Objective: Non-adhesive textile electrode belts offer several advantages over adhesive electrodes and are increasingly used in neonatal patients during continuous electrical impedance tomography (EIT) lung monitoring. However, non-adhesive belts may rotate in unsedated patients and discrepancies between chest circumference and belt sizes may result in a gap between electrodes near the sternum. This project aimed to determine the effects of belt rotation and sternal electrode gap on commonly used lung EIT parameters. Approach: We developed a simulation framework based on a 3D finite-element model and introduced lung regions with little or no ventilation that could be changed according to a decremental positive end-expiratory pressure (PEEP) trial. Four degrees of sternal gap and belt rotation were simulated and their effect on the EIT parameters silent spaces, centre of ventilation, global inhomogeneity index and overdistension/collapsed lung (OD/CL) analysed. Additionally, seven premature infants were examined to assess the influence of leftward and rightward belt rotations in a clinical setting. Main results: Small violations of the electrode equidistance criterion and rotations of the belts less than one electrode space exert only minor effects on the EIT parameters and do not impede the interpretation. Rotations of two and three electrode spaces induce non-negligible effects that might lead to flawed interpretations. The 'best PEEP' determined with the OD/CL approach was robust and identifiable with all studied sternal gaps and belt rotations. Significance: We revealed an important challenge for neonatal EIT applications related to a wide electrode gap at the sternum and belt rotation, which should be avoided in clinical application.