Spatial standardization is a key part of many structural and functional neuroimaging studies. To accurately reproduce the target point locations within and between treatment centers and define them by coordinates, as well as to allow for standardization in mapping, different systems to standardize the size and structures of the brain have been introduced. The general process comprises registering multiple brains to one template and applying nonlinear deformation algorithms to the images (Klein et al., 2009).

In 2001, after several iterations of different brain atlases, the MNI-152 coordinate system was developed by the Montreal Neurological Institute (Mazziotta et al., 1995, 2001a,b). This template used high-resolution T1-weighted scans from 152 young adults (86 males and 66 females), gathered across three sites (MNI, UCLA, UTHSCSA). MNI-152 became the standard template due to its complete coverage of the brain and increased resolution (Collins et al., 1994; Evans et al., 2012). The MNI-152 atlas space is commonly used, especially for stereotactic procedures, and an increasing number of neurosurgery studies report results in MNI space (Ewert et al., 2018; Horn & Blankenburg, 2016; Schӧnecker et al., 2009).

Though templates can be quite useful for group-level comparisons, there are several drawbacks. Previous studies analyzing the effects of transformation into standard space have found that it affects structural and functional analysis results and have suggested that caution be used when applying standard space regions of interest to volumetric functional magnetic resonance imaging (fMRI) data (Hutchison et al., 2014). Moreover, analysis in the subject's real, or ‘native,’ space is much more accurate due to the partial volume effects introduced by analysis of magnetic resonance imaging (MRI) scans in standard space (Aribisala et al., 2011). It is also known that the MNI brain is significantly larger than the average brain (Allen et al., 2002). MNI brains are known to be higher, deeper, and longer, with a maximum in the order of 10mm, especially at points that are further from the center of the brain (Brett, 1999). Most notably, normalizing MRI data in MNI space found distortions in the position of the reconstructed lead used for deep brain stimulation (DBS) patients in the subthalamic nucleus, which is one of the areas encompassed in the subcortical structures we examined (Nowacki et al., 2018).

Important to note is that during spatial normalization of data with lesions, cost function masking can be used to exclude the lesion location from computation to prevent substantial errors in the resulting scan (Brett et al., 2001). However, if the effects of a lesion's necrosis extend into nearby areas, distortion could be introduced, and these areas could be improperly normalized. A more recent report also examined the necessity of cost function masking in the normalization of scans with focal lesions and found that unmasked normalization resulted in a significant underestimation of lesion volume, especially in patients who had larger lesions, implying that cost function masking is still imperative when normalizing scans with lesions (Andersen et al., 2010).

When data from neurosurgical procedures that utilize templates are analyzed, results can guide targeting and treatments for patients. For this study, MRI scans of patients were examined after treatment with gamma ventral capsulotomy for obsessive-compulsive disorder (OCD). OCD has a lifetime prevalence of about 1.9-2.5% worldwide, and has a serious impact on employment, marital satisfaction, and quality of life (Matsunaga & Seedat, 2007; Stein et al., 2019). Characterized by recurring thoughts or sensations known as obsessions, and the drive to do certain things repetitively, also known as compulsions, OCD can be a very disabling, and sometimes unmanageable mental disorder (Stein et al., 2019). A select number of OCD patients with severe symptoms who do not improve after conventional treatments may choose to have neurosurgery (Miguel et al., 2019). The most updated procedure for treatment of refractory OCD is ventral capsulotomy, in which bilateral lesions are produced in the internal capsule. The internal capsule, a white matter structure, is connected to both the thalamus and prefrontal cortex. Creating lesions in these fiber tracts disrupts connections between the frontal and cingulate cortex and subcortical nuclei (Lv et al., 2021). Following a study conducted with severely ill patients with OCD who received bilateral lesions in the ventral portion of the anterior limb of the internal capsule (ALIC) over a period of 20 years, gamma ventral capsulotomy was found to be effective for several patients (Rasmussen et al., 2018). Notably, 56% of patients had an improvement in their Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) score of greater than 35% during the 3 years after the procedure. Moreover, patients showed improvements in depression, anxiety, quality of life, and global functioning.

Past techniques have utilized standard space MRI scans to determine typical lesion size and location in relation to response rate, without adequate knowledge about how these measurements might have changed. As noted, consequences arise when using group-level data to analyze results without considering the potential error related to the imaging transformation to standard space. The importance of correctly targeting the internal capsule for lesions to ameliorate severe OCD should not be understated, given that slight changes in location can influence a patient's response (Lippitz et al., 1999). If significant distortions are introduced, studies involving MNI-based results stand to be invalidated. This invalidation could then impact the accuracy of results illustrating which brain regions are the most effective targets for surgery and other less invasive interventions. Neurosurgical procedures ranging from DBS to cingulotomy, as well as other ablative surgeries that adjust target location based on studies could benefit from these results.

The objective of this study was to analyze lesion size and location in individuals who have undergone capsulotomy for intractable OCD, comparing deviations in the native scans to those in standard space. Our hypothesis was that there would be a significant difference between the size of structures in native and standardized scans, given the varying effects that MNI introduces. To assess this, native and MNI scans from patients who underwent gamma ventral capsulotomy were compared using measurements in the coronal and axial planes.