Introduction Dizziness in the presence of strong magnetic fields has been noticed ever since the first magnetic resonance experiments at high field strengths (>1 Tesla) have been conducted. It was suggested that this was due to the changing magnetic flux in the inner ear induced by the gross subject-motion within the magnetic field and that the motionless presence in the static magnetic field could not affect subjects significantly ( Schenck, 1992 ). Recently however, Roberts et al. (2011) showed that healthy subjects at rest who are kept in total darkness and exposed to the static magnetic field of an MR tomograph developed a persistent nystagmus, while patients with bilateral peripheral vestibular failure did not. Furthermore, the nystagmus’ slow phase velocity was modulated systematically depending on the subject’s head orientation relative to the field. The authors argued that the ionic fluids in the inner ear, which are constantly flowing as cells maintain resting activity, will be diverted by a (magnetic) Lorentz-force and this creates a pressure onto the cupula “the rotatory motion sensor” in the inner ear, thus leading to a nystagmus akin to a constant (accelerating) rotatory stimulation. This model was further supported by a simulation study ( Antunes et al., 2012 ) and a recent study of patients with unilateral labyrinthine disorders ( Ward et al., 2014 ). It was speculated that this magnetic vestibular stimulation (MVS) effect might influence fMRI results, because a nystagmus is indicative for a vestibular imbalance. Methods The aim of the current study was to investigate if this MVS effect does indeed modulate the fluctuations of the BOLD signal in areas related to the vestibular system. Here we treat the case of MVS influencing resting-state networks, presenting results from resting-state experiments conducted at two different field strengths (1.5 Tesla & 3 Tesla) and focus in particular on the influence of MVS on the default mode network. Results Those subparts of the default mode network that were significantly modulated between field strengths were also commonly associated with vestibular functions. Furthermore, the signal-scaling behavior of these subparts was different from the other parts of the default mode network that were not significantly modulated. In fact, the signal-scaling behavior was in agreement with a modulation due to MVS as expected from a Lorentz-model perspective (i.e. linear increase) while the other parts that were not significantly modulated by the MVS effect showed a scaling behavior that was expected simply from considering fMRI signal-scaling behavior due to increased field strength (i.e. approximately square root increase, e.g. see ( Duyn, 2012 )). Conclusion The MVS effect significantly affected BOLD signal modulation. Those parts which were more closely connected to the vestibular peripheral sense-organs did modulate in accordance with the Lorentz-model prediction and those subparts which modulated less, but still modulated significantly above the expected increase due to field-strength related signal-increase, were further removed in terms of connectivity from the vestibular peripheral sense-organs and therefore were modulated less strongly by MVS.
Read full abstract