Abnormalities in the white matter tracts in patients with Parkinson disease and psychosis

Abstract

Objective

The objective of the current study was to compare the microstructural integrity of the white matter (WM) tracts in patients having Parkinson disease (PD) with and without psychosis (PD-P and PD-NP) through diffusion tensor imaging (DTI).

Methods

This cross-sectional study involved 48 PD-NP and 42 PD-P who were matched for age, sex, and education. Tract-based spatial statistics (TBSS) was used to compare several DTI metrics from the diffusion-weighted MRIs obtained through a 3-Tesla scanner. A set of neuropsychological tests was used for the cognitive evaluation of all patients.

Results

The severity and stage of PD were not statistically different between the groups. The PD-P group performed poorly in all the neuropsychological domains compared with the PD-NP group. TBSS analysis revealed widespread patterns of abnormality in the fractional anisotropy (FA) in the PD-P group, which also correlated with some of the cognitive scores. These tracts include inferior longitudinal fasciculus, inferior fronto-occipital fasciculus, right parieto-occipital WM, body of the corpus callosum, and corticospinal tract.

Conclusion

This study provides novel insights into the putative role of WM tract abnormalities in the pathogenesis of PD-P by demonstrating significant alterations in several WM tracts. Additional longitudinal studies are warranted to confirm the findings of our research.

Patients with Parkinson disease (PD) develop an array of nonmotor symptoms.¹ Psychosis is one such symptom, which commonly manifests through formed visual hallucinations (VHs) and/or minor hallucinations (MHs) that encompass a false sense of presence or passage and illusions.^2,3 Although several risk factors for the emergence of psychosis in PD have been identified,^4–6 the underlying pathogenesis remains unclear. As psychosis is associated with increased health care utilization⁷ and mortality in PD,⁸ it is essential to gain deeper insights into the neural correlates of PD-associated psychosis.

Several volumetry-based studies on PD-associated psychosis have demonstrated atrophy of the cholinergic brain structures,⁹ visuoperceptive regions,¹⁰ and hippocampus.¹¹ However, the literature on white matter (WM) involvement is sparse. Two studies demonstrated abnormal WM integrity in patients having PD with psychosis (PD-P).^12,13 Lee et al.¹² have reported microstructural abnormalities in the left optic nerve and optic radiations; however, Yao et al.¹³ illustrated alterations in the posterior hippocampal WM in patients having PD with VH. However, these studies have focused only on specific tracts¹² or structures¹³ and have been performed on small sample size.

Moreover, several functional imaging studies have implicated multiple neural networks in the pathogenesis of psychosis,¹⁴ which include the default mode network and dorsal attention network.¹⁵ It is therefore probable that psychosis is associated with widespread connectivity abnormalities. Hence, to explore the underlying pathogenesis of psychosis in PD, our study based on diffusion tensor imaging (DTI) compares the whole-brain WM tract abnormalities of PD-P and PD-NP using tract-based spatial statistics (TBSS).

Methods

Subject recruitment and clinical evaluation

This study included 48 patients having PD without psychosis (PD-NP) and 42 patients having PD with VH or MH (PD-P). The patients were recruited from the general neurology outpatient clinics and movement disorders services of the National Institute of Mental Health and Neurosciences (NIMHANS), Bangalore, India. Diagnosis of PD was based as per the United Kingdom Parkinson’s Disease Society Brain Bank Criteria. The Unified Parkinson’s Disease Rating Scale (UPDRS-III) was used to assess the severity of motor symptoms, whereas the modified Hoehn and Yahr (H&Y) system was used to assess the stage of PD. Semistructured interviews (by authors A.L. and S.S.A.) were conducted to explore the presence of psychiatric symptoms, and the diagnosis of PD-P was based on the criteria proposed by the National Institutes of Neurological Disorders and Stroke and National Institute of Mental Health.¹⁶ As per these criteria, continuous presence (at least for 1 month) or recurrent emergence of illusions or a false sense of presence or hallucinations or delusions are considered as the characteristic symptoms of PD-P. For the diagnosis of PD-P, any of the aforementioned symptoms must have occurred after the diagnosis of PD, and other diseases that could contribute toward the psychotic symptom (dementia with Lewy bodies, psychiatric disorders such as schizophrenia, schizoaffective disorder, delusional disorder, or mood disorder with psychotic features, or a general medical condition including delirium) should be excluded.¹⁶

The severity of the psychosis was assessed using the scale proposed by Ondo et al.¹⁷ Details of the antiparkinsonian medications were obtained from all the patients, and the total levodopa equivalent dose per day (LEDD) was calculated. Anxiety and depression were assessed by the Hamilton Anxiety Rating Scale (HAM-A) and Hamilton Depression Rating Scale, respectively.

Standard protocol approvals, registrations, and patient consents

The institutional ethics committee of the NIMHANS, Bangalore, India, approved the recruitment of patients for the current study. All the patients provided written informed consent before their recruitment for this study.

Cognitive evaluation

Cognitive functions across various domains were briefly evaluated using the Montreal Cognitive Assessment (MoCA) scale. Also, the Frontal Assessment Battery (FAB) was used to assess the frontal executive functions. The Rey Auditory Verbal Learning Test (RAVLT) was used to assess verbal learning, memory, and recall, the Complex Figure Test (CFT) andCorsi block-tapping test¹⁸ were used to assess the visuospatial functions, and the Trail Making Test B (TMT-B) was used to assess the executive functions.¹⁹ Table 1 summarizes the demographic and clinical characteristics of the participants.

Table 1. Demographic and clinical characteristics of the subjects.

Parameters	PD-NP (n = 48)	PD-P (n = 42)	Significance, p
Sex (women: men)	7:41	7:35	NS
Mean age, y	57.9 ± 7.0	58.5 ± 7.8	NS
Years of education	12.1 ± 2.4	11.3 ± 2.5 y	NS
Age at onset of PD, y	51.0 ± 9.9	51.9 ± 7.1	NS
Duration of PD	5.7 ± 2.4	6.5 ± 3.2 y	NS
UPDRS-III (OFF)	35.2 ± 8.4	36.3 ± 8.1	NS
UPDRS-III (ON)	12.1 ± 5.9	12.6 ± 5.6	NS
H&Y stage	2.3 ± 0.3	2.4 ± 0.2	NS
Total LED/da	577.5 ± 222.6	722.8 ± 314.5	0.003
Use of DA (%)	54.5	68.6	NS
Pramipexole (mg/d)	3.1 ± 1.1	3.6 ± 1.1	NS
Ropinirole (mg/d)	6.7 ± 2.7	7.2 ± 2.6	NS
Patients on amantadine (%)	25	38	NS
Mean dose (mg/d)a	150.1 ±79.5	185.4 ± 35.2	0.07
Patients on THP (%)a	12.5	40.4	0.002
Mean dose (mg/d)	4.5 ± 1.6	5.0 ± 1.3	NS
HAM-Aa	8.4 ± 5.3	12.6 ± 5.3	<0.001
HAM-D	6.9 ± 4.8	8.6 ± 5.5	NS
MoCA	26.0 ± 2.0	25.7 ± 2.4	NS
FABa	15.5 ± 1.6	14.5 ± 2.1	0.01
Corsi block tappinga
Forward	4.7 ± 0.7	4.3 ± 0.2	<0.001
Backward	3.9 ± 0.6	3.6 ± 0.6	0.01
Complex Figure Testa
Copy	33.6 ± 2.0	30.9 ± 5.6	0.001
Immediate recall	24.4 ± 5.6	20.7 ± 6.0	0.001
Delayed recall	21.1 ± 5.6	17.0 ± 5.3	<0.001
RAVLTa
Total learning	48.5 ± 10.5	44.0 ± 9.0	0.01
Immediate recall	10.9 ± 2.6	9.7 ± 2.1	0.01
Delayed recall	8.6 ± 2.6	7.2 ± 2.0	0.005
Trail Making Test Ba	175.9 ± 53.1	208.9 ± 46.8	0.01
Stroop effect	185.5 ± 99.5	185.1 ± 60.8	NS
Psychosis severity score	—	10.8 ± 3.9	—

Abbreviations: DA = dopamine agonist (pramipexole/ropinirole); FAB = Frontal Assessment Battery; HAM-A = Hamilton Rating Scale for Anxiety; HAM-D = Hamilton Rating Scale for Depression; HC = healthy control; H&Y = Hoehn and Yahr; LED = levodopa equivalent dose; MoCA = Montreal Cognitive Assessment; PD = Parkinson disease; PD-P = patients having PD with psychosis; PD-NP = patients having PD without psychosis; RAVLT: Rey Auditory Verbal Learning Test; THP = trihexyphenidyl hydrochloride; UPDRS-III = Motor Section of the Unified Parkinson’s Disease Rating Scale.

^aRepresents the parameters that are significantly different between PD-NP and PD-P.

Image acquisition

Imaging data were acquired at the NIMHANS, Bangalore, India, using a Philips Achieva 3T MRI scanner with a 32-channel head coil. The diffusion MRI data for each subject were acquired using a single-shot spin-echo, echo-planar sequence in axial sections with the following settings: TR = 8,783 ms, TE = 62 ms, field of view = 224 × 224 mm, section thickness = 2 mm, and voxel size = 1.75 × 1.75 × 2 mm (without any intersection gap) in 15 diffusion-sensitive gradient directions with b value = 1,000 s/mm². Besides, a single image without diffusion weighting was acquired corresponding to b = 0 s/mm².

Image processing and statistical analysis

The preprocessing of diffusion images included manual quality assessment, after converting from DICOM to NIFTI format and were processed using FSL 5.0.9. Eddy current and head motion were corrected using Eddy correct that uses an affine transformation between the baseline b = 0 image and the gradient images.²⁰ Based on the rotation parameter in the affine transformation, the gradients were rotated to match the transformed images. The nonbrain tissue was stripped using the brain extraction tool for each DWI, and a brain mask was created. The diffusion tensors were then computed using least square approximation via the dtifit tool in FSL, and subsequently, fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) maps were generated.

The TBSS pipeline in FSL 5.0.9 was used to compare the DTI metrics between the PD-P and PD-NP groups. First, all the subjects’ FA maps were nonlinearly aligned to FMRIB-58 FA map from Montreal Neuroimaging Institute (MNI) template space. Following the deformable registration, the mean FA skeleton was computed that represented the center of the WM tracts common to all subjects. For MD, AD, and RD, the deformation fields from FA maps were used, and the registered maps were projected onto the FA skeleton. Voxel-wise statistical analysis was performed through a general linear model with contrasts to test for group differences between the PD-P and PD-NP groups. This analysis used the TBSS framework with nonparametric permutation testing (5,000 permutations) for multiple comparisons correction and a threshold-free cluster enhancement (TFCE). Age, sex, LEDD, HAM-A, and duration of disease were used as nuisance covariates. Results were considered significant at p value <0.05, TFCE-corrected for multiple comparisons. The clusters with statistically significant differences in the results of TBSS analysis were chosen as the region of interest, and the mean FA values of these were extracted.

To perform the correlations, the significant clusters from TBSS were used, and the mean values of these measures (e.g., FA and MD) were extracted from each cluster and correlated using Pearson correlation to the neurocognitive scores (RAVLT-delayed recall, Stroop effect, TMT-B, and CFT-delayed recall) as well as the psychosis severity scores separately in the PD-P and PD-NP groups. The neurocognitive scores were adjusted a priori for age, sex, LEDD, HAM-A, and duration of disease by linear regressions, and the resulting standard residuals were used in correlations. Pearson r was computed, and the significance threshold of the correlation was maintained at p value <0.05. Correction for multiple comparisons was performed using False Discovery Rates,²¹ with a q value threshold at 0.1. Student t tests and χ² tests were used to compare the continuous variables and categorical variables in table 1, respectively.

To gain a deeper understanding of the anatomic areas and attain insights into the gross changes, we mapped the Johns Hopkins University (JHU) atlas²² to MNI space (in which TBSS was performed) and extracted the mean FA of each of the 21 tracts that are defined in the JHU atlas. The list of the tracts is given in table 2.

Table 2. FA values of the 21 tracts registered in the JHU atlas.

Tracts	PD-P	PD-NP	q Value
AThR-L	0.466 ±0.015	0.475 ± 0.014	1.84E-04
AThR-R	0.468 ±0.014	0.477 ± 0.014	6.26E-07
CST-L	0.547 ±0.019	0.551 ± 0.015	1.84E-04
CST-R	0.539 ±0.018	0.547 ± 0.015	2.12E-04
Cg-L	0.453 ± 0.022	0.464 ± 0.018	2.74E-04
Cg-R	0.433 ± 0.020	0.443 ± 0.019	9.29E-05
Hpc-L	0.396 ± 0.019	0.402 ± 0.019	5.95E-03
Hpc-R	0.404 ± 0.020	0.413 ± 0.021	2.52E-03
FMa	0.531 ± 0.018	0.539 ± 0.018	2.34E-04
FMi	0.481 ±0.021	0.489 ±0.019	2.52E-03
IFOF-L	0.464 ± 0.021	0.473 ± 0.019	6.10E-04
IFOF-R	0.457 ± 0.020	0.465 ± 0.019	2.34E-04
ILF-L	0.396 ± 0.016	0.402 ± 0.016	3.14E-05
ILF-R	0.389 ± 0.015	0.395 ± 0.015	1.96E-04
SLF-L	0.423 ± 0.020	0.431 ± 0.015	2.34E-04
SLF-R	0.421 ± 0.020	0.429 ± 0.015	4.04E-04
Unci-L	0.405 ± 0.018	0.411 ± 0.016	1.16E-04
Unci-R	0.409 ± 0.017	0.416 ± 0.016	1.81E-04
SLFt-L	0.401 ± 0.021	0.409 ± 0.017	4.73E-02
SLFt-R	0.443 ± 0.022	0.449 ± 0.017	1.90E-03
BCC	0.662 ± 0.025	0.671 ± 0.030	2.52E-03

Abbreviations: AThR = anterior thalamic radiation; bCC = body of the corpus callosum; Cg = cingulum (cingulate gyrus); CST = corticospinal tract; FA = fractional anisotropy; FMa = forceps major; FMi = forceps minor; Hpc = cingulum (hippocampus); IFOF = inferior fronto-occipital fasciculus; ILF = inferior longitudinal fasciculus; JHU = Johns Hopkins University; PD-P = patients having PD with psychosis; PD-NP = patients having PD without psychosis; SLF = superior longitudinal fasciculus; SLFt = superior longitudinal fasciculus (temporal part); Unci = uncinate fasciculus.

Results

Demographics, clinical characteristics, and cognitive evaluations

The 2 groups were matched for age and sex and were not significantly different in terms of age at onset of PD and the duration of PD. Of those 42 PD-P, 11 had isolated VH, 27 had isolated MH (isolated presence hallucination: 15, a combination of presence and passage hallucination: 10, and a combination of presence hallucination and illusion: 2), and 4 had a combination of VH and MH (all had presence hallucinations). The severity of motor symptoms (assessed by the UPDRS-III) and stage of PD (assessed by the H&Y scale) were similar in 2 PD subgroups. The PD-P group had greater HAM-A compared with the PD-NP group, whereas the LEDD was also significantly higher in the PD-P group compared with the PD-NP group. The mean psychosis severity score of the PD-P group was 10.8 ± 3.9 (table 1). The global cognitive performance, as measured by the MoCA, revealed no difference between the 2 PD groups. The PD-P group performed poorly in all neuropsychological tests including the FAB, Corsi block-tapping test, RAVLT, CFT, and TMT-B compared with the PD-NP group. A detailed comparison of the neuropsychological parameters is provided in table 1.

Image analysis

PD-P had significantly lower FA in 2 clusters. Cluster 1 included the corpus callosum, right inferior longitudinal fasciculus (ILF), corticospinal tract (CST), and right occipitoparietal WM (p value <0.05, after correction). The other cluster (cluster 2) comprised the left ILF and inferior fronto-occipital fasciculus (IFOF) (p value <0.05, after correction). Figure 1 presents the WM tracts showing significant group differences in FA. No significant differences were observed in the MD, AD, and RD features. Figure 2 demonstrates the mean and SDs of the FA values over the 21 tracts defined in table 2. It can be observed that the PD-P group has a lower FA in all the tracts than the PD-NP group. Table 3 illustrates the mean cluster values extracted from TBSS and provides the significance between the mean values of these clusters.

Figure 1. Cluster of white matter tracts having a significant difference in fractional anisotropy in patients having PD with hallucinations.

Cluster 1: corpus callosum + right inferior longitudinal fasciculus and inferior fronto-occipital fasciculus + corticospinal tract + right occipitoparietal white matter. Cluster 2: left inferior longitudinal fasciculus–inferior fronto-occipital fasciculus. PD = Parkinson disease

Figure 2. FA values of all 21 tracts registered in the Johns Hopkins University white matter atlas.

AThR = anterior thalamic radiation; bCC = body of the corpus callosum; CST = corticospinal tract; Cg = cingulum (cingulate gyrus); Hpc = cingulum (hippocampus); FA = fractional anisotropy; FMa = forceps major; FMi = forceps minor; IFOF = inferior fronto-occipital; ILF = inferior longitudinal fasciculus; SLF = superior longitudinal fasciculus; SLFt = superior longitudinal fasciculus (temporal part); Unci = uncinate fasciculus.

Table 3. Fractional anisotropy values of the significant clusters.

WM cluster	PD-P	PD-NP	p Value
Cluster 1	0.512 ± 0.025	0.536 ± 0.023	4.32E-06
Cluster 2	0.448 ± 0.024	0.470 ± 0.019	4.31E-06

Abbreviations: PD-P = patients having PD with psychosis; PD-NP = patients having PD without psychosis; WM = white matter.

Cluster 1: corpus callosum + right inferior longitudinal fasciculus and inferior fronto-occipital fasciculus + corticospinal tract + right occipitoparietal white matter.

Cluster 2: left inferior longitudinal fasciculus–inferior fronto-occipital fasciculus.

In PD-P, the FA value of both the clusters correlated with the Stroop effect (p value <0.05, r = −0.337 for cluster 1 and r = −0.328 for cluster 2), whereas the FA value of only cluster 2 correlated with the RAVLT (p value <0.05, r = 0.228) (figure 3).

Figure 3. Correlation of the fractional anisotropy values of the white matter tracts with the neuropsychological scores.

FA = fractional anisotropy; RegRAVLT = Regressed Rey Auditory Verbal Learning Test; RegStroopE = Regressed Stroop Effect.

Discussion

This study aimed at exploring the microstructural abnormalities in the WM of PD-P (with VH and MH) in comparison to a cohort of patients without psychosis, having a comparable duration of disease, age at onset, UPDRS, and H&Y staging. We used TBSS on diffusion-weighted MRIs to attain insights into the abnormalities in the WM tracts. The key findings of this study were widespread WM abnormalities involving the bilateral ILF, bilateral IFOF, right parieto-occipital WM, the body of the corpus callosum, and the right CST. These results also correlated with the scores of the RAVLT and Stroop test (figure 3). Moreover, when the mean FA values of each of the tracts were computed, an overall trend of lower FA in PD-P compared with PD-NP was observed.

WM tract abnormalities in PD-P are relatively understudied compared with major psychiatric illness such as schizophrenia. Although the exact neural correlates remain elusive, several studies have suggested a key association of WM abnormalities in schizophrenia.^23,24 These studies, in particular, have demonstrated myelin dysfunction and alterations in the oligodendrocyte numbers in schizophrenia. FA is a measure of anisotropy of free water diffusion in WM tracts, and the anisotropy has uneven contributions from the axon and myelin. Abnormalities in the WM tracts would augment the diffusion of water molecules in random directions resulting in a reduction in anisotropy, resulting in a reduced FA. Auditory hallucinations are common in schizophrenia, and in 1 study, patients having schizophrenia with auditory verbal hallucinations had increased FA of WM fibers of the corpus callosum connecting the auditory cortex in both sides.²⁵ Several other studies have reinforced abnormalities in WM tracts related to speech and language processing areas of the brain.^26,27 Similar to the well-established fact of an overall reduction of FA in schizophrenia,²⁸ WM tract abnormalities in the pathogenesis of hallucinations in PD are grossly expected but have not been studied extensively till now.

The involvement of the ILF and IFOF as demonstrated in figure 1 reinforces the theory that altered processing of visual information plays an essential role in the genesis of VH/MH in PD. The ILF connects the temporal and occipital lobes, running along the lateral walls of the inferior and posterior cornua of the lateral ventricle. The IFOF runs superomedial to the optic pathways and spatially overlaps with the ILF along part of their pathways. Functionally, the ventral visual pathway, which follows the ILF, extends from the occipital lobe into the anterior and ventral portions of the temporal lobe. This pathway has a vital role in representing the aspects of visual information critical for encoding object identity, whereas the IFOF is speculated to play a crucial role in attention and visual processing. As abnormalities in attention and visual processing have been reported in PD-P,²⁹ structural abnormalities of the ILF and IFOF are vital in this context. Alteration in these tracts might hamper the accurate relaying of the visual information from the occipital cortex to the temporal regions and subsequently to the orbitofrontal cortex. Moreover, recent findings indicate that the ILF probably mediates the fast transfer of visual signals forward from the occipital cortex to the parahippocampal gyrus and neuromodulatory projections from the amygdala back to visual areas aiding in fine-tuning of visual information processing.³⁰ A study on patients having schizophrenia with VH has also reported abnormalities in the ILF.³¹ In another DTI-based study, developing children with visuoperception abnormalities were found to have altered microstructural integrity in the left temporal tracts compared with those without visuoperceptual problems,³² whereas impaired visual memory was shown to correlate with the right ILF.³³ It is interesting to note that visuoperceptual abnormalities have been reported in patients having PD with VH in this study and in several previous studies.³⁴ Reports of visual agnosia, prosopagnosia, and impaired visual recent memory in animals and humans with lesions in the ILF further support the role of the ILF in visuospatial information processing.^35,36 In an interesting case report, compromised structural characteristics were demonstrated in a patient with occipital stroke who developed VH.³⁷ WM tracts connecting the visual cortex with the frontal and temporal regions had structural abnormalities in the same patient, suggesting large-scale disruption of WM tracts from the occipital cortex in the setting of VH. Considering all the aforementioned facts, structural alterations in the ILF and IFOF seem to play a significant role in the pathogenesis of hallucinations in PD.

The FA value in the cluster having the right occipitoparietal WM was also reduced in the PD-P group. Although there is no previous evidence to suggest any role of WM abnormality in the parietal lobe in the genesis of hallucinations in PD, structural abnormalities in the parietal lobe have been reported in schizophrenia.³⁸ Complex VHs have been observed in parietal lobe pathology³⁹; however, the exact mechanism is still not clearly understood. It is postulated that the structural alterations in the parietal lobe in patients presenting with VH perhaps represent the lesions in the retino-geniculo-calcarine tract responsible for carrying visual information from the retina to the visual cortex. This tract traverses through various areas of the brain, including parts of the occipital cortex, parietal cortex, and lesions of this tract at different levels have been associated with VH.^40,41

Previous studies in schizophrenia have indicated an association of transcallosal disconnectivity in the genesis of positive and negative symptoms.⁴¹ As mentioned above, a study on patients having schizophrenia with auditory verbal hallucinations had increased FA of the WM fibers of the CC connecting the auditory cortex in both sides.²⁵ It is unclear how exactly CC abnormalities lead to the genesis of VH/MH in PD. However, it can be speculated that transcallosal dysconnectivity probably leads to a mismatch of visual information in both hemispheres resulting in faulty visual processing. It is warranted that future studies focus on various segments of the CC to obtain more in-depth insight regarding the role of the CC in the emergence of psychotic symptoms in PD. Structural abnormalities in the CC and parietal lobe have been shown to associate with cognitive deficits in PD.⁴² Given the correlation between specific cognitive deficits and FA in the above structures and the higher prevalence of cognitive deficits in PD-P, a confounding effect cannot be ruled out.

In the PD-P group, the FA values of the 2 clusters of WM tracts were correlated with the cognitive scores. The FA values of both the clusters correlated negatively with the Stroop effect, whereas the FA value of only cluster 2 correlated positively with the RAVLT. This result indicates that significant executive dysfunction (Stroop effect) and deficits in memory and learning (RAVLT) could potentially be associated with the genesis of hallucinations in PD.

We recognize the cross-sectional design of this study as a limitation because the patients who did not have hallucinations at the time of participation may develop hallucinations in the future. This warrants future studies with longitudinal design to confirm the novel insights obtained from our research. Besides, the PD-P group in this study had more significant cognitive impairment compared with PD-NP, which could potentially confound the imaging results. However, as there is a strong association of cognitive impairment with psychosis in PD,⁴³ matching 2 groups for the cognitive status along with age, sex, and education would not be practical and would remove a robust natural association as well.

In terms of imaging, because we did not acquire another diffusion data set in posterior to anterior phase encoding direction, which is required to compute the susceptibility off-resonance field, distortion correction was not performed. Furthermore, the data acquired were not perfectly isotropic (with a voxel size of 1.75 × 1.75 × 2 mm), which may perhaps affect the FA values.⁴⁴ However, based on earlier studies, the alterations in FA are seen in voxel sizes with a larger anisotropy. In our case, the anisotropy of the voxel size was very low that may not alter the FA values significantly.⁴⁵

Considering significant heterogeneity in the motor and nonmotor symptoms of PD, especially the profound effects of sex and ethnicities on the nonmotor symptoms,⁴⁶ results of the current study may not be generalizable to all the patients with PD.

To conclude, our comprehensive voxel-based analysis, which compared the microstructural WM integrity of PD-P and PD-NP, revealed altered structural integrity in several tracts including the ILF, IFOF, right parietal WM, CC, and parts of the motor tract. This study provides novel insights into the putative role of WM tract abnormalities in patients having PD with VH.

Glossary.

AD

axial diffusivity

CFT

Complex Figure Test

CST

corticospinal tract

DTI

diffusion tensor imaging

FA

fractional anisotropy

FAB

Frontal Assessment Battery

H&Y

Hoehn and Yahr

HAM-A

Hamilton Anxiety Rating Scale

IFOF

inferior fronto-occipital fasciculus

ILF

inferior longitudinal fasciculus

JHU

Johns Hopkins University

LEDD

levodopa equivalent dose per day

MD

mean diffusivity

MH

minor hallucination

MNI

Montreal Neuroimaging Institute

MoCA

Montreal Cognitive Assessment

NIMHANS

National Institute of Mental Health and Neurosciences

PD

Parkinson disease

PD-NP

patients having PD without psychosis

PD-P

patients having PD with psychosis

RAVLT

Rey Auditory Verbal Learning Test

RD

radial diffusivity

TBSS

tract-based spatial statistics

TFCE

threshold-free cluster enhancement

TMT-B

Trail Making Test B

VH

visual hallucination

WM

white matter.

Data availability

Anonymized data can be obtained with the request from any qualified investigator for purposes of replicating procedures and results.