Introduction
Obsessive–compulsive disorder (OCD) is a disabling, often chronic, psychiatric disorder with a lifetime prevalence of 1–2% (Halvorsen et al., Reference Halvorsen, Samuels, Wang, Greenberg, Fyer, McCracken and Goldstein2021). The most typical clinical manifestations of OCD include time-consuming obsessions (intrusive and unwanted thoughts, urges, or images that could cause anxiety or distress) and compulsions (the repetitive behaviors or mental acts, which patients feel driven to execute in response to the anxiety or distress) (Grant, Reference Grant2014). Current pathophysiological theories of OCD highlight the role of dysconnectivity in this disorder; that is, the wide spectrum of OCD symptoms arises from abnormal interactions among interconnected brain regions rather than from abnormality in isolated brain regions (Anticevic et al., Reference Anticevic, Hu, Zhang, Savic, Billingslea, Wasylink and Pittenger2014; Chen et al., Reference Chen, Ou, Lv, Yang, Li, Jia and Li2019; Gürsel et al., Reference Gürsel, Reinholz, Bremer, Schmitz-Koep, Franzmeier, Avram and Koch2020; Pagliaccio, Durham, Fitzgerald, & Marsh, Reference Pagliaccio, Durham, Fitzgerald and Marsh2021). In parallel, diffusion-tensor imaging (DTI) studies have revealed altered fractional anisotropy (FA) in numerous white matter (WM) fibers in patients with OCD, providing further structural evidence for the dysconnectivity in OCD (Koch, Reess, Rus, Zimmer, & Zaudig, Reference Koch, Reess, Rus, Zimmer and Zaudig2014; Li et al., Reference Li, Zhao, Huang, Guo, Long, Luo and Gong2021). Such network dysconnectivity likely arises in early development as polymorphism of an important regulator gene for the development of myelin-producing cells is associated with OCD (Stewart et al., Reference Stewart, Platko, Fagerness, Birns, Jenike, Smoller and Pauls2007). If this is the case, the early occurring WM abnormalities in OCD might have an impact on the cortical structure, especially the folding patterns.
Human cerebral cortex is extensively folded, allowing a larger surface area (hence a greater number of neurons) to be accommodated in the space-limited skull, improving efficient connectivity among neuronal units (Chklovskii, Schikorski, & Stevens, Reference Chklovskii, Schikorski and Stevens2002). While we do not know what the precise determinants of folding (or gyrification) are. During the early development of outward bulging gyri, the tension exerted by WM axonal fibers appears to play a crucial role (tension-based morphogenesis) (Essen, Reference Essen2020). As such, cortical gyrification has been thought of as a stable surrogate measure of the state of neural connectivity in early life. This notion is highly relevant to understanding neurodevelopmental disorders such as OCD where neural disruptions may precede symptoms by several years.
Many studies have reported aberrant gyrification in OCD with some increases (Fan et al., Reference Fan, Palaniyappan, Tan, Wang, Wang, Li and Liddle2013; Park et al., Reference Park, Ha, Kim, Lho, Moon, Kim and Kwon2022) and some decreases (Rus et al., Reference Rus, Reess, Wagner, Zaudig, Zimmer and Koch2017; Shim et al., Reference Shim, Jung, Choi, Jung, Jang, Park and Kwon2009; Wobrock et al., Reference Wobrock, Gruber, McIntosh, Kraft, Klinghardt, Scherk and Moorhead2010) as well; such mixed results are likely to arise if WM tension is weakened, leading to certain areas of hypergyria and other areas of hypogyria, as shown by experimental axonal lesions by Goldman-Rakic (Goldman-Rakic, Reference Goldman-Rakic1980). Considering the putative causal role of WM tension in the development of cortical gyrification (Van Essen, Reference Van Essen1997), concurrent examination of cortical gyrification and WM microstructural properties such as FA and fiber complexity is warranted to gain mechanistic insights into OCD. Nevertheless, to the best of our knowledge, no prior studies have been performed in this regard. Indeed, there are two main obstacles challenging the study of microstructure in WM development. One of them is the confounding effect of medications on certain WM deficits; serotonergic drugs in particular have notable effects on microstructure even after limited exposure, with a partial normalization of WM changes reported when treating OCD. The other is the limitation of commonly-used DTI metrics such as FA; they are calculated based on WM diffusive properties, which can be altered for various reasons that are not related to a developmental perturbation.
Director field analysis (DFA) is a newly-developed mathematical framework that could be applied to quantify the WM microstructural geometric properties (Cheng & Basser, Reference Cheng and Basser2017). This framework provides a general measure of orientational distortion and specific measures, which respectively capture three fundamental types of orientational distortions (i.e. splay, bend, and twist). Indeed, the fine-tuned local geometry of white matter fibers is of vital importance to brain maturation (Chang et al., Reference Chang, Owen, Pojman, Thieu, Bukshpun, Wakahiro and Mukherjee2015; Zhao et al., Reference Zhao, Shi, Dai, Wei, Zhang, Yu and Wang2021) and has been shown to be altered in neurodevelopmental disorders such as autism and schizophrenia (Kraguljac, Guerreri, Strickland, & Zhang, Reference Kraguljac, Guerreri, Strickland and Zhang2023). However, the alterations of orientational complexity in WM fibers remain unknown in OCD. Investigation of WM utilizing DFA might provide new important insights into the developmental dysconnectivity in patients with OCD.
In this study, we examined the cortical gyrification changes and their association with WM microstructural abnormalities in a large sample of medication-naïve patients with OCD (n = 64) in comparison to matched HCs (n = 61). We first conducted a vertex-wise comparison of LGI between the two groups. We then utilized FA and DFA metrics (splay, bend, twist, and total distortion) to examine both diffusive property and geometrical distortion of WM tracts in OCD. Third, we related cortical gyrification to WM microstructural abnormalities, expecting patients with significant disruptions in cortical gyrification to display robust WM microstructural abnormalities.
Materials and methods
Participants
Sixty-four medication-naive patients with OCD and 61 HCs were recruited from the Department of Psychiatry in the First Affiliated Hospital of Zhengzhou University. The clinical diagnosis of OCD was made by two experienced psychiatrists through a structured clinical interview provided in the Diagnostic and Statistical Manual of Mental Disorders (fifth edition) (DSM-5). Patients were assessed for their symptom severity using the Yale-Brown Obsessive–Compulsive Scale (Y-BOCS). Both patients with OCD and HCs were evaluated using the 14-item Hamilton Anxiety Scale (HAMA) and the 24-item Hamilton Depression Rating Scale (HAMD) to assess their anxiety and depressive symptoms. Exclusion criteria for patients included inability to tolerate scanning procedures, contraindication to scanning, current or past substance abuse or dependence, history of psychosis or bipolar disorder, current anxiety (individuals meeting DSM criteria and/or having a HAMA score >14), current depression (individuals meeting DSM criteria for Major Depressive Episode and/or have an HAMD score >20) and current neurological disorders in patients. Exclusion criteria for HCs included inability to tolerate scanning procedures, contraindication to scanning, current or past substance abuse or dependence, and history of neurological disorders. This study was approved by the Ethics Committee of the first affiliated hospital of Zhengzhou University, and written informed consent was obtained from all participants.
MRI data acquisition
All the MRI data were collected on a GE 3.0 T Discovery MR750 scanner. The T1-weighted MRI images were collected using a three-dimensional brain volume (BRAVO) imaging sequence with the following parameters: repetition time (TR) = 8.232 ms, echo time (TE) = 3.184 ms, inversion time = 450 ms, flip angle = 12°, matrix = 256 × 256, thickness = 1.0 mm, no gap, 188 slices, and voxel size = 1 × 1 × 1 mm3. The diffusion-weighted images were collected using a single-shot echo-planar imaging sequence with the following parameters: TR = 7100 ms, TE = 61 ms, 64 diffusion directions, b value = 1000 s/mm2, matrix = 128 × 128, field of view = 256 × 256 mm2, thickness = 2.0 mm, no gap, and 70 axial slices.
Cortical gyrification analysis
T1-weighted brain MRI images of each participant were processed using FreeSurfer (https://surfer.nmr.mgh.harvard.edu/) to generate the LGI map using Schaer's approach (Schaer et al., Reference Schaer, Cuadra, Schmansky, Fischl, Thiran and Eliez2012) used in many of our prior studies (Ajnakina et al., Reference Ajnakina, Das, Lally, Di Forti, Pariante, Marques, ，Mondelli and Dazzan2021; Papini et al., Reference Papini, Palaniyappan, Kroll, Froudist-Walsh, Murray and Nosarti2020). Specifically, the individual T1-weighted images were segmented to estimate the voxel-based gray/white matter boundary, which was triangulated to obtain a triangle-based gray/white matter boundary surface. This triangle-based gray/white matter surface was then topologically corrected to generate a refined gray/white matter surface (i.e. the white surface). The white surface was deformed outward with a deformable surface algorithm to generate the pial surface. These surfaces were visually checked and manually edited if necessary, according to established routines (Schaer et al., Reference Schaer, Cuadra, Schmansky, Fischl, Thiran and Eliez2012). Subsequently, the LGI map was obtained by calculating the ratio of the area of a circular region on the pial surface to the area of a corresponding circular region on the triangulated convex hull of the pial surface. Before entering into statistical analysis, the individual LGI map was resampled and smoothed using a heat kernel with a full-width-at-half-maximum (FWHM) of 15 mm.
Vertex-wise contrasts of the LGI maps between patients with OCD and HCs were performed using the SurfStat package (http://www.math.mcgill.ca/keith/surfstat/) in MATLAB. Specifically, for each vertex on the pial surface, we fitted a generalized linear model (GLM) with age and sex as the covariates of no interest. A vertex-wise p < 0.001 was used to define potential clusters of difference. Using random field theory (RFT), the resulting clusters were then corrected at the cluster level for multiple comparisons. The significance level for clusters was set at RFT-corrected p < 0.05. In addition, the global mean values of the significant clusters were extracted and correlated with clinical measures (including HAMD, HAMA, Y-BOCS, and duration of illness) in these patients using Pearson's correlation coefficient. The results of the correlation analyses were further corrected for multiple comparisons using the Bonferroni correction method, with the adjusted p value threshold of 0.0125 (i.e. 0.05/4).
DFA
Diffusion-weighted MRI images were preprocessed using the FMRIB's Software Library (FSL; https://fsl.fmrib.ox.ac.uk/fsl). Principal preprocessing procedures included eddy-current and head motion correction, and brain-tissue extraction. We used DMRITool (https://diffusionmritool.github.io) to reconstruct diffusion tensors and calculate FA and four DFA metrics: bend, splay, twist, and total distortion. Briefly, the bend index measures the spatial bending parallel to the fiber direction; the splay index measures the spatial bending perpendicular to the fiber direction; the twist index measures the extent to which neighboring directions are rotated with respect to one another, rather than aligned; the total distortion index is a global measure of geometric distortion of WM tracts, and it is defined as the square root of the sum of the squares of the bend, splay, and twist indices. The DFA metrics (bend, splay, twist, and total distortion) represent three different local geometric patterns of fibers formed by directions of tensors (i.e. eigenvectors), while traditional FA and mean diffusivity (MD) of tensors represent diffusivity properties formed by shapes of tensors (i.e. eigenvalues). Online Supplementary Fig. S1 in the supplementary materials shows four DFA metrics and FA, MD for the three synthetic tensor fields with splay, bend, and twist patterns, respectively. All tensors in these three tensor fields have the same FA and MD values, but their directions demonstrate splay, bend, and twist geometric patterns, which can be distinguished by splay, bend, and twist indices.
We extracted the WM tracts connected to the regions with atypical gyrification. Specifically, cortical regions with atypical gyrification were converted into 1-mm-thick volumetric masks below the white surface using ‘mri_label2vol’ command in Freesurfer. These masks were transformed to diffusion space and used as seed masks for probabilistic tractography. Fiber tracking was initiated from voxels of individual seed mask to the whole brain to generate 5000 streamline samples. Voxels with streamline count >10 were excluded, giving rise to a volume of WM tracts for each seed mask. For each participant, the volumes of WM tracts of the seed masks were transformed into Montreal Neurological Institute (MNI) space and merged into a single volume through a union operation. The resulting volumes were averaged across all participants to derive a population-based probability map of WM tracts. This population-based probability map was binarized with a threshold of p > 50% to yield a population-based mask of WM tracts connected to regions with atypical gyrification.
Voxelwise analyses of the FA and DFA metrics were conducted using the tract-based spatial statistics (TBSS) procedure in FSL. All the FA and DFA maps were aligned to the MNI space using FNIRT. A mean FA image was derived (from the aligned FA images) and thinned to generate a WM skeleton. The aligned FA and DFA maps were projected onto the WM skeleton and fed into subsequent voxel-wise statistics. With the introduction of the population-based WM mask described above, the voxel wise analyses were confined to the WM tracts connected to the regions with atypical gyrification. Specifically, we fitted a voxel-wise GLM with age and sex as the covariates of no interest. Group differences were examined using nonparametric permutation test with 5000 repetitions and subject to a correction for multiple comparisons using threshold-free cluster enhancement (TFCE). The level of significance was set at TFCE-corrected p < 0.05. In addition, the global mean values of the significant clusters were extracted and correlated with clinical measures (including HAMD, HAMA, Y-BOCS, and duration of illness) in these patients using Pearson's correlation coefficient. The results of the correlation analyses were further corrected for multiple comparisons using the Bonferroni correction method, with the adjusted p value threshold of 0.0125 (i.e. 0.05/4).
Results
Participant characteristics
We excluded two patients and two HCs due to LGI computation errors resulting from poor surface reconstruction that required notable manual intervention and poor reliability. Therefore, a total of 62 patients with OCD and 59 HCs were included in this study. The detailed demographic and clinical characteristics of these patients with OCD and HCs are shown in Table 1. There was no significant difference in age, sex, or years of education between the two groups. Patients with OCD had higher HAMA and HAMD scores than HCs.
Table 1. Demographic and clinical data of the participants
NA, not applicable; Y-BOCS, Yale-Brown Obsessive–Compulsive Scale; HAMD, Hamilton Depression Scale; HAMA, Hamilton Anxiety Scale.
Note: Data represent mean ± standard deviation. Between-group differences were examined using two-sample t test and χ2 test.
Increased LGI in OCD
Compared to HCs, patients with OCD had higher LGI in the left medial frontal cortex, paracentral gyrus (cluster size = 1952 vertices, peak Talaraich coordinates: x = −3.65, y = 21.90, z = 60.07, and peak t value = 3.5679), right medial frontal gyrus, dorsal anterior cingulate cortex (dACC) (cluster size = 7103 vertices, peak Talaraich coordinates: x = 5.00, y = 51.96, z = 41.39, and peak t value = 4.0249), and right precuneus (cluster size = 2283 vertices, peak Talaraich coordinates: x = 11.55, y = −45.00, z = 63.72, and peak t value = 3.6196) (Fig. 1). There were no regions of lower LGI in patients than HCs.
Figure 1. Brain regions with increased LGI in the patients with OCD compared with HCs. The result was corrected for multiple comparisons using RFT. The color bar indicates the RFT-corrected p value.
Decreased WM distortion in OCD
Compared with HCs, patients with OCD showed no significant FA alterations but significant distortion reduction in the WM tracts connected to regions with atypical gyrification. Specifically, patients showed a reduction in total distortion in the body, genu, and splenium of the CC, indicating orientational deficits that likely originated during development. (Fig. 2). As distortion is the geometric mean of bend, twist and splay, we parsed these features further and noted (1) decreased bend in nearly the whole callosum, left posterior corona radiata and right anterior corona radiata; (2); increased splay in the genu and the body of the callosum and decreased splay in left posterior corona radiata; (3) decreased twist in the genu and the body of the callosum in patients with OCD (online Supplementary Fig. S2 in the supplementary materials). In summary, this indicates a deviation in the microstructural properties of the callosal bundle in OCD.
Figure 2. WM tracts with decreased total distortion in medication-naïve patients with OCD compared with HCs. The result was corrected for multiple comparisons using TFCE.
Relationships between imaging and clinical data
Higher mean LGI of the affected cortical regions related to higher HAMA scores (r = 0.3561, p = 0.0052) (Fig. 3a), while the lower distortion of affected WM tracts related to higher HAMA scores (r = −0.3529, p = 0.0057) in patients with OCD (Fig. 3b). There was no significant correlation between imaging data and Y-BOCS scores in patients with OCD.
Figure 3. Relationships between imaging and clinical data in medication-naïve patients with OCD. The scatterplots show (a) the positive correlation between the mean LGI of the affected cortical regions and HAMA scores and (b) the negative correlation between the decreased total distortion and HAMA scores in medication-naïve patients with OCD.
Association between gyrification and WM distortion
Higher mean LGI of the affected cortical regions related to the reduced mean distortion of the affected WM tracts in patients with OCD (r = −0.5532, p < 0.001) (Fig. 4).
Figure 4. Relationship between gyrification and WM total distortion. The scatterplot shows the negative correlation between the decreased distortion and increased LGI in medication-naïve patients with OCD.
Discussion
Using a surface-based LGI and the newly developed approach of DFA, this study examined the alterations in cortical gyrification and WM microstructure in medication-naive patients with OCD. Compared with HCs, patients with OCD showed significantly increased LGI in bilateral medial frontal gyrus and the right precuneus, where the mean LGI was higher in patients with higher anxiety scores. Patients also showed significantly decreased total distortion in the body, genu, and splenium of the corpus callosum, where the reduced distortion was more pronounced in patients with higher anxiety scores. Moreover, patients with a higher mean LGI of the affected cortical regions also had lower mean distortion in the affected WM tracts. These observations indicate that developmental dysconnectivity-driven alterations in cortical folding are one of the neural substrates underlying the clinical manifestations in OCD.
In the present study, we found significantly increased LGI in medication-naive patients with OCD compared with HCs, which is consistent with two previous studies (Fan et al., Reference Fan, Palaniyappan, Tan, Wang, Wang, Li and Liddle2013; Park et al., Reference Park, Ha, Kim, Lho, Moon, Kim and Kwon2022). The neural mechanism underlying such hypergyrification is not well understood and may relate to neurodevelopmental irregularities such as excessive dendritic or terminal axon arborization, synaptic hyperplasia, and/or insufficient synaptic pruning. This concept is in agreement with a number of studies reporting significant hyperactivation in these regions in OCD (Ahmari & Rauch, Reference Ahmari and Rauch2022; Essen, Reference Essen1997; Melcher, Falkai, & Gruber, Reference Melcher, Falkai and Gruber2008). Notably, however, there were also studies showing significantly reduced LGI in medicated patients with OCD (Rus et al., Reference Rus, Reess, Wagner, Zaudig, Zimmer and Koch2017; Shim et al., Reference Shim, Jung, Choi, Jung, Jang, Park and Kwon2009) compared with HCs. Such discrepancy may reflect a prominent effect of antipsychotic medications on the gyrification pattern, in keeping with a previous report of relatively lower mean LGI in medicated patients compared with medication-naive patients (Fan et al., Reference Fan, Palaniyappan, Tan, Wang, Wang, Li and Liddle2013). Nonetheless, we cannot exclude the involvement of age differences among studies (Fan et al., Reference Fan, Palaniyappan, Tan, Wang, Wang, Li and Liddle2013; Rus et al., Reference Rus, Reess, Wagner, Zaudig, Zimmer and Koch2017; Shim et al., Reference Shim, Jung, Choi, Jung, Jang, Park and Kwon2009; Wobrock et al., Reference Wobrock, Gruber, McIntosh, Kraft, Klinghardt, Scherk and Moorhead2010) in causing the divergent cortical gyrification findings in OCD. Future longitudinal studies are warranted to further investigate this issue.
Compared to HCs, medication-naïve patients with OCD showed increased LGI in the dACC. Prior studies provided abundant evidence implicating the dACC in the pathogenesis of OCD. Specifically, electrical stimulation of dACC was found to cause compulsive goal-directed behaviors (Parvizi, Rangarajan, Shirer, Desai, & Greicius, Reference Parvizi, Rangarajan, Shirer, Desai and Greicius2013), while high-frequency transcranial stimulation of dACC was reported to alleviate OCD symptoms (Carmi et al., Reference Carmi, Alyagon, Barnea-Ygael, Zohar, Dar and Zangen2018). In addition, previous studies also reported hyperactivation and increased temporal variability of activity in dACC as well as a correlation between hyperactivation in dACC and symptom severity (Cavedini, Gorini, & Bellodi, Reference Cavedini, Gorini and Bellodi2006) in patients with OCD. The dACC plays important roles in conflict- and error-detection and performance monitoring (Fitzgerald et al., Reference Fitzgerald, Stern, Angstadt, Nicholson-Muth, Maynor, Welsh and Taylor2010; Gilbert, Barclay, Tillman, Barch, & Luby, Reference Gilbert, Barclay, Tillman, Barch and Luby2018). The increased LGI in the dACC may underlie the intrusive feeling of incompleteness (Summerfeldt, Reference Summerfeldt2004) and imperfection (Frost & Steketee, Reference Frost and Steketee1997) in OCD. Compared with HCs, patients with OCD also showed increased LGI in the left paracentral lobule which belongs to the supplementary motor complex (SMC) and the cognitive control network that plays a crucial role in the executive control of movement (Nachev, Kennard, & Husain, Reference Nachev, Kennard and Husain2008). Repetitive transcranial magnetic stimulation on the SMC could increase the resting motor threshold and improve the overall symptoms in OCD (Mantovani et al., Reference Mantovani, Lisanby, Pieraccini, Ulivelli, Castrogiovanni and Rossi2006). As such, it is plausible that the observed LGI increase in the left paracentral lobule relates to inhibitory or cognitive control deficits in this illness. In addition, the present study also showed significantly increased LGI in the medial prefrontal cortex (mPFC) and the precuneus in medication-naïve patients with OCD, which is consistent with previous studies reporting increased gyrification in patients with OCD (Fan et al., Reference Fan, Palaniyappan, Tan, Wang, Wang, Li and Liddle2013). The medial prefrontal gyrus and the precuneus are key hubs of the default mode network (DMN), which is activated at rest working on self-referential process and deactivated while performing attention-demanding tasks (Cavanna & Trimble, Reference Cavanna and Trimble2006). Insufficient deactivation of DMN was believed to reflect a failure to switch from self-referential processing to attention-demanding tasks in OCD (Gonçalves Ó et al., Reference Gonçalves Ó, Soares, Carvalho, Leite, Ganho-Ávila and Fernandes-Gonçalves2017). This finding indicates that the hypergyrification of mPFC and the precuneus may, to some extent, have contributed to the intrusive thoughts in OCD.
Using tractography techniques, we examined the microstructural characteristics of the tracts connected to regions associated with significant LGI increases. In these tracts, patients with OCD showed no significant FA change but significantly decreased total distortion in the body, genu, and splenium of the CC compared with HCs, indicating the presence of significant geometric alterations yet a lack of diffusional abnormalities in the CC in patients with OCD. As a general measure of WM microstructural geometry, the distortion quantifies three geometric characteristics (i.e. splay, twist, and bend) of WM fibers (Cheng & Basser, Reference Cheng and Basser2018). Notably, the decrease in distortion seems to be mainly driven by a lower bend index, which signifies more straight commissural fibers through the CC in patients with OCD. Such WM microstructural abnormalities may relate to increased tangential pressure and radial tension that occur as a result of neurodevelopmental abnormalities such as shortened commissural fibers (Gan et al., Reference Gan, Zhong, Fan, Liu, Niu, Cai and Zhu2017), local atrophy (Di Paola et al., Reference Di Paola, Luders, Rubino, Siracusano, Manfredi, Girardi and Spalletta2013; Lopez et al., Reference Lopez, Lalonde, Mattai, Wade, Clasen, Rapoport and Giedd2013; Piras et al., Reference Piras, Vecchio, Kurth, Piras, Banaj, Ciullo and Spalleta2021) and aberrant growing trajectory of the CC (Friedlander & Desrocher, Reference Friedlander and Desrocher2006; Rosenberg et al., Reference Rosenberg, Keshavan, Dick, Bagwell, Mac Master and Birmaher1997) that likely arise from factors such as prenatal hypoxia, perinatal trauma, among others. As increased LGI accompanies the decreased distortion in the tracts connected to these regions, we infer a pattern of consistency with the tension-based morphogenesis notion which implies the tension along WM tracts as the main diver of cortical folding. However, we were unable to explore the exact causal relationship between the two due to the cross-sectional nature of our study.
In patients with OCD, we observed a significant correlation between anxiety score and the LGI of the affected cortical regions. The functional connectivity within the affected regions (midline DMN nodes) relates to anxiety in healthy individuals (Coutinho et al., Reference Coutinho, Fernandesl, Soares, Maia, Gonçalves and Sampaio2016). In OCD, the presence of gyrification defects may suggest that this subnetwork is developmentally primed for pathological anxiety in OCD (Mogg & Bradley, Reference Mogg and Bradley2018). Meanwhile, we also found a negative correlation between the decreased distortion in the CC and anxiety score, which is partially consistent with a previous report of a negative correlation between the FA in the CC and HAMA score in patients with general anxiety disorder (Wang et al., Reference Wang, Peng, Wang, Wang, Li, Wang and Liu2019) Thus, a crucial but mild, likely developmental, aberration in interhemispheric communication reflected in reduced distortion metrics appears to play a role in the pathogenesis of anxiety in OCD. Nevertheless, we found no significant correlation of changes in LGI/DFA metrics with Y-BOCS, in keeping with a previous study on OCD (Rus et al., Reference Rus, Reess, Wagner, Zaudig, Zimmer and Koch2017). We concur that changes in gyrification and DFA metrics likely represent the underlying anxiety-related trait characteristics of OCD, whereas the expressed obsessive–compulsive symptom burden (state variable) has other distinct, likely functional neural correlates (Rus et al., Reference Rus, Reess, Wagner, Zaudig, Zimmer and Koch2017)
There are several limitations that should be acknowledged in this study. Firstly, the relatively small sample size restricted our ability to conduct subgroup analyses based on factors such as age of onset or symptom patterns. Secondly, the cross-sectional design of the study prevented us from identifying the developmental trajectory of cortical folding and WM connectivity throughout the progression of OCD. Thirdly, it is worth noting that statistical correlations alone cannot establish a direct causal relationship between WM geometric alterations and cortical gyrification in OCD.
In conclusion, this study revealed increased cortical gyrification, decreased WM distortion as well as a negative correlation between the two in medication-naive patients with OCD. Furthermore, these features were related to anxiety scores of these patients. Our results imply that developmental dysconnectivity-driven alterations in cortical folding are one of the neural substrates underlying the clinical manifestations of OCD, in particular the phenomenon of accompanying anxiety in this illness.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291723003422
Acknowledgements
The authors sincerely thank all patients and their family who participated in this study.
Funding statement
This work was partly funded by the National Natural Science Foundation of China (Grant No. 61971017) and Provincial and Ministry Co-construction Youth Funds of Henan Provincial Health Commission (Grant No. SBGJ202003048). In addition, L. Palaniyappan's work is supported by research support from the Canada First Research Excellence Fund, awarded to the Healthy Brains, Healthy Lives initiative at McGill University (through New Investigator Supplement to LP); Monique H. Bourgeois Chair in Developmental Disorders and Graham Boeckh Foundation (Douglas Research Centre, McGill University) and a salary award from the Fonds de recherche du Quebec-Sante (FRQS).
Competing interests
L. Palaniyappan reports personal fees for serving as chief editor from the Canadian Medical Association Journals, speaker/consultant fee from Janssen Canada and Otsuka Canada, SPMM Course Limited, UK, Canadian Psychiatric Association; book royalties from Oxford University Press; investigator-initiated educational grants from Janssen Canada, Sunovion and Otsuka Canada outside the submitted work.
