Background: Perinatal stroke causes hemiparetic cerebral palsy and lifelong disability. Arterial (AIS) and venous (PVI) lesions damage motor pathways, but differences in timing and location dictate unique developmental plasticity trajectories that determine functional outcome. Emerging models are defining central therapeutic targets, but lack understanding of the integrated network that ultimately determines motor function. We demonstrate that resting state functional magnetic resonance imaging (rs-fMRI) can identify motor neural networks in children with perinatal stroke.

Methods: Ten children (6 males, mean 14.5yrs, 4 AIS, 6 PVI) were included from a population-based cohort with MRI-confirmed unilateral perinatal stroke and mild-to-severe motor disability (Melbourne Assessment: 74-100%; Assisting Hand Assessment: 52-100%). Participants underwent 3T MRI (GE750W) including anatomical imaging and six minutes rs-fMRI. The time-course of rs-fMRI signal in the primary motor cortex (M1) of the non-lesioned hemisphere was extracted and compared to all brain voxels using a General Linear Model. Descriptive analyses assessed relative connectivity between lesioned and non-lesioned M1 and supplementary motor area (SMA) and potential associations with motor function.

Results: Rs-fMRI motor networks from the non-lesioned M1 were obtainable in all subjects. In PVI children, non-lesioned M1 appeared more connected to contralateral, lesioned M1 than to SMA. The opposite trend was suggested for AIS (figure). Motor outcome appeared positively associated with relative connectivity to contralateral M1. Seeding from lesioned M1 demonstrated motor networks but was challenging in AIS.

Conclusions: Rs-fMRI can assess motor networks in children with perinatal stroke. Understanding developmental plasticity motor network patterns could further define central therapeutic targets.