The periaqueductal gray (PAG) is a brainstem area that is assumed to serve as a relay station in the central nervous system pathways involved in control of micturition. The PAG is proposed to be responsible for projecting afferent information from the bladder to cortical and subcortical brain areas and acts as a gatekeeper projecting efferent information from cortical and subcortical areas to the pons and spinal cord [1]. The role of the PAG in lower urinary tract symptoms (LUTS) is not well understood, but a better understanding could provide a framework for new therapies or diagnostic procedures. Studies using animal models have shown that the PAG is organized in a columnar fashion. Technological advances during recent years have made ultra high-field magnetic resonance imaging (MRI) scanners more readily available, and the use of these scanners will allow for the investigation of PAG activity in sub-millimeter detail in human participants. Additionally, the advancement of MRI analysis methods allows for more data-driven analyses and systematical approaches.
The current study aims to use 7 Tesla functional magnetic resonance imaging (fMRI) to parcellate the PAG into clusters of voxels with a higher within-cluster connectivity than between-cluster connectivity in, both, an empty and a full bladder state. We will assess the similarity between parcellation results of both bladder states and expect a significant spatial overlap between the cluster constellations found in both states. A significantly higher agreement between parcellations based on fMRI data obtained during empty and full bladder states indicates that parcellation of the PAG is independent of the level of bladder fullness and experienced bladder sensations. This is necessary to qualify this novel methodological approach to be utilized for studying  PAG activity related to bladder functioning.

This study was approved by the local ethical committee, and informed consent was obtained from of our participants. We evaluated 7 Tesla fMRI parcellations for 6 female participants (mean age: 50). We acquired fMRI data during an empty bladder state. We then filled our participants’ bladder at a rate of 30ml/min until they indicated they experienced a strong desire to void using a joystick they could control from the scanner. Once the participant indicated a strong desire to void we started a second fMRI scan.  
Functional data were preprocessed using BrainVoyager 21. Using a mask drawn on the anatomical images for each individual subject we selected all voxels within the PAG, on which we computed a voxel-by-voxel correlation matrix based on the functional time course of each voxel during the full bladder scan. We parcellated this matrix using an implementation of the Louvain module detection algorithm in MATLAB (Fig. 1.A), which outputs clusters with stronger within-cluster connectivity than between-cluster connectivity [2]. Using the empty bladder parcellation we computed random parcellations using a script that randomly places an anchor point in the 3D space of the PAG mask and grows modules, similar in size to the original modules, from this anchor point. We iterated this script to obtain 1000 different random parcellations. These random parcellations were subsequently used to assess the agreement between our parcellations based on fMRI data acquired during an empty and full bladder state compared to what could be expected based on chance.

The Sørensen–Dice coefficient is a useful measure of spatial overlap which can be applied to studies of reproducibility and accuracy in image segmentation, it is based on the percentage of spatial overlap between two images  [3]. For each participant, we computed the Sørensen–Dice coefficient between each cluster in the empty bladder parcellation and each cluster in the full bladder parcellation. Next, we computed the Sørensen–Dice coefficient between each cluster from the empty and full bladder parcellations and each cluster from the 1000 iterations of the random parcellation. This allowed us to statistically test the extent to which the similarity between empty and full bladder parcellations was higher than could be expected based on chance on a single subject level. For each of our participants we found that the agreement between at least one of the clusters in both states resulting from the parcellation procedure was higher than could be expected based on chance (p = <0.05, corrected for multiple comparisons using a false discovery rate (q=0.05)) (table 1). For two of our participants we managed to find a counterpart for each module in both parcellations that had a significantly higher agreement than could be expected on chance.

Our results indicate that PAG activity in an empty and full bladder state can reliably be subdivided in clusters that are largely independent on bladder state, and show a higher similarity between each other than could be expected based on chance (figure 1). Thereby supporting the idea that fMRI based parcellation of the PAG allows for replication of neuro-anatomical findings based on animal work. These results imply that PAG parcellations can be used to study changes in dynamic connectivity between PAG clusters related to bladder fullness and paticipants’ experienced bladder sensations. These connectivity changes can serve to create dynamic response profiles which, when established in healthy controls, could serve as a benchmark to which LUTS patients’ response profiles can be compared. This opens new possibilities to investigate the effects of treatments of LUTS on signal processing in the PAG. It will also provide a new and necessary framework for translational research into new therapeutic targets, treatment effectivity predictors, and diagnostic measurements.

The human PAG can reliably be parcellated into functional modules that are largely independent of bladder state using ultra high-field fMRI. This method of functionally subdividing the PAG can be used to assess connectivity changes between PAG clusters in relation to bladder sensations. This will enable us to take an interdisciplinary and translational approach to find new and more optimally targeted treatment and diagnostic options for LUTS patients.

Fowler, C.J., D. Griffiths, and W.C. de Groat, The neural control of micturition. Nat Rev Neurosci, 2008. 9(6): p. 453-66.
Fortunato, S. (2010). Community detection in graphs. Physics reports, 486(3), 75-174.
Zou, K. H., Warfield, S. K., Bharatha, A., Tempany, C. M., Kaus, M. R., Haker, S. J., ... & Kikinis, R. (2004). Statistical validation of image segmentation quality based on a spatial overlap index 1: Scientific reports. Academic radiology, 11(2), 178-189.