Search

Background: Diffusion kurtosis imaging provides in vivo measurement of microstructural tissue characteristics and could help guide management of Parkinson’s disease. Objective: To investigate longitudinal diffusion kurtosis imaging changes on magnetic resonance imaging in the deep grey nuclei in people with early Parkinson’s disease over two years, and whether they correlate with disease progression. Methods: We conducted a longitudinal case-control study of early Parkinson’s disease. 262 people (Parkinson’s disease: n = 185, aged 67.5±9.1 years; 43% female; healthy controls: n = 77, aged 66.6±8.1 years; 53% female) underwent diffusion kurtosis imaging and clinical assessment at baseline and two-year timepoints. We automatically segmented five nuclei, comparing the mean kurtosis and other diffusion kurtosis imaging indices between groups and over time using repeated-measures analysis of variance, and Pearson correlation with the two-year change in Movement Disorder Society Unified Parkinson’s Disease Rating Scale Part III. Results: At baseline, mean kurtosis was higher in Parkinson’s disease than controls in the substantia nigra, putamen, thalamus and globus pallidus when adjusting for age, sex, and levodopa equivalent daily dose (p < 0.027). These differences grew over two years, with mean kurtosis increasing for the Parkinson’s disease group while remaining stable for the control group; evident in significant “group ×time” interaction effects for the putamen, thalamus and globus pallidus ( η p 2 = 0.08–0.11, p < 0.015). However, we did not detect significant correlations between increasing mean kurtosis and declining motor function in the Parkinson’s disease group. Conclusion: Diffusion kurtosis imaging of specific grey matter structures shows abnormal microstructure in PD at baseline and abnormal progression in PD over two years.

Parkinson's disease (PD) is a chronic, non-reversible neurodegenerative disorder, and freezing of gait (FOG) is one of the most disabling symptoms in PD as it is often the leading cause of falls and injuries that drastically reduces patients' quality of life. In order to monitor continuously and objectively PD patients who suffer from FOG and enable the possibility of on-demand cueing assistance, a sensor-based FOG detection solution can help clinicians manage the disease and help patients overcome freezing episodes. Many recent studies have leveraged deep learning models to detect FOG using signals extracted from inertial measurement unit (IMU) devices. Usually, the latent features and patterns of FOG are discovered from either the time or frequency domain. In this study, we investigated the use of the time-frequency domain by applying the Continuous Wavelet Transform to signals from IMUs placed on the lower limbs of 63 PD patients who suffered from FOG. We built convolutional neural networks to detect the FOG occurrences, and employed the Bayesian Optimisation approach to obtain the hyper-parameters. The results showed that the proposed subject-independent model was able to achieve a geometric mean of 90.7% and a F1 score of 91.5%.

Introduction Subjective cognitive complaints (SCC) and affective symptoms are highly prevalent in Parkinson's Disease (PD). In early PD, SCC prevalence and its affective correlates, using recommended Movement Disorders Society (MDS) Level II Criteria to define the underlying cognitive impairment, has not been previously explored. Methods We recruited 121 participants with early PD from two tertiary hospitals in Singapore. The presence of SCC was defined using a Non-Motor Symptoms Scale Domain-5 Score ≥1. Comprehensive neuropsychological testing was conducted with Mild Cognitive Impairment (PD-MCI) defined using recommended MDS Level II Criteria. Affective symptoms were assessed using the Hospital Anxiety Depression Scale (HADS), Geriatric Depression Scale (GDS) and Apathy Scale (AS). Analysis using multivariable linear regression model was performed. Results In our early PD cohort, SCC prevalence independent of underlying cognitive status was 38.8%. Prevalence of SCC in cognitively impaired and cognitively normal participants was 10.7% and 28.1% respectively (р = 0.241). In cognitively normal PD participants, multivariable linear regression analysis revealed that SCC was significantly associated with anxiety (β = 0.28, 95% CI = 0.09–0.79, p = 0.014), depression (β = 0.31, 95% CI = 0.10–0.59, p = 0.006) and apathy (β = 0.32, 95% CI = 1.15–5.98, p = 0.004). Such an association was not found in cognitively impaired PD participants. Conclusion SCC is highly prevalent even in early PD. Its implications in early PD differ depending on underlying cognitive status. SCC in cognitively impaired participants underestimates the true prevalence of PD-MCI. In contrast, SCC in cognitively normal participants is suggestive of an underlying affective disorder.

Brain donations are imperative for research; understanding possible barriers to entry is required to improve brain donation rates. While a few surveys have studied attitudes towards brain banking in patients with neurodegenerative disorders, none have surveyed patients with chronic neurological disorders but without neurodegeneration. This cross-sectional study was conducted on 187 participants, with both neurodegenerative (n = 122) and non-neurodegenerative disorders (n = 65), to compare their attitudes and preferences towards brain donation. Encouragingly, patients with non-neurodegenerative disorders were just as likely to consider brain donation as those with neurodegenerative diseases. Approximately half of each group were willing to consider brain donation, and majority of participants across both groups would not be offended if asked to participate in brain donation (71%). Across both groups, altruistic reasons such as desire to advance medical knowledge and benefit to other patients were the main motivating factors for brain donation, while perceived stress for family members, fears of body disfigurement and religious reasons were the main reasons against brain donation. Of note, nearly two-thirds of all participants were agreeable to allow their family to decide on their behalf. Overall, participants with non-neurodegenerative disorders appeared equally likely to consider brain donation as participants with neurodegenerative disorders. This is an important finding as they represent a significant population seen in specialist neurology clinics who may be overlooked in brain donor recruitment and awareness efforts. Healthcare professionals involved in brain banking should consider actively approaching these potential donors and involving their family members in these discussions.

Parkinson’s disease (PD) is characterized by Lewy bodies containing α-synuclein and ubiquitin aggregates, their co-occurrence possibly linked to a failure of the ubiquitin proteasome system. Ubiquitin C-terminal hydrolase L1 (UCHL1) plays an important role in maintenance of nervous system integrity, and overexpression of UCHL1 has been shown to increase ubiquitin levels within neurons. While cerebrospinal fluid ubiquitin levels were reported to be lower in PD vs controls, plasma UCHL1 levels and their relationship with clinical measures in PD has not been reported. We measured plasma UCHL1 levels using single molecule array (Simoa) in 291 subjects (242 PD and 49 healthy controls, HC). We found that UCHL1 levels were significantly higher in PD patients at moderate stages (Hoehn and Yahr, H&Y stage >2) vs milder PD (H&Y ≤2, p

Graphical abstract, Highlights • Putamen susceptibility value was higher in PD than controls one year into diagnosis. • Putamen susceptibility value was associated with clinical motor scores in early PD. • Mean diffusivity revealed greater cellular loss in the lateral substantial nigra. • Putamen and caudate microstructural degradation were driven by radial diffusivity. • A composite putamen-caudate DKI-QSM marker classified early PD from controls., Objective To investigate the utility of quantitative susceptibility mapping (QSM) and diffusion kurtosis imaging (DKI) as complementary tools in characterizing pathological changes in the deep grey nuclei in early Parkinson’s disease (PD) and their clinical correlates to aid in diagnosis of PD. Method Patients with a diagnosis of PD made within a year and age-matched healthy controls were recruited. All participants underwent clinical evaluation using the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS III) and Hoehn & Yahr stage (H&Y), and brain 3 T MRI including QSM and DKI. Regions-of-interest (ROIs) in the caudate nucleus, putamen, globus pallidus, and medial and lateral substantia nigra (SN) were manually drawn to compare the mean susceptibility (representing iron deposition) and DKI indices (representing restricted water diffusion) between PD patients and healthy controls and in correlation with MDS-UPDRS III and H&Y, focusing on susceptibility value, mean diffusivity (MD) and mean kurtosis (MK). Results There were forty-seven PD patients (aged 68.7 years, 51% male, disease duration 0.78 years) and 16 healthy controls (aged 67.4 years, 63% male). Susceptibility value was increased in PD in all ROIs except the caudate, and was significantly different after multiple comparison correction in the putamen (PD: 64.75 ppb, HC: 44.61 ppb, p = 0.004). MD was significantly higher in PD in the lateral SN, putamen and caudate, the regions with the lowest susceptibility value. In PD patients, we found significant association between the MDS-UPDRS III score and susceptibility value in the putamen after correcting for age and sex (β = 0.21, p = 0.003). A composite DKI-QSM diagnostic marker based on these findings successfully differentiated the groups (p

Freezing of Gait is the most disabling gait disturbance in Parkinson’s disease. For the past decade, there has been a growing interest in applying machine learning and deep learning models to wearable sensor data to detect Freezing of Gait episodes. In our study, we recruited sixty-seven Parkinson’s disease patients who have been suffering from Freezing of Gait, and conducted two clinical assessments while the patients wore two wireless Inertial Measurement Units on their ankles. We converted the recorded time-series sensor data into continuous wavelet transform scalograms and trained a Convolutional Neural Network to detect the freezing episodes. The proposed model achieved a generalisation accuracy of 89.2% and a geometric mean of 88.8%.

Background The main motor subtypes of Parkinson’s disease (PD) include tremor-dominant (TD) and postural instability gait disorder (PIGD), with varying disease course that warrant the development of biomarkers capable of predicting progression according to motor subtype. The PIGD subtype is associated with a poorer prognosis, hence identification of a biomarker associated with PIGD is clinically relevant. Neurofilament light (NfL) chain is a potential biomarker of disease severity in neurological disorders including PD. However, no study has investigated NfL and PD motor subtypes. Here, we aimed to investigate the diagnostic and prognostic utility of plasma NfL for PD motor subtypes in early Parkinson’s disease. Given the higher risk for cognitive and motor decline in PIGD, we hypothesized that plasma NfL is a potential biomarker for PIGD. Methods Plasma NfL was measured in 199 participants (149 PD and 50 healthy controls, HC) using an ultrasensitive single molecule array. Patients were classified into TD or PIGD based on MDS-UPDRS components. After 2 years, 115 patients were reassessed. Association between NfL and clinical measures in PIGD and TD at baseline and at 2-year follow-up were analysed. Results At baseline, plasma NfL levels were higher in PD than HC (8.8 ± 3.4 vs 16.2 ± 7.6 pg/ml, p p p = 0.039). Within the PIGD group, higher NfL associated significantly with worse global cognition and UPDRS motor scores at baseline, and was able to predict motor and cognitive decline at a mean follow-up duration of 1.9 years, controlled for age, sex and disease duration. Conclusions In this longitudinal study, we demonstrated for the first time the potential utility of plasma NfL as a diagnostic and prognostic biomarker in PIGD even at early stages of PD. These important novel findings will require further confirmation in larger, longitudinal PD cohorts.

Objective Uric acid has been found to be potentially neuroprotective in Parkinson's disease (PD). We investigated the relationship between serum uric acid levels and both motor and non-motor features in a prospective early PD cohort study. Methods Fasting serum uric acid levels were measured from 125 early PD patients. Demographic, clinical characteristics, motor and non-motor assessments were performed. Patients were categorized into three motor subtypes: tremor-dominant (TD), postural instability/gait difficulty (PIGD), and mixed. Non-motor symptoms were classified as present or absent based on the appropriate cut-offs for each non-motor instrument. Results Most patients had TD (n = 51, 40.8%) and mixed (n = 63, 50.4%) motor subtypes, while a minority had PIGD (n = 11, 8.8%) motor subtype. The mean serum uric acid levels were significantly different between the three motor subtypes (p = 0.0106), with the mixed subtype having the lowest serum uric acid levels. Using the TD subtype as reference, patients with higher serum uric acid levels were less likely to have the mixed (OR = 0.684; p = 0.0312) subtype as opposed to the TD subtype. Uric acid levels were not significantly different between the TD and PIGD subtypes. For non-motor symptoms, higher serum uric acid levels were significantly associated with less fatigue (OR = 0.693; p = 0.0408). Conclusion Higher serum uric acid levels were associated with TD motor subtype and less fatigue in early PD, which could be related to its anti-oxidative properties. Uric acid could be an important biomarker for specific motor features and symptoms of fatigue in PD.

Background and purpose This study quantified the total brain and periventricular white matter hyperintensity (WMH) burdens in patients with early Parkinson's disease (PD) and explored their associations with cardiovascular risk factors and cognitive performance. Methods A total of 175 non-demented patients with early PD who had undergone baseline brain magnetic resonance imaging were included. Comprehensive neurocognitive testing was conducted to identify PD with mild cognitive impairment (PD-MCI) and to evaluate performances in individual cognitive domains. Cardiovascular risk was expressed as a modified Framingham 10-year cardiovascular risk score (mFRS). Results A total of 53.7% of this early PD cohort fulfilled the diagnostic criteria for PD-MCI. An increase in mFRS was significantly associated with increases in the total brain WMH (P = 0.015) and periventricular WMH (P = 0.040) burden, independent of age and gender. The periventricular WMH burden was significantly associated with PD-MCI (P = 0.046) in early PD, independent of cardiovascular risk factors. Patients in the 5th quintile of periventricular WMH burden were 8.6 times more likely to have PD-MCI compared with patients in the 1st quintile of periventricular WMH burden (P = 0.004). However, total brain WMH burden was not associated with PD-MCI (P = 0.158). In individual cognitive domains, heavier periventricular WMH burden was associated with worse executive function and visuospatial function independent of cardiovascular risk factors. Conclusion Periventricular WMHs are a useful imaging biomarker for cognitive impairment in early PD. Cardiovascular risk factors, although associated with periventricular WMHs, were unable to fully explain the association between periventricular WMHs and cognitive impairment in early PD.

Freezing of Gait (FoG) is a common motor-related impairment among Parkinson's disease patients, which substantially reduces their quality of life and puts them at risk of falls. These patients benefit from wearable FoG detection systems that provide timely biofeedback cues and hence help them regain control over their gait. Unfortunately, the systems proposed thus far are bulky and obtrusive when worn. The objective of this paper is to demonstrate the first integration of an FoG detection system into a single sensor node. To achieve such an integration, features with low computational load are selected and dedicated hardware is designed that limits area and memory utilization. Classification is achieved with a neural network that is capable of learning in real time and thus allows the system to adapt to a patient during run-time. A small form factor FPGA implements the feature extraction and classification, whereas a custom PCB integrates the system into a single node. The system fits into a 4.5 × 3.5 × 1.5 cm $^3$ housing case, weighs 32 g, and achieves 95.6% sensitivity and 90.2% specificity when adapted to a patient. Biofeedback cues are provided either through auditory or somatosensory means and the system can remain operational for longer than 9 h while providing cues. The proposed system is highly competitive in terms of classification performance and excels with respect to wearability and real-time patient adaptivity.

BACKGROUND While the association between alpha-synuclein gene promoter (Rep1) variability and risk of PD is well established, its association with cognition is unclear. OBJECTIVES To investigate the association between Rep1 and motor and cognitive outcomes in PD. METHODS Rep1 allele lengths were determined in 172 PD patients who were grouped into "long" and "short" carriers according to previous methods. Multivariable regression analysis was performed to investigate the effect of Rep1 length on cognitive and motor scores. RESULTS Long Rep1 allele carriers had significantly lower MMSE (P = 0.010) and higher UPDRS Part III (P = 0.026) and H & Y (P = 0.008) scores compared to short allele carriers (controlled for age, sex, and disease duration). Interaction analyses of Rep1 with apolipoprotein 4 revealed no significant effect on clinical outcomes. CONCLUSIONS PD patients carrying long Rep1 alleles are more impaired on cognitive and motor function independent of apolipoprotein 4 genotype. © 2019 International Parkinson and Movement Disorder Society.

Parkinson’s disease patients are at risk of falls due to freezing of gait (FoG). Wearable detection systems providing biofeedback for aid rely on accurate FoG classifiers, but such algorithms have yet to propose a dedicated hardware implementation. This paper is a first proposal of a dedicated hardware for a real-time FoG feature extractor and classifier on a single chip. The neural network classifier and FoG feature extractor exploit an FoG system’s inherent time-sharing affinity to reduce area consumption which is crucial in wearable applications. The design has been fitted onto a small form-factor 10×10 mm2 Artix-7 FPGA. It consumes 6.716 μJ per classification which surpasses non-dedicated FoG system implementations in terms of efficiency, while remaining competitive with 91.32% and 93.64% in sensitivity and specificity, respectively.

We utilized ultrasensitive single molecule technology to measure plasma alpha‐synuclein in 221 subjects (51 controls, 170 PD). Plasma alpha‐synuclein levels were significantly higher in PD than controls (15506.3 vs. 13057.0 pg/mL, P = 0.037), adjusting for age and gender. In PD, alpha‐synuclein levels did not vary by H&Y stage or UPDRS motor scores but were significantly higher in PD patients with poorer cognition (MMSE ≤ 25) than controls (P = 0.016, Bonferroni corrected P = 0.047). Alpha‐synuclein levels quantified using ultrasensitive single molecule technology discriminate PD from controls and correlate with cognitive severity. These preliminary findings require independent validation to determine the utility of this assay.

Freezing of gait (FoG) is a sudden and episodic inability to generate effective stepping among Parkinson's disease patients. It poses a risk for falls and deteriorates a patient's quality of life. Aid is sought after by the implementation of wearable systems that detect FoG and provide biofeedback cues in real-time. Detection is predominantly attempted with patient-independent classifiers which have difficulties to account for some patient's inimitable walking styles. Such gait peculiarities can be addressed with patient-adaptive classifiers. However, the patient-specific adaptations proposed thus far are retrospective and require a patient's labeled data. We propose to provide patient adaptivity in real-time through semi-supervised neural networks which exploit the stream of unlabeled data generated during usage. Using supervised learning, a patient-independent neural network is designed to serve as a base model. Upon a new patient's utilization of the system, the base model's parameters are adapted in real-time through unsupervised learning from the generated stream of unlabeled data. On average, patient adaptivity augmented sensitivity by 4.58% for the price of 0.59% in specificity. Moreover, it accounted for inimitable walking styles of patients that had been inadequately classified by the patient-independent base model. For such patients, sensitivity increased up to 42.01%. The overall patient-adaptive classifier resulted in 95.90% and 93.05% in sensitivity and specificity, respectively.

Freezing of gait (FoG) is a common gait impairment in Parkinson's disease that puts patients at risk of falls and deteriorates their quality of life. Relief is sought after by evaluating the possibility of wearable systems that detect FoG in real-time and provide gait-reinforcing biofeedback cues. The successful detection relies on the extraction of high quality features, which have to be computed from recent samples of an inertial measurement unit in order to ensure real-time applicability. Unfortunately, the amount of samples considered for a feature's computation, i.e. the data window length, has been subjected to widespread disagreement: With no thorough analysis available, employed window lengths differed by several seconds among implementations. We derive optimal window lengths for a broad number of features used throughout literature by using mutual information as an evaluation metric, and elaborate on a window length's significance in affecting classification performance. With conventional feature selection methods, feature subsets tailored to various machine learning algorithms are established. Relying on these feature subsets for FoG classification, whereby all features are extracted with optimal window lengths, F1-scores increase up to 17.1% for individual classifiers and up to 12.7% on average when compared to previously proposed feature sets that are extracted with sub-optimal window lengths.