Page 68 - AN-4-3

Advanced Neurology Brain regions in olfactory dysfunction in PD

Given Iran’s diverse ethnicities, the goal was to identify the modules “Display slices” and “Check sample.”
odors familiar to all Iranians. A group of 90 students Both options are located in the CAT12 window
from various cultural backgrounds residing in Tehran under “Check Data Quality.” Furthermore, quality
dormitories was asked to review the 40 odors included parameters are estimated and saved in XML files for
in the UPSIT and identify those most familiar. To ensure each dataset during pre-processing. These quality
better linguistic comprehension, the original UPSIT was parameters are also printed on the report PDF page
translated into Farsi, and participants were asked to suggest and can be used in the module “Check sample.” Before
local odors commonly encountered in different regions of inputting the GM images into a statistical model, the
Iran. The development of ISIT involved the following steps: image data needs to be smoothed.
(i). Identifying and replacing unfamiliar odors. (ii) Statistical analysis: The smoothed GM images are
(ii). Compiling a preliminary list of 40 odorants, which input into a statistical model. This requires building
included both natural and synthetic options, while also a statistical model (e.g., t-tests, analysis of variance
producing fragrance microcapsules when necessary. (ANOVAs), and multiple regressions). This is done
(iii). Designing scratch-and-sniff stickers by mixing by the standard SPM modules “Specify 2 Level”
nd
microcapsules with varnish ink and printing them or preferably “Basic Models” in the CAT12 window,
using a silk screen printer on sticker paper. covering the same function but providing additional
A pilot study was then conducted with 43 participants options and a simpler interface optimized for
(23 females and 20 males, aged 20 – 40) using this structural data. The statistical model is estimated. This
initial version of ISIT. The pilot study aimed to identify is done with the standard SPM module “Estimate”
any deficiencies in the procedure and to select the (except for surface-based data, where the function
most appropriate odors among the 40 items and their “Estimate Surface Models” should be used instead). If
alternatives. 24 total intracranial volume (TIV) is used as a confound
in a model to correct for different brain sizes, it is
2.3. VBM pre-processing necessary to check whether TIV reveals a considerable
Preprocessing analysis for VBM was performed using correlation with any other parameter of interest, and
the CAT12 toolbox on high-resolution T1-weighted rather uses global scaling as an alternative approach.
25
structural images, acquired from all patients on a 1.5 After estimating the statistical model, contrasts are
Tesla MRI scanner (Avanto, Siemens Healthineers, defined to get the results of the analysis. This is done
Germany), within the Statistical Parametric Mapping with the standard SPM module “Results”. 25,26
(SPM12) framework developed by the Department of 2.4. MRI data acquisition
Imaging Neuroscience Group (http://www.fil.ion.ucl.
ac.uk/spm). This analysis was conducted using MATLAB All patients underwent MRI scans using a 1.5T MRI
R2023b software (CAT, Structural Brain Mapping Group, scanner. The choice of a 1.5T scanner is based on its
Germany). Initially, the anatomical images were segmented widespread availability and effectiveness in clinical settings
into GM, WM, and cerebrospinal fluid using the unified for a variety of neurological assessments. This model has
segmentation module. After segmentation, the GM been well-studied and is known for providing high-quality
26
images were normalized to the Montreal Neurological images while maintaining patient comfort.
Institute (MNI) standard space using the diffeomorphic The scans were performed utilizing the HE1_4 coil
anatomical registration through exponentiated lie algebra element, specifically designed for this scanner. Coil
(DARTEL) algorithm. Following the affine and non- elements play a crucial role in MRI examinations, as they
27
linear registration of the GM templates in MNI space, the determine the sensitivity of the MRI equipment to the
images were modulated to preserve the relative GMVs magnetic fields produced by the scanned tissues. The HE1_4
post-spatial normalization. The resulting GM images were coil element is optimized for head imaging, resulting in an
then smoothed with a Gaussian kernel featuring a full improved signal-to-noise ratio and enhancing the clarity
width at half maximum of 10 mm. In summary, a VBM and resolution of the images captured. This optimization is
analysis comprises the following steps:
(i) Pre-processing: T1 images are normalized to a essential for accurately visualizing fine anatomical details
template space and segmented into GM, WM, and and potential pathological changes in brain structures.
cerebrospinal fluid. The pre-processing parameters To ensure patient comfort and minimize movement
can be adjusted through the module “Segment Data.” during the scanning process, all participants were
After the pre-processing is finished, a quality check is instructed to lie supine on the MRI table. This position
highly recommended. This can be achieved through helps in achieving a stable and reproducible imaging setup,

Volume 4 Issue 3 (2025) 62 doi: 10.36922/AN025110024