Spinal cord microstructural changes are connected with the aberrant sensorimotor cortical oscillatory activity in adults with cerebral palsy

Abstract

Previous animal models have illustrated that reduced cortical activity in the developing brain has cascading activity-dependent effects on the microstructural organization of the spinal cord. A limited number of studies have attempted to translate these findings to humans with cerebral palsy (CP). Essentially, the aberrations in sensorimotor cortical activity in those with CP could have an adverse effect on the spinal cord microstructure. To investigate this knowledge gap, we utilized magnetoencephalographic (MEG) brain imaging to quantify motor-related oscillatory activity in fourteen adults with CP and sixteen neurotypical (NT) controls. A subset of these participants also underwent cervical-thoracic spinal cord MRI. Our results showed that the strength of the peri-movement beta desynchronization and the post-movement beta rebound were each weaker in the adults with CP relative to the controls, and these weakened responses were associated with poorer task performance. Additionally, our results showed that the strength of the peri-movement beta response was associated with the total cross-sectional area of the spinal cord and the white matter cross-sectional area. Altogether these results suggest that the altered sensorimotor cortical activity seen in CP may result in activity-dependent plastic changes within the spinal cord microstructure, which could ultimately contribute to the sensorimotor deficits seen in this population.

Figure 1

Flanker task design. For each trial, the participants fixated on a crosshair for 1500 ± 50 ms, then a series of five arrows appeared for 2500 ms. Upon arrows appearing, the participants responded with their right hand regarding the center arrow pointing to the left (2nd digit) or right (3rd digit). In the congruent condition, the flanking arrows pointed in the same direction as the middle arrow. In the incongruent condition, the flanking arrows pointed in the opposite direction than the middle arrow.

Figure 2

Accuracy between groups and conditions in the flanker task. (A) Accuracy (in %) was higher in the congruent versus the incongruent condition across all participants. (B) Accuracy was also higher in the neurotypical (NT) controls compared with the individuals with cerebral palsy (CP). Error bars reflect the standard error of the mean. *P < 0.05.

Figure 3

Sensorimotor spectrograms and group averaged beamformer images. (A) Averaged time–frequency spectrogram from a sensor located near the hand region of the contralateral (left) motor cortex averaged across all participants depicting the motor-related beta responses in the 0–50 Hz range. Time is denoted on the x-axis in milliseconds with 0.0 ms being movement onset, and frequency (Hz) is denoted on the y-axis. There was a beta ERD (18–26 Hz) that began about − 300 ms prior to movement onset and lasted until about 300 ms when the movement finished. After movement termination, there was a post-movement beta rebound (16–22 Hz) that lasted from about 500 to 800 ms before beginning to dissipate. (B) Averaged time–frequency spectrogram from a sensor located near the hand region of the contralateral (left) motor cortex depicting the 50–100 Hz range averaged across all participants. Beginning before movement onset around − 100 ms and lasting until about 100 ms after movement was a gamma ERS. Averaged beamformer images of the beta ERD (C) and the PMBR (D) depict that these responses were both located within the left sensorimotor cortices. (E) Group averaged image of the gamma ERS located in the contralateral (left) motor cortices.

Figure 4

Beta ERD neural timecourse. (A) Neural timecourses of the beta ERD response in individuals with CP (red) and the NT group (blue) extracted from the peak voxel of the beta ERD response. The black line at 0.0 ms denotes movement onset, and the gray box is the window of interest (− 300 to 300 ms). (B) Bar graph depicting the difference in beta ERD strength between the NT and CP groups. As depicted, the beta ERD response was stronger in the NT group relative to those with CP. *P < 0.05.

Figure 5

PMBR neural timecourse (A) Neural timecourses of the PMBR response in individuals with CP (red) and the NT group (blue) extracted from the peak voxel of the PMBR. The black line at 0.0 ms denotes movement onset, and the gray box is the window of interest (500 to 800 ms). (B) Bar graph depicting the difference in PMBR strength between the NT and CP groups. Overall, those with CP had a weaker PMBR compared to controls. *P < 0.05.

Figure 6

Total cross-sectional area (CSA) of the spinal cord. Individuals with cerebral palsy (CP) had a smaller total CSA than the neurotypical controls (NT). *P < 0.05.

Figure 7

Cortical oscillations drive changes in spinal cord microstructure. Left: Representative T1 weighted images of the brain and spinal cord from an adult with CP. The light blue circles denote the area of the brain and spinal cord that the strength of the beta ERD and white matter cross sectional area (CSA) were taken, respectively (Talairach coordinates: − 42, − 24, 21; spinal cord section: C6–T3). Middle: Representative beta ERD response (top) and CSA of the spinal cord (white matter depicted in light blue and gray matter depicted in yellow). Right: Scatterplot depicting the relationship between the strength of the beta ERD and the white matter CSA. Individuals with a stronger beta ERD tended to have more white matter within their upper spinal cord (r = − 0.40, P = 0.048) and more total CSA (r = − 0.43, P = 0.031). Blue squares denote controls while red triangles reflect those with CP.