Reorganization of the Action Observation Network and Sensory-Motor System in Children with Unilateral Cerebral Palsy: An fMRI Study

Abstract

Little is known about the action observation network (AON) in children with unilateral cerebral palsy (UCP). Using fMRI, we aimed to explore AON and sensory-motor network (SMN) in UCP children and compare them to typically developed (TD) children and analyse the relationship between AON (re-)organization and several neurophysiological and clinical measures. Twelve UCP children were assessed with clinical scales and transcranial magnetic stimulation (TMS). For the fMRI study, they underwent a paradigm based on observation of complex and simple object-manipulation tasks executed by dominant and nondominant hand. Moreover, UCP and TD children carried out a further fMRI session to explore SMN in both an active motor and passive sensory task. AON in the UCP group showed higher lateralization, negatively related to performances on clinical scales, and had greater activation of unaffected hemisphere as compared to the bilateral representation in the TD group. In addition, a good congruence was found between bilateral or contralateral activation of AON and activation of SMN and TMS data. These findings indicate that our paradigm might be useful in exploring AON and the response to therapy in UCP subjects.

Figures

Figure 1

(a) Examples, taken from a single frame, of the six videoclips showing object manipulation performed by the right hand in three different contexts (“cubes,” “piano,” and “key”): three complex actions (A, B, C) and three simple actions with the same object (D, E, F). (G, H, I) Initial static frames of the corresponding action types, used as BASELINE conditions. (b) Diagram of the functional series presented to children: the block design comprises two TASK blocks for each of the four different conditions (CR, SR, CL, and SL) for complex (C) or simple (S) actions performed by the right (R) or left (L) hand, alternating with the same number of BASELINE blocks. Each block lasts 24 seconds and is composed of the random sequence of the 8-second videoclips of hand actions or still pictures of the resting hands. The presentation of the different conditions in the TASK blocks was completely randomized. Each functional series included four initial extra scans (12 s) to allow the stabilization of signal. Reproduced with permission (Copyright © 2015 John Wiley & Sons Ltd) from Biagi et al. [5].

Figure 2

Representative slices of 3D T1-weighted images depicting the brain lesion of each UCP child. Numbers correspond to the ID of UCP children as reported in Table 1.

Figure 3

Probabilistic functional maps of the action observation circuit for the ALL TASKS > BASELINE contrast in TD children (as previously published in Biagi et al. 2015, top [5]) and in UCP children (bottom). Colour bar represents different levels of probability of activation for the action observation task from 33% (red, meaning that a brain region appeared in the map only if it was activated in at least 4 subjects) to 83% (cyan, equivalent to areas activated by more than 10 subjects). For each transversal slice, the z Tailarach's coordinate is indicated. On the right, we represent the left hemisphere for TD children (radiological convention; R = right, L = left) and the unaffected hemisphere, that is, the hemisphere contralateral to the unaffected hand, in UCP children (unA = unaffected, A = affected); on the left, the right hemisphere for TD and the hemisphere contralateral to the affected hand in UCP.

Figure 4

Box plots of the laterality indices of each single subject for AON (top row) of the dominant hand (DH) (a) and of the nondominant hand (non-DH) (b), as well as for the primary sensory-motor cortex, pSMC (bottom row), for the stimulation of the dominant hand (c) and of the nondominant hand (d). UCP children are indicated by light grey triangles, TD children by black circles. Each box is defined by the 25th and 75th percentiles. The whiskers are determined by the minimum and maximum values; the square indicates the mean value, while the line corresponds to the median value.

Figure 5

Lateralization index (LI) for the observation of all object-related actions versus BASELINE for each UCP child, plotted against his/her clinical scores (MUUL scale on panel a; AHA scale on b). Children were represented with different colours and symbols according to the classification of their lesions (type I = red circles, type II = green squares, type III = blue triangles). Data were fitted with a standard linear function. Considering all the subjects, the correlations are not significant (Pearson's value R = −0.55, p = 0.06 for MUUL; R = −0.34, p = 0.28 for AHA; pink dotted line). They become significant when only children with lesions of type I and type II are used in the fit (R = −0.90, p = 0.0007 for MUUL; R = −0.68, p = 0.04, for AHA; cyan solid line), due to big variability of data from children with type III lesion.

Figure 6

Box plots of LI values obtained by different contrasts (AON task: all TASK>BASELINE, panel a; AON task: (C-ND + S-ND) > BASELINE, panel b; and sensory-motor task: (Sens-ND + Mot-ND) > REST, panel c) in UCP children, grouped according to TMS data (CL = contralesional reorganization, IL = ipsilesional). As in Figure 5, children were represented with different colours and symbols with respect to the classification of their lesions (type I = red circles, type II = green squares, type III = blue triangles). Grey dotted lines represent the threshold value of |0.20| for hemispheric lateralization.

Figure 7

Four-quadrants charts of LI values of pSMC in the sensory motor task (abscissae) and of the AON (ordinates) of the two hands: dominant hand (DH) on panel a, nondominant hand (non-DH) on b. The first and third quadrants represent the congruence of the sign (both positive in the first, I; both negative in the third, III), while the second and the fourth represent the discordance of the sign (pSMC-negative and AON-positive in the second, II; pSMC-positive and AON-negative in the fourth, IV). TD children were represented by grey diamonds, UCP children by black stars. In the chart for the nondominant hand, labels on data of UCP children are used to identify subjects, according to Table 1.