somatosensory cortical activity might be enhanced after children with CP undergo an intensive gait training protocol (Kurz et al. 2012). Further testing of these therapeutic concepts would be laudable, and may have the potential to alter the treatment strategies currently being used to improve the mobility and lower extremity motor actions of children with CP.

It is well appreciated that the time that the developing infant incurs a noxious event often predicts the type of brain abnormality seen on an MRI (Krageloh-Mann 2004; Himmelmann et al. 2017). For example, preterm infants that incur insults during the first to second trimester display maldevelopments on their MRIs, while insults that occur during the early third trimester more often present white mater injuries. The nature of these time specific brain injuries have been suggested to account for the different sensorimotor presentation seen in children with CP (Himmelmann and Uvebrant 2011; Himmelmann et al. 2017). The children that participated in this investigation primarily had white matter injuries on their MRIs, which suggests that their brain injuries occurred during the early portion of the third trimester. It is possible that the hyper-gating somatosensory cortical response seen in these children may be related to the timing of the preterm brain injury. However, this conjecture should be interpreted with caution because the relationship between the timing of the brain insults seen in children with CP and the resultant alterations in their brain activity is not well understood.

Funding

This work was partially supported by grants from the National Institutes of Health (1R01-HD086245) and the National Science Foundation (NSF 1539067).