Introduction

Cerebral palsy (CP) is a sensorimotor dysfunction caused by damage to the not yet developed brain. This leads to weakness and spasticity of the affected muscles (1). Time-resolved phase contrast (PC) MRI can be used to study muscle function and was applied to assess voluntary leg muscle contraction in adults, e.g. in (2,3). Children with CP cannot always freely control movement of their limbs, therefore voluntary exercise paradigms are often not adequate for diagnosis and examination. A cine PC MRI method to measure involuntary muscle contraction induced by electrical muscle stimulation (EMS) has been recently developed (4), however this MRI technique was never applied to children. Aim of this study is to check the feasibility of cine PC MRI of evoked muscle contraction in pediatric patients with spastic CP.

Methods

Two groups of children were included in the study: to date, 11 healthy controls (5 males, age range 9 to 16 years) and 4 CP patients with hemiparesis (3 male, age 11-12 years) were scanned. For the experimental setup, a two-channel commercially-available EMS device was used to induce involuntary periodic muscle contraction of the triceps surae muscle group of the dominant leg (in the controls) or of the most affected leg (in the patients) through self-adhesive electrodes attached to the muscle belly. The evoked force was recorded with a foot pedal device with a built-in force sensor during the scan (5). The maximum voluntary force (MVF) (without EMS) was also measured. Cine PC MRI was performed at 3T whole body clinical MRI scanner. A three-directional cine gradient echo PC velocity encoding sequence was applied with the following parameters: 3 parasagittal slices, spatial resolution 2.2x2.2x5.0 mm³, velocity encoding (VENC) 25 cm/s, temporal resolution 126 ms, TE=7.2 ms, bandwidth = 400 Hz/pixel, FOV= 280x140 mm², 3 k-space lines per segment, acquisition time 3 min, 27 temporal phases. The MRI acquisition was synchronized to the EMS cycle (1 s ramp up, 1 s plateau, 1 s ramp down, 1 s recovery). The stimulus consisted of 30 pulses/s (pulse width 300 µs). The amplitude of the stimulation current was in the range of 15 – 28 mA and was individually adjusted for each child within its comfort levels to a point that induced visible force output of at least 5% MVF if possible. To visualize the contraction of the muscle, a ROI was manually drawn on the calf muscles and the velocity vector field was displayed for every temporal phase (4). The contraction speed was calculated for each phase as median of the magnitude of the velocity vectors in the ROI. Strain rate and strain vectors were extracted from the velocity and displacement fields. Maps of the principal strain and strain rate were displayed.

Results

The maximum tolerated stimulation current and the developed force at this current showed a relatively large variability within both groups. In patients, the range of the average evoked force was 2 N – 53 N (2 – 21 % MVF). Consequently, in some children the data quality of the velocity fields was not sufficient for analysis of velocity and strain/strain rate. Data quality was sufficient in healthy children if developed force was at least 5% of the MVF. The time course of the contraction speed showed two distinctive peaks for the moments of contraction and release in 8 of the 11 healthy children and in 2 of the 4 patients. The peak contraction speeds were below 1 cm/s. Strain and strain rate maps showed activation in the whole ROI with higher localized activation in some children.

Discussion

In general, the EMS was well tolerated in both groups. Even the stiff legs of the CP patients could be stimulated to develop a periodic force over 3 min. The variability of the results can be attributed to the heterogeneity of our children population (age, physical development, compliance), to the different grades of paresis in patients, and to a lower tolerance of children for EMS, which prevented the use of higher currents. Therefore, the protocol needs to be further optimized for this population, for example by reducing the VENC and thus increasing the sensitivity to low velocities. Nevertheless, the extraction of quantitative parameters related to the muscle biomechanics was already feasible and they can provide useful insight in the assessment of this disease.

Conclusion

This study showed that synchronous dynamic PC MRI is feasible in electrically stimulated muscle in children, even in CP patients. This functional assessment may detect single muscles or muscle groups to be major factors for a given motor dysfunction. Work is in progress to compare the findings with physical/clinical data.

Acknowledgements

The study was supported by the Swiss National Science Foundation (Grant Number 173292).