2.1 Participants
The dataset for this study was amassed from 507 patients, who were enlisted via outpatient services at Shanghai Children's Medical Center (SCMC) and through public advertisements. The Institutional Review Board (IRB) of SCMC, affiliated with the School of Medicine, Shanghai Jiao Tong University, provided ethical approval for our study (Approval No. SCMC-201014). This research adhered to the principles outlined in the Declaration of Helsinki, as well as the ethical standards set forth by the IRB. Recruitment spanned from July 2018 to November 2022, during which time consent was obtained in writing from all participants. Eligibility for participation in our study was contingent upon individuals having data from resting-state functional magnetic resonance imaging (rs-fMRI), T1-weighted structural MRI, and assessments of MDA function levels. Additionally, the intellectual capacities of all participants were assessed by trained professional developmental and behavioral pediatricians utilizing the Wechsler Intelligence Scale for Children-Revised in China (WISC-RC) [19]. This scale, which evaluates verbal IQ (VIQ), performance IQ (PIQ), and full scale IQ (FIQ), is tailored to the developmental standards of school-aged children in China. Recognizing the significant influence of IQ on brain functional connectivity, the administration of IQ tests to all children participating in our study was designed to control for the potential confounding effects of this variable on our research outcomes. Each candidate was required to undergo an extensive clinical evaluation, covering physical, psychiatric, and neurological assessments. These evaluations were to be conducted by a minimum of 3 experienced developmental and behavioral pediatricians.
Inclusion Criteria: This study targeted individuals aged between 5 to 16 years, with a stipulation for right-handedness. A minimum FIQ score of 80 was essential for inclusion. Participants must have no recorded medical history of conditions that are associated with the onset of NE, such as urinary tract infections or diabetes mellitus, among others. Additionally, individuals should not have received any form of enuresis treatment prior to their MRI evaluations.
Exclusion Criteria: The study excluded individuals with psychiatric comorbidities or neurological disorders, including, but not limited to, attention-deficit/hyperactivity disorder (ADHD), tic disorders, intellectual disabilities, and cerebral palsy. Candidates exhibiting daytime lower urinary tract symptoms, those with contraindications to MRI, significant head movement during scans, or a history of current or previous antipsychotic medication use were also not eligible for participation.
The diagnosis of PNE was conducted through a comprehensive evaluation by at least 3 senior developmental and behavioral pediatricians. This assessment included a detailed medical history review, physical examinations, and was supported by urinalysis, renal, ureteral, and bladder ultrasonography, and developmental, psychological, and behavioral testing. When necessary, lumbar-sacral MRI scans were performed to exclude potential underlying causes of secondary enuresis. The diagnostic process adhered strictly to the ICCS standards.
Furthermore, all participants designated as healthy children (HC) had not experienced any instances of bedwetting and, following the age of approximately 3 years, regularly demonstrated the ability to awaken due to the urge to urinate and then get out of bed to void, aligning with the developmental milestones observed in the vast majority of children. We requested that parents closely monitor and report on their child's ability to independently awaken and urinate during the night for a duration of at least one month. Those children who frequently exhibited the capability for independent nocturnal awakening to urinate were classified as HC. This classification indicates the existence of a robust and well-developed MDA function in HC. After a comprehensive screening process to exclude certain samples, the study ultimately encompassed 133 patients diagnosed with PNE and 40 HC possessing intact MDA functionality (Fig. 1).
2.2 Auxiliary diagnostics and behavioral assessments
The urinalyses of the participants demonstrated normative results, devoid of glucosuria and leukocytes, indicating an absence of immediate metabolic or infectious concerns. Comprehensive ultrasonographic evaluations of the renal and urinary systems revealed no organic abnormalities within the kidneys, ureters, or bladder. Furthermore, in individuals diagnosed with PNE, bladder volumes were assessed via ultrasonography at the juncture of pronounced micturition urgency.
Detailed clinical data encompassing age, gender, duration of illness, and bed-wetting frequency were meticulously documented. Significantly, we developed a grading system to evaluate the strength of MDA function. This system was established based on parental observations and reports of children's nocturnal urination behavior spanning a period of at least one month. The grades are defined as follows: grade 1 = always wakes up before bed-wetting occurs; grade 2 = often wakes up after a minor amount of urination; grade 3 = often wakes up after a significant amount of urination; grade 4 = often wakes up after completely emptying the bladder; and grade 5 = frequently fails to wake up even after the bladder is emptied. This grading system indicates that a higher MDA score correlates with diminished MDA function. This scale was formulated based on our detailed clinical observations and the enhancement of methodologies from a preceding study [20].
2.3 Imaging acquisition and preprocessing
The MRI data collection was conducted at the Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, Shanghai, China, utilizing a Siemens 3.0 T MRI system (Prisma). The parameters for the rs-fMRI sequence were established as follows: participants were instructed to keep their eyes closed; repetition time was set at 2000 ms, echo time at 30 ms, in-plane resolution at 3.5 × 3.5 mm2, slice thickness at 3.5 mm, flip angle at 90°, with a total of 240 volumes, leading to an acquisition time of 486 s. For T1-weighted images, crucial for segmentation and normalization processes, the sequence parameters were defined as: repetition time at 2530 ms, echo time at 2.98 ms, in-plane resolution at 1 × 1 mm2, slice thickness at 1 mm, and flip angle at 7°.
The preprocessing of neuroimaging data is detailed within the Supplementary Materials section. Notably, a total of 29 participants with PNE and 4 HC were excluded due to excessive head movement during imaging (specifically, translations exceeding 3 mm, rotations greater than 3°, or a mean framewise displacement above 0.2 mm), as depicted in Fig. 1.
2.4 Definition of regions of interest within the PAG
As delineated in preceding sections, the PAG encompasses four principal subdivisions: the vlPAG, lPAG, dlPAG, and dmPAG [13]. Leveraging the corpus of extant literature, the PAG's subdivisions have been methodically identified and classified into seven specific regions, hereinafter referred to as 'seeds' [18, 21]. The precise spatial coordinates and anatomical locations of these seeds are cataloged in Table 1 and visually represented in Figure S1 of the Supplementary Materials.
Table 1. The coordinates of PAG subregions.
Abbreviations: PAG = Periaqueductal gray; MNI = Montreal Neurological Institute.
2.5 Analysis of functional connectivity
Utilizing voxel-wise correlation coefficients, we generated correlation maps predicated on the aggregated time series data derived from each identified seed within the PAG subdivisions. In an endeavor to enhance the statistical normality of our findings, these correlation maps underwent a transformation into z maps through the application of Fisher's r-to-z conversion. This methodological approach was executed utilizing the RESTplus 1.27 software [22], ensuring precision and reliability in our analysis.
2.6 Statistical analysis
The demographic and clinical characteristics of all participants, including key variables such as age, sex, and MDA functionality levels, were characterized through the use of descriptive statistics, facilitated by the SPSS 26.0 software suite. To investigate the potential linkage between the nuanced connectivity of the PAG subregions and cortical areas with respect to MDA functionality levels, we embarked on a thorough whole-brain correlation analysis. This analysis entailed a comparative assessment of MDA functionality levels against z maps, while adjusting for age and sex as covariates to bolster the methodological integrity of our examination.
In confronting the pivotal challenge posed by multiple comparisons, our approach leveraged the AlphaSim correction technique, adopting a rigorous significance threshold of P < 0.001 and specifying a minimum cluster size of 351 mm3 (equivalent to 13 voxels, each with dimensions of 3 × 3 × 3 mm3). This procedural rigor facilitated the attainment of a corrected significance level of P < 0.05. These methodological strategies were implemented via the RESTplus software.