Simultaneous EEG data with passive motor stimulus from preterm infants

Preprocessed EEG data - infants were fitted with neonatal 25-electrode EEG caps (29 - 32 cm head circumference) and connected to an MR-compatible system (EASYCAP and Brain Products GmbH). EEG data preprocessing was performed using Analyzer 2 software (Brain Products GmbH), with an initial 0.2 Hz high-pass filter used to remove slow frequency drift in the EEG data. After exclusion of TRs with visible motion on the raw EEG, MR gradient artifact was cleaned using a 25 TR sliding window template subtraction. A 40 Hz lowpass and 50 Hz notch filter were applied. Electrodes with poor signal or bridged to the reference (FCz) were removed. MRI data were acquired with a 3 Tesla Philips Achieva scanner (Best, Netherlands) and a 32 channel adult head coil. Infants were scanned following feeding, during natural sleep and were fitted with ear protection (molded dental putty and adhesive earmuffs: Minimuffs, Natus Medical Inc, San Carlos CA, USA) and immobilized in a vacuum cushion (Med-Vac, CFI Medical Solutions, Fenton, MI, USA). fMRI data were acquired using T2*-weighted single-shot gradient echo echo-planar imaging (GRE-EPI) sequence (resolution: 2.5*2.5*3.25mm; 21 slices; TE: 30ms; TR: 1500ms, flip angle: 90°, lasting up to 13.5 minutes). A custom-built MR compatible robotic device (Dall'Orso et al., 2018) was fitted to the right wrist to deliver blocks of 1Hz passive right wrist flexion-extension for 7.5 to 10.5 seconds (5 to 7 TRs, up to 24 epochs of stimulation), with a variable inter-stimulus interval (21 to 24 seconds) to minimize anticipatory responses.