No observation of DIANA signals in rats at 7.0 and 17.2 Tesla

Abstract

Abstract Recently, a new method was introduced to detect neuronal activity using MRI. The method, referred to as DIANA, showed MRI signals with millisecond temporal resolution that correlated with local field potentials measured invasively in mice. Troublingly, attempts by other groups to detect the DIANA signals in humans at 7 Tesla and mice at 15.2 Tesla have failed. So far, attempts to reproduce DIANA in small rodents have focused on paradigms using whisker pad stimulation, which were expected to produce a 0.1-0.15% signal change. However, the supplementary material accompanying the original DIANA paper showed that visual stimulation produced a three times larger signal, which should be much easier to detect. Therefore, we attempted to find the DIANA signal in rats using a visual stimulation paradigm. Experiments were performed at 17.2 Tesla but also at 7.0 Tesla to see if the DIANA signal appears at a lower field strength where T2 is longer and BOLD contributions are reduced. In addition, simulations were performed to investigate the theoretical detectability of synthetic DIANA signals in noisy data. Although our data indicated that a 0.1% signal change would have been detectable, we did not observe a DIANA signal. We did observe neuronally driven hemodynamic signal variations that were much larger than the anticipated DIANA signal. The amplitude of these signal changes was relatively similar at 7.0 and 17.2 Tesla (0.7% vs 1.1%). Numerical simulations indicated, however, that the measured hemodynamic signal changes would not interfere with the detection of DIANA signals. Therefore, it is reasonable to expect that measurements at higher field strength with improved SNR would have a better chance to detect the DIANA signal. Yet, we, among others, were unable to find it.