Background Functional magnetic resonance imaging (fMRI) in preterm infants has been introduced as a non-invasive method to study information processing in the developing brain and to identify early signs of dysfunction. Studying the sensorimotor system, fMRI has so far delivered inconsistent brain activation patterns regarding laterality and direction of BOLD signal changes.

Aims To study evoked responses to unilateral passive sensorimotor stimulation in preterm infants using a customised neonatal head coil.

Patients/methods 14 preterm infants born less 30 weeks gestation were scanned using a block design (10 × 30 sec ON-OFF) at 3 T (Achieva, Philips, Best, NL) using a customised neonatal head coil (J. Nordmeyer-Massner and K. Pruessmann, ETH Zürich) at term equivalent age. Data were pre-processed (slice-time-correction, motion correction, anatomical co-registration with high-resolution T2-weigthed structural images, 6mm³ smoothing) and analysed on an individual level using SPM8. Only activation clusters surviving p < 0.05 (FWE-corrected) were considered as statistically significant.

Results 5/14 scans had to be withdrawn due to movement artefacts (n = 4) or technical failure (n = 1). In the remaining sample of 9/14 subjects, unilateral passive sensorimotor stimulation elicited primarily positive BOLD responses, which were located in contralateral (8/9) and ipsilateral (1/9) primary sensori-motor cortex (SMC).

Conclusions The study results indicate that improved sensitivity of a size-optimised neonatal head coil is crucial for detection of primary SMC activity on an individual level and question the hypothesis of unfocused negative and bilateral BOLD responses in the premature brain.