Locus Coeruleus Integrity from 7T MRI Relates to Apathy and Cognition in Parkinson's Disease and Progressive Supranuclear Palsy

Background and Objectives The loss of noradrenergic neurons in the locus coeruleus (LC) contributes to various cognitive and neuropsychiatric symptoms in Parkinson's disease (PD) and progressive supranuclear palsy (PSP). The spatial precision of in vivo LC imaging is improved using a magnetisation transfer sequence combined with ultra-high field 7T MRI. This study aimed to test the sensitivity of LC imaging in PD and PSP, to characterise the spatial pattern of LC atrophy in patients, and its relationship to cognition and apathy. Methods This cross-sectional observational study recruited patients with idiopathic PD, probable PSP-Richardson's syndrome and age-matched healthy controls (HC) via specialist clinics and volunteer registries. All participants underwent clinical assessments for cognition and apathy, and high-resolution (0.08 mm3) LC scans. To quantify LC integrity, the contrast-to-noise ratio (CNR) relative to a pons reference region was calculated and extracted using a probabilistic atlas. Subregional mean CNRs were summarised to test group differences and to correlate LC integrity with apathy and cognition scores. LC clusters were identified to confirm the spatial pattern of the effect (threshold free cluster enhancement, 10000 permutations, p<0.05, corrected for family-wise error). Results Twenty-five patients with PD, 14 with PSP and 24 controls with completed dataset were included in the study. Patients with PSP were more impaired on global cognition and apathy scores, compared to controls and PD. Clusters with reduced contrast were observed in the caudal LC for both PD and PSP patients relative to controls (HC>PD, right caudal LC, 37 voxels; HC>PSP, bilateral caudal LC, 206 voxels). PSP and PD patients showed similar levels of LC degeneration, but this was more extensive in PSP. Across both disease groups, LC integrity was associated with cognitive performance (MoCA: F(1,37)=5.339, p=0.027, bilateral LC, 200 voxels; ACE-R: F(1,37)=4.297, p=0.045, left LC, 70 voxels) and apathy (Apathy Scale: F(1,37)=4.335, p=0.044, left LC, 99 voxels). Discussion We confirm the sensitivity of 7T LC imaging to detect sub-regional LC changes in PD and PSP. The relationship between LC integrity and non-motor symptoms highlights the potential for noradrenergic therapeutic strategies to ameliorate cognitive and behavioural features of PD and PSP.


Introduction
Parkinson's disease (PD) and progressive supranuclear palsy (PSP) result from distinct pathologies, yet they share a number of non-motor symptoms. These include executive dysfunction and neuropsychiatric features, with apathy [1][2][3] . While the causes of cognitive and neuropsychiatric symptoms in these diseases are multi-factorial, there is increasing evidence that noradrenergic dysfunction plays a prominent role [4][5][6][7] .
In both PD and PSP, the noradrenergic locus coeruleus is a site of selective vulnerability, undergoing early and severe pathological changes along its rostrocaudal axis [8][9][10][11] . Highly collateralised projections enable the locus coeruleus to release noradrenaline in widespread brain regions. The effect of noradrenaline at target regions is to alter the gain of neurons and facilitate reorganisation of functional networks 12,13 . Through these actions, noradrenaline influences many cognitive functions, in particular promoting attentional focus and flexible shifting between goal-directed behaviours 14 . In PD and PSP, vulnerability of locus coeruleus neurons and consequent loss of noradrenergic projections has a wide-ranging impact on cognition and behaviour 15 .
We propose that noradrenergic deficits are determinants of cognitive and behavioural impairments in PD and PSP, and are therefore a promising treatment target for cognitive and neuropsychiatric symptoms. However, individual differences in brain structure and neurochemistry determine a patient's response to noradrenergic treatment, with a non-linear (U-shaped) pattern similar to dopaminergic and serotonergic interventions 6,7,16,17 .
Specifically, it is proposed that an individual's need for -and response to -noradrenergic treatment depends on the state of their locus-coeruleus noradrenergic system. Progress in noradrenergic therapeutics has been hindered by the need for in vivo characterisation of individuals' locus coeruleus degeneration. The locus coeruleus is small and elongated (~1x16 mm) with few neurons (~50,000 in humans), which challenges standard imaging techniques. Specialised magnetisation-transfer sequences have been developed, that are sensitive to the intrinsic contrast generated by neuromelanin-rich noradrenergic cells of the locus coeruleus 18,19 . These correlate with the density of neuromelanin-positive neurons in the locus coeruleus 20,21 , providing a biomarker of locus coeruleus degeneration 22 that confirms reductions in PD and PSP 23,24 . Such sequences are . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 27, 2021. ; https://doi.org/10.1101/2021.04. 19.21255762 doi: medRxiv preprint most sensitive at ultra-high field 7T MRI 25,26 , providing both enhanced contrast-to-noise and improved spatial resolution.
Here, we test the hypothesis that apathy and cognitive function correlate with the integrity of the locus coeruleus in PD and PSP, exploiting ultra-high field neuromelanin-sensitive 7T MRI. The high resolution (0.08 mm 3 ) enabled localisation of degeneration to sub-regions of the locus coeruleus. Given the prominent role the locus coeruleus-noradrenergic system plays in cognition and goal-directed behaviours, we predict that clinical assessments of cognition and apathy will relate to locus coeruleus integrity, with a rostro-caudal gradient of degeneration.

MRI Acquisition
Participants underwent MR scans on a 7T Magnetom Terra (Simens, Erlangen, Germany) with a 32-channel head coil (Nova Medical, Wilmington, USA). Two controls had incidental structural abnormalities discovered on diagnostic scans. The imaging data for one PSP patient was missing due to coil reconstruction failure. These participants were excluded from the following analyses. The locus coeruleus was imaged using a sensitive 3-D magnetisation transfer (MT) weighted sequence 25   CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 27, 2021. ; https://doi.org/10.1101/2021.04.19.21255762 doi: medRxiv preprint enhance the signal-to-noise ratio. An additional scan without MT pre-saturation (MT off ) was acquired for registration purposes. A high resolution T1-weighted structural image was acquired using MP2RAGE sequence: TE = 2.58 ms, TR = 3500 ms, BW = 300 Hz/px, voxel size = 0.7 x 0.7 x 0.7 mm 3 , FoV = 224 x 224 x 157 mm 3 , acceleration factor (A>>P) = 3, flip angles = 5/2° and inversion times (TI) = 725/2150 ms for the first/second images.

Locus Coeruleus Integrity Estimation
To limit potential confounds in standard intensity-based segmentation and increase the sensitivity, an atlas-based segmentation approach was adopted to estimate LC integrity using a 5% probabilistic LC atlas in the standard space, generated from an independent 7T data of 29 older healthy volunteers 7 . The MT images were co-registered to the high-resolution (0.5 mm isotropic) ICBM152 2009b template following a T1-driven, intra-modality coregistration pipeline using ANTs (v2.2.0) 26 . In summary, MT images were off-line bias field corrected and averaged; the roadmap used for image standardisation were generated using registration parameters estimated in the following steps (in order): MT off to MT, individual T1 to MT off , individual T1 to T1 study-wise group template, and T1 group template to ICBM152 template.
The standardisation of MT images was completed by applying all the transformation parameters in one step. Voxelwise contrast-to-noise ratio (CNR) of the LC was computed against a pontine reference region using the difference between a given voxel and the mean intensity of the reference region divided by the standard deviation of the reference signals.
The LC mask was applied on CNR maps to extract averaged contrast for each slice, three equidistantly subdivided regions (rostral, central and caudal), left and right LC, and the whole structure.
Potential confounds that might affect MT signals were explicitly measured and compared . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Statistical Analyses
Statistical analyses were performed in JASP (v0.11.1) and R (version 3.6.1). Group differences in clinical assessments and LC contrast across the subregions were tested using ANOVA and post hoc tests for controls and patients. The relationships between regional mean LC contrast and clinical assessments were specifically tested in patient groups with linear regression models with clinical scores as the dependent variable and LC CNR as fixed effect. Additional linear models were tested with age as a covariate of no interest and to test the group interaction within patient groups. We report the results for the hypothesis testing with a significance level of p = 0.05. The effect size was estimated using Cohen's d in pairwise tests and R 2 for linear regression models. The 95% confidence interval (CI) was provided for variables.

Voxelwise Analyses
Voxelwise analyses provide a more sensitive and spatially more accurate localisation for the LC group differences and clinical correlations, as averaged CNR is insensitive to voxel-byvoxel variations in LC integrity. The co-registered CNR maps were smoothed with a 1 mm FWHM Gaussian kernel, masked by the 5% 7T LC atlas, then subjected to general linear models for testing group differences. Relationships with clinical variables were assessed by regression models where clinical scores were mean-centred across patient groups. Voxelwise analyses used the randomise function in FSL (v6.0.0). A threshold free cluster enhancement (TFCE) method was adopted for cluster inference combined with permutation tests (10000 iterations). Analogous regression models used age included as covariate of no interest. The family-wise error (FWE) corrected p-value (<0.05) was used to determine significant clusters.

Data Availability
The MT and MP2RAGE data used in this study are available for non-commercial academic purposes upon request.

Demographics and Clinical Assessment
Demographics and clinical data are summarised in Table 1. Controls were age-and sexmatched to patient groups. Patients with PSP had fewer years of education compared to controls, and they were more impaired on the MMSE, MoCA, ACE-R, Apathy Scale and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 27, 2021. ; https://doi.org/10.1101/2021.04.19.21255762 doi: medRxiv preprint HADS-depression assessments relative to both controls and PD. Patients with PD had higher LEDD and RBDSQ scores than PSP.

Locus Coeruleus Group Differences
LC integrity assessment benefited from the atlas-based approach, which had clear advantages over classic segmentation-based methods. Pathology-related signal reduction in the LC is accompanied by increased segmentation failure and reduced spatial precision, leading to inflated contrast estimation. Using regional averaged LC contrast, there were group differences in the caudal subregion (F(2,60)=4.05, p=0.02) between PSP patients and controls Voxelwise analyses revealed clusters with significantly reduced contrast (threshold free cluster enhancement, p<0.05, FWE-corrected) in both disease groups compared to controls (Table 2). A small cluster with reduced contrast was identified for PD patients in the right caudal LC (Fig 1C, 37 voxels) whereas PSP patients had extensive bilateral caudal damage in the LC (Fig 1D, 206 voxels). There was no group difference between PD and PSP by voxelwise analysis.

Locus Coeruleus Integrity and Clinical Assessments in Patients
MoCA performance was associated with whole LC contrast (Fig 2A, F(1,37 (Fig 2F), that is, reduced LC integrity was associated with worse apathy. The details for the LC clusters associated with cognition and apathy were summarised in Table 2. Adding age as a covariate of no interest did not meaningfully change the results of voxelwise analyses.

Discussion
Using ultra-high field 7T magnetisation-transfer imaging, we confirm in vivo loss of locus coeruleus integrity in two neurodegenerative diseases (PD and PSP) that have well established locus coeruleus pathology at post mortem. We confirm the regionally specific pattern of predominantly caudal locus coeruleus changes in PD and PSP, which was associated with cognitive impairment and apathy. These findings highlight the importance of the locus coeruleus-noradrenaline system in cognitive and behavioural features of these diseases. They demonstrate the utility of neuromelanin-sensitive imaging at ultra-high field to measure the locus coeruleus in neurodegenerative disease, which we suggest will be an important biomarker for use in future clinical trials, where patients could be stratified according to locus coeruleus integrity 7,26 .
The simple group-wise effect of PD and PSP on the locus coeruleus was most apparent in the caudal subregion. This rostro-caudal gradient has been noted in some neuropathological studies of PD 31 , although not all 32 . More severe caudal degeneration in conditions with prominent behavioural and motor features, such as PD and PSP, is consistent with the . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 27, 2021 33,34 , while Alzheimer's disease, Down's syndrome 32 , and healthy ageing 35 have been associated with greater cell loss from the rostral locus coeruleus. In our PD and PSP participants, the relationship of LC signal to behavioural (Apathy scale) and cognitive (e.g., MoCA) change was not confined to caudal regions, but extended to mid and rostral LC respectively (Fig 2).
Whereas the caudal locus coeruleus has some topographic specific projections to subcortical regions, rostral projections preferentially target prefrontal cortex 33,34 , consistent with their role in diverse cognitive functions. Locus coeruleus degeneration and reduced noradrenaline in neurodegenerative disease is linked with a higher risk of progression to dementia 5,36 , and faster cognitive decline in older adults 37,38 . The correlation of apathy to LC integrity -most clearly in the PSP group -is in keeping with extensive human and preclinical evidence linking goal directed behaviour with the noradrenergic system 14,39,40 . Our results highlight that whilst cognitive and motivation deficits in neurodegenerative disease have multifactorial causes, the locus coeruleus-noradrenaline system plays a key role via its widespread modulation of target structures supporting these processes.
Cognitive decline and disorders of goal directed behaviour, including the co-occurrence of both apathy and impulsivity 41,42 , are common in both PD and PSP. The link we have shown between locus coeruleus integrity and these non-motor symptoms supports the rationale for noradrenergic therapies in selected patients. Preliminary work in PD suggests that the noradrenergic reuptake inhibitor atomoxetine can improve goal directed behaviour 6 and that drug responsivity depends on locus coeruleus integrity 7 . The ability to detect localised changes in the locus coeruleus that we have demonstrated here in PD and PSP, supports the hypothesis that this technique could be used to stratify patients in future clinical trials of noradrenergic therapy in neurodegenerative disease.
The current study exploits the benefit of ultra-high field 7T MRI, for assessment of the locus coeruleus in vivo. The short imaging sequence 25 was well-tolerated in a clinical setting. For locus coeruleus signal extraction, we advanced a cross-modality co-registration pipeline and an atlas-based approach, to maximise spatial precision and avoid estimation biases that occur with classic segmentation-based methods. Potential confounds specific to locus coeruleus imaging were also controlled and explicitly tested, including the consistency of noise level . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 27, 2021. ; https://doi.org/10.1101/2021.04.19.21255762 doi: medRxiv preprint along the rostrocaudal axis and the effect of age. Moreover, the spatial localisation of effects was further improved by voxelwise analyses that exploited the high resolution afforded by 7T MRI. Not only do these advances allow for more accurate locus coeruleus quantification, they permit the localisation of changes to locus coeruleus sub-regions, which have previously been identified in post mortem studies 32,43,44 . The histological validation of neuromelaninsensitive imaging 20 provides additional evidence that the results we report here reflect the underlying pathological changes that have been observed at post mortem. Together, these findings pave the way for future clinical studies to investigate contributions of the locus coeruleus-noradrenaline system in the pathogenesis, progression and treatment of neurodegenerative diseases.
There is significant heterogeneity in the expression of motor, neuropsychiatric and cognitive features across both PD 1 and PSP 45 . Locus coeruleus pathology may be more prominent in certain PD phenotypes, including patients with cognitive impairment or dementia 46, 47 , depression 24 , and REM sleep behaviour disorder 48 . Future studies using ultra-high field locus coeruleus imaging in PD and PSP may focus on specific phenotypes, to define the role of regionally specific locus coeruleus changes in specific non-motor symptoms. For example, the caudal sub-region may be especially vulnerable as it lies adjacent to the 4 th ventricle, exposed to environmental insults from the cerebrospinal fluid 49 . Also, the caudal locus coeruleus receives projections from the vagal nerve via the nucleus of the solitary tract which might speculatively make it more vulnerable to transmission of misfolded proteins from the periphery 50 . Future studies would be required to establish the relationship to peripheral molecular pathology.

Conclusions
The locus coeruleus is a site of selective vulnerability across neurodegenerative diseases, and is hypothesised to influence the development of some cognitive and behavioural symptoms 15,22 . We provide in vivo confirmation of locus coeruleus degeneration in Parkinson's disease and Progressive Supranuclear Palsy and demonstrate its association with cognition and apathy. Ultra-high field 7T neuromelanin imaging is sensitive to the pathological changes in the locus coeruleus and its associated clinical features. This MRI technique can provide a valuable tool for investigating non-motor symptoms and stratifying patients in clinical trials of noradrenergic therapy.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 27, 2021. ; https://doi.org/10.1101/2021.04.19.21255762 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Potential Conflicts of Interest
The authors declared no conflict of interest.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 27, 2021  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.     . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 27, 2021. ; https://doi.org/10.1101/2021.04.19.21255762 doi: medRxiv preprint