Unravelling the Alteration of Brain Structure and Function in Parkinson´s Disease With Ultra-high Field MRI

BACKGROUND Parkinson's disease (PD) is a multi-system neurodegenerative disease affecting a subset of vulnerable cells in various brain structures, including brain stem nuclei. Selective degeneration of dopamine (DA) producing neurons in the pars compacta of the substantia nigra (SNc) is a key neuropathological feature in PD and gives rise to the classical motor symptoms which can be symptomatically treated by dopamine replacement therapy. Yet the link between DA and PD extends far beyond the motor system, pertaining to core aspects of cognition that are supported by dopamine signalling: For instance, there is converging evidence that reward computations and motivational drive are significantly affected by PD and its dopaminergic treatment. Noradrenaline (NA) producing cells of the pontine nucleus locus coeruleus (LC) are also affected in PD with early and severe degeneration. Although the LC is the main source of NA in the brain, and NA has been shown to be an important neuromodulator, it is unclear how neurodegeneration of the LC contributes to motor and non-motor symptoms in PD.

PD and the dopaminergic system Several midbrain nuclei contain cell groups that synthesize the neurotransmitter DA and project from the ventral tegmental area (VTA) and SNc to cortical areas and the striatum. It has recently been suggested that VTA and SNc have distinct roles in reinforcement learning: dopamine bursts from VTA are thought to reinforce stimulus values, while the SNc is specifically important for action values. Since PD preferentially affects the ventral tier of the SNc leading to the motor symptoms, it has been hypothesized that dopaminergic treatment leads to an "overdosing" of the less impaired pathways from VTA to striatum.

PD and the noradrenergic system Parallel to the degeneration of dopaminergic neurons in SNc, noradrenergic neurons in the pontine locus coeruleus (LC) exhibit Lewy pathology and undergo a marked degeneration early on in PD which exceeds that of dopaminergic SNc neurons. The LC is the largest source of noradrenaline (NA) to the brain, and exerts neuromodulatory effects through extensive ascending projections to the entire cerebral cortex. The loss of NA neuromodulation associated with PD might contribute to both motor and non-motor symptoms and could potentially be a therapeutic target, however the effect and significance of NA deficiency in PD pathogenesis is to date not fully understood. The LC is thought to be involved in the modulation of attention and arousal and also in the processing of emotion. The role of LC neurodegeneration and the resulting decrease in noradrenaline levels to the pathophysiology of PD is not fully understood. However, a large number of studies in humans and animals support a double role of noradrenergic cell loss in PD; a contribution to motor- and non-motor symptoms as well as a facilitation of DA neurodegeneration.

AIMS The overall purpose of the study is to map structural and functional changes in the subcortical dopaminergic and noradrenergic networks that mediate cognition, attention and emotion in PD.

The study aims to:

• Measure PD related neurodegeneration on 7T sMRI in ROIs corresponding to neuromodulatory brainstem nuclei: Substantia nigra pars compacta, nigrosome-1, ventral tegmental area \& locus coeruleus
• Predict disease severity by multivariate pattern of neurodegeneration measures in these areas.
• Map the relationship between the autonomic arousal response to arousing stimuli, neural activity in the noradrenergic arousal network and the degree of neurodegeneration in noradrenergic brain regions.
• Investigate the effects of neurodegeneration in the core dopaminergic nuclei (SNc and VTA) and dopaminergic medication on the midbrain striatal network mediating action-value and stimulus-value learning.

HYPOTHESES

Structural changes of SN and LC integrity:

1. Participants with PD will have significantly decreased signal on MTw-imaging in the SN and LC quantified as reduced contrast ratios in SN/cerebral peduncle and LC/pontine tegmentum.
2. SN and LC volumes, calculated as the sum of supra-threshold voxels in the regions of SN and LC, will be significantly decreased in participants with PD.
3. Accelerated iron accumulation in SN and nigrosome-1 will lead to an increased hypointensity of the SN and a loss of a distinct nigrosome-1 region on SWI, R2* and QSM.
4. R2* and QSM values in SN on Quantitative Susceptibility Mapping will be significantly higher in participants with PD.
5. Motor impairment (UPDRS-3 subscore) can be predicted by multivariate pattern analysis of MTw- and SWI-images of the SN.
6. Measures of LC structural integrity will correlate with non-motor symptoms, specifically apathy and depression. Functional changes:
7. Participants with PD will have significantly decreased functional activation in a region corresponding to the LC when exposed to arousing stimuli.
8. Functional activation in the LC will correlate with LC structural integrity.
9. Autonomic arousal responses (pupil response and skin conductance response) will correlate with LC structural integrity and functional activation.
10. In healthy controls, there will be increased SN-dorsal striatal connectivity for action value learning, increased VTA-ventral striatal connectivity for stimulus value learning.
11. In PD patients off their medication, there will be reduced SN-dorsal striatal connectivity. This will impair action value learning while stimulus value learning will only be mildly affected.
12. Dopaminergic medication restores dopamine levels in the dorsal striatum and "overdoses" ventral striatum. Medication will restore (at least partially) SN-dorsal striatal connectivity and dorsal striatal activity which will lead to largely unimpaired action value learning. "Overdosing" will result in exaggerated VTA-ventral striatal connectivity and exaggerated ventral striatal activity. In subjects with strong ventral striatal "overdosing" inappropriate stimulus value learning might occur.

RESEARCH PLAN Part 1: 7T and 3T structural MRI to map structural changes in SN and LC and their clinical correlates in PD.

Participants: 60 patients with idiopathic PD (aged 18 or more), 20 age-matched healthy control subjects.

Ultra-high field MRI. All participants will be scanned using a research-only 7T Achieva MR System (Philips, Best, The Netherlands) located at Hvidovre Hospital. MR-scanning will be performed with a dual transmit, 32-channel receive head coil (Nova Medical Products), including a neuromelanin sensitive, magnetization transfer weighted (MTw) sequence at 0.4x0.4x1.0mm resolution, a susceptibility weighted imaging (SWI) sequence at 0.4mm isotropic resolution and a multi-echo quantitative R2* sequence at 1.0mm isotropic resolution. Additionally, all participants will be scanned using a 3T Siemens Prisma MR system.

Additional examinations: Outside the scanner, motor and non-motor symptoms (UPDRS, Unified Parkinson's Disease Rating Scale \& NMSS, Non-Motor Symptom Scale), cognition (MoCA, Montreal Cognitive Assessment), apathy (LARS, Lille Apathy Rating Scale), depression (BDI-II, Beck's Depression Inventory), impulse control disorders (QUIP, Parkinson's Disease Impulsive-Compulsive Disorders Questionnaire), impulsivity (BIS-11, Barratt Impulsiveness Scale), and handedness (EHI, Edinburgh Handedness Inventory) will be assessed.

Part 2: 7T fMRI to assess DA and NA dysfunction in PD Participants: A sub-group of individuals who participated in part 1 will be asked to participate in Part 2: It is expected that 20 patients and all the healthy participants are able to lie sufficiently still to cooperate in a fMRI experiment.

Participants will be scanned again with the 7T Achieva MR system using a reduced transverse field-of-view covering the SN and LC. Structural imaging will include a 1.0 mm isotropic T1w sequence for co-registration. High-spatial resolution functional Blood Oxygen Level Dependent (BOLD) (fMRI) will be used to map the functional activations of SN and LC in two task-fMRI paradigms. For the assessment of NA activation, a paradigm utilizing auditory stimuli and visual stimuli with arousing content (IAPS, International Affective Picture System) will be administered in two separate sessions; first inside the scanner, and secondly outside the scanner, during which autonomic responses to the stimuli (pupil responses and skin conductance responses) are recorded. In the second paradigm, participants will be gambling on a virtual slot machine with two handle bars in two different colours where one of the handles can be pushed, the other pulled. Participants have MRI-compatible joysticks to pull or push the handles of the virtual slot machine. Depending on the block, the action performed (pulling or pushing the handles, irrespective of their colour) or the stimulus colour (yellow or blue, irrespective of whether they can be pushed or pulled) are predictive of reward and punishment (one option has a higher reward probability than the other, these have to be learnt via feedback). Reward probabilities are changing repeatedly and unannounced.

Patients will undergo the experiments twice on two separate days, on and off PD medication, in a randomized, balanced order. Healthy participants will also undergo the experiments on two separate days.