MR mapping of the effect of ultrasonic neuromodulation in the amygdala

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
The possibility to directly but non-invasively modulate neural activity would open up new horizons for treatment for example for anxiety disorders. This promise can be met by Transcranial Ultrasonic Stimulation (TUS) with the potential to module plasticity in deep brain structures non-invasively and induce transient changes outlasting the stimulation. The goal of this study was to develop MR protocols to enable effective and personalized mapping and guidance of TUS. Specifically, we advanced state-of-the-art MR skull imaging pipelines and quantified optimal parameters for accelerated resting-state functional MRI. We focused on the amygdala, a deep, small brain structure with a well-known functional and structural connectivity yet out of reach for conventional neuromodulation tools and susceptible to multiple MR artifacts. To quantify signal specificity, sensitivity, and fidelity of different protocols we used seed-based connectivity analyses, comparing number of echoes, multiband acceleration, voxel size, and susceptibility distortion correction methods. Specifically, across multiple tests, we found that multiband acceleration should be limited to a factor 4, with multiple echoes benefitting signal-to-noise ratio, where available smaller voxels outperforming more standard sizes, and spin-echo distortion correction to be preferred over alternatives. These investigations were targeted at a deep brain region, the amygdala, but our results show the critical principle of how the inferences that can be drawn are strongly dependent on the choice of protocol. In fact, well-established features of amygdaloid connectivity, such as the coupling with subgenual anterior cingulate regions, a critical biomarker of mental health, are entirely absent in some protocols. These findings highlight the importance of careful protocol selection in resting-state fMRI.
Faculteit der Sociale Wetenschappen