The Parkinson Progression Marker Initiative (PPMI)

Assessment of U-Net in the segmentation of short tracts: Transferring to clinical MRI routine

Accurately studying structural connectivity requires precise tract segmentation strategies. The U-Net network has been widely recognized for its exceptional capacity in image segmentation tasks and provides remarkable results in large tract segmentation when high-quality diffusion-weighted imaging (DWI) data are used. However, short tracts, which are associated with various neurological diseases, pose specific challenges, particularly when high-quality DWI data acquisition within clinical settings is concerned. Here, we aimed to evaluate the U-Net network ability to segment short tracts by using DWI data acquired in different experimental conditions. To this end, we conducted three types of training experiments involving 350 healthy subjects and 11 white matter tracts, including the anterior, posterior, and hippocampal commissure, fornix, and uncinated fasciculus. In the first experiment, the model was exclusively trained with high-quality data of the Human Connectome Project (HCP) dataset. The second experiment focused on images of healthy subjects acquired from a local hospital dataset, representing a typical clinical routine acquisition. In the third experiment, a hybrid training approach was employed, combining data of the HCP and local hospital datasets. Then, the best model was also tested in unseen DWIs of 10 epilepsy patients of the local hospital and 10 healthy subjects acquired on a scanner from another company. The outcomes of the third experiment demonstrated a notable enhancement in performance when contrasted with the preceding trials. Specifically, the short tracts within the local hospital dataset achieved Dice scores ranging between 0.60 and 0.65. Similar intervals were obtained with HCP data in the first experiment, and a substantial improvement compared to the scores between 0.37 and 0.50 obtained with the local hospital dataset at the same experiment. This improvement persisted when the method was applied to diverse scenarios, including different scanner acquisitions and epilepsy patients. These results indicate that combining datasets from different sources, coupled with resolution standardization strengthens the neural network ability to generalize predictions across a spectrum of datasets. Nevertheless, short tract segmentation performance is intricately linked to the training composition, to validation, and to testing data. Moreover, curved tracts have intricate structural nature, which adds complexities to their segmenting. Although the network training approach tested herein has provided promising results, caution must be taken when extrapolating its application to datasets acquired under distinct experimental conditions, even in the case of higher-quality data or analysis of long or short tracts.

Analysis of cognitive dysfunction in Parkinson's disease using voxel based morphometry and radiomics

Cognitive impairment in Parkinson's disease (PD) is associated with changes in the brain anatomical structures. The objective of this study, is to identify the atrophy patterns based on the severity of cognitive decline and evaluate the disease progression. In this study, gray matter alterations are analysed in 135 PD subjects under 3 cognitive domains (91 Cognitively normal PD (NC-PD), 25 PD with Mild Cognitive Impairment (PD-MCI) and 19 PD with Dementia (PD-D)) by comparing them with 58 Healthy Control (HC) subjects. Voxel Based Morphometry (VBM) is used to segment the gray matter regions in magnetic resonance images and analyse the atrophy patterns statistically. Significant patterns of gray matter variations observed in the middle temporal and medial frontal region differentiate between HC and PD subject groups based on the severity of cognitive decline. Abnormalities in gray matter is substantiated through radiomic features extracted from the significant gray matter clusters. Significant radiomic features of the clusters are able to differentiate between the HC and PD-D subjects with an accuracy of 81.82%. Higher atrophy levels identified in PD-D subjects compared to NC-PD and PD-MCI group enables early diagnosis and treatment procedures. The combined and comprehensive analysis of gray matter alterations through VBM and radiomic features gives better assessment of cognitive impairment in PD.

Neutrophil-to-lymphocyte ratio and longitudinal cognitive performance in Parkinson's disease

OBJECTIVE

Previous studies have suggested a link between peripheral inflammation and cognitive outcomes in the general population and individuals with Parkinson's disease (PD). We sought to test the association between peripheral inflammation, measured by the neutrophil-to-lymphocyte ratio (NLR), cognitive performance, and mild cognitive impairment (MCI) status in individuals with PD.

METHODS

A retrospective, longitudinal analysis was carried out using data from the Parkinson's Progression Markers Initiative (PPMI), including 422 participants with PD followed over 5 years. Cognitive performance was assessed using a neuropsychological battery including the Montreal Cognitive Assessment (MoCA) and tests of verbal learning, visuospatial function, processing speed, and executive function. Mixed-effect regression models were used to analyze the association between NLR, cognitive performance, and MCI status, controlling for age, sex, education, APOE genotype, and motor severity.

RESULTS

There was a negative association between NLR and MoCA, even after adjusting for covariates (b = -0.12, p = 0.033). MoCA scores for individuals in the high NLR category exhibited a more rapid decline over time compared to the low NLR group (b = -0.16, p = 0.012). Increased NLR was associated with decreased performance across all cognitive domains. However, NLR was not associated with MCI status over 5 years of follow-up.

INTERPRETATION

This study demonstrates a link between elevated NLR and cognitive performance in PD, but not with MCI status over 5 years. This suggests that NLR is more strongly associated with day-to-day cognitive performance than with incident MCI, but this requires further study in more heterogeneous cohorts.

Radiomics incorporating deep features for predicting Parkinson's disease in 123I-Ioflupane SPECT

PURPOSE

123I-Ioflupane SPECT is an effective tool for the diagnosis and progression assessment of Parkinson's disease (PD). Radiomics and deep learning (DL) can be used to track and analyze the underlying image texture and features to predict the Hoehn-Yahr stages (HYS) of PD. In this study, we aim to predict HYS at year 0 and year 4 after the first diagnosis with combined imaging, radiomics and DL-based features using 123I-Ioflupane SPECT images at year 0.

METHODS

In this study, 161 subjects from the Parkinson's Progressive Marker Initiative database underwent baseline 3T MRI and 123I-Ioflupane SPECT, with HYS assessment at years 0 and 4 after first diagnosis. Conventional imaging features (IF) and radiomic features (RaF) for striatum uptakes were extracted from SPECT images using MRI- and SPECT-based (SPECT-V and SPECT-T) segmentations respectively. A 2D DenseNet was used to predict HYS of PD, and simultaneously generate deep features (DF). The random forest algorithm was applied to develop models based on DF, RaF, IF and combined features to predict HYS (stage 0, 1 and 2) at year 0 and (stage 0, 1 and ≥ 2) at year 4, respectively. Model predictive accuracy and receiver operating characteristic (ROC) analysis were assessed for various prediction models.

RESULTS

For the diagnostic accuracy at year 0, DL (0.696) outperformed most models, except DF + IF in SPECT-V (0.704), significantly superior based on paired t-test. For year 4, accuracy of DF + RaF model in MRI-based method is the highest (0.835), significantly better than DF + IF, IF + RaF, RaF and IF models. And DL (0.820) surpassed models in both SPECT-based methods. The area under the ROC curve (AUC) highlighted DF + RaF model (0.854) in MRI-based method at year 0 and DF + RaF model (0.869) in SPECT-T method at year 4, outperforming DL models, respectively. And then, there was no significant differences between SPECT-based and MRI-based segmentation methods except for the imaging feature models.

CONCLUSION

The combination of radiomic and deep features enhances the prediction accuracy of PD HYS compared to only radiomics or DL. This suggests the potential for further advancements in predictive model performance for PD HYS at year 0 and year 4 after first diagnosis using 123I-Ioflupane SPECT images at year 0, thereby facilitating early diagnosis and treatment for PD patients. No significant difference was observed in radiomics results obtained between MRI- and SPECT-based striatum segmentations for radiomic and deep features.

Identification of Parkinson's disease PACE subtypes and repurposing treatments through integrative analyses of multimodal data

Parkinson's disease (PD) is a serious neurodegenerative disorder marked by significant clinical and progression heterogeneity. This study aimed at addressing heterogeneity of PD through integrative analysis of various data modalities. We analyzed clinical progression data (≥5 years) of individuals with de novo PD using machine learning and deep learning, to characterize individuals' phenotypic progression trajectories for PD subtyping. We discovered three pace subtypes of PD exhibiting distinct progression patterns: the Inching Pace subtype (PD-I) with mild baseline severity and mild progression speed; the Moderate Pace subtype (PD-M) with mild baseline severity but advancing at a moderate progression rate; and the Rapid Pace subtype (PD-R) with the most rapid symptom progression rate. We found cerebrospinal fluid P-tau/α-synuclein ratio and atrophy in certain brain regions as potential markers of these subtypes. Analyses of genetic and transcriptomic profiles with network-based approaches identified molecular modules associated with each subtype. For instance, the PD-R-specific module suggested STAT3, FYN, BECN1, APOA1, NEDD4, and GATA2 as potential driver genes of PD-R. It also suggested neuroinflammation, oxidative stress, metabolism, PI3K/AKT, and angiogenesis pathways as potential drivers for rapid PD progression (i.e., PD-R). Moreover, we identified repurposable drug candidates by targeting these subtype-specific molecular modules using network-based approach and cell line drug-gene signature data. We further estimated their treatment effects using two large-scale real-world patient databases; the real-world evidence we gained highlighted the potential of metformin in ameliorating PD progression. In conclusion, this work helps better understand clinical and pathophysiological complexity of PD progression and accelerate precision medicine.

Biomarkers of Parkinson's disease in perspective of early diagnosis and translation of neurotrophic therapies

Parkinson's disease (PD) is a common neurodegenerative disorder characterized by progressive loss of dopamine neurons and aberrant deposits of alpha-synuclein (α-syn) in the brain. The symptomatic treatment is started after the onset of motor manifestations in a late stage of the disease. Preclinical studies with neurotrophic factors (NTFs) show promising results of disease-modifying neuroprotective or even neurorestorative effects. Four NTFs have entered phase I-II clinical trials with inconclusive outcomes. This is not surprising because the preclinical evidence is from acute early-stage disease models, but the clinical trials included advanced PD patients. To conclude the value of NTF therapies, clinical studies should be performed in early-stage patients with prodromal symptoms, that is, before motor manifestations. In this review, we summarize currently available diagnostic and prognostic biomarkers that could help identify at-risk patients benefiting from NTF therapies. Focus is on biochemical and imaging biomarkers, but also other modalities are discussed. Neuroimaging is the most important diagnostic tool today, but α-syn imaging is not yet viable. Modern techniques allow measuring various forms of α-syn in cerebrospinal fluid, blood, saliva, and skin. Digital biomarkers and artificial intelligence offer new means for early diagnosis and longitudinal follow-up of degenerative brain diseases.

AI-based differential diagnosis of dementia etiologies on multimodal data

Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an artificial intelligence (AI) model that harnesses a broad array of data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations and multimodal neuroimaging, to identify the etiologies contributing to dementia in individuals. The study, drawing on 51,269 participants across 9 independent, geographically diverse datasets, facilitated the identification of 10 distinct dementia etiologies. It aligns diagnoses with similar management strategies, ensuring robust predictions even with incomplete data. Our model achieved a microaveraged area under the receiver operating characteristic curve (AUROC) of 0.94 in classifying individuals with normal cognition, mild cognitive impairment and dementia. Also, the microaveraged AUROC was 0.96 in differentiating the dementia etiologies. Our model demonstrated proficiency in addressing mixed dementia cases, with a mean AUROC of 0.78 for two co-occurring pathologies. In a randomly selected subset of 100 cases, the AUROC of neurologist assessments augmented by our AI model exceeded neurologist-only evaluations by 26.25%. Furthermore, our model predictions aligned with biomarker evidence and its associations with different proteinopathies were substantiated through postmortem findings. Our framework has the potential to be integrated as a screening tool for dementia in clinical settings and drug trials. Further prospective studies are needed to confirm its ability to improve patient care.

Radiomics score derived from T1-w/T2-w ratio image can predict motor symptom progression in Parkinson's disease

OBJECTIVES

To clarify the association between a radiomics score (Rad-score) derived from T1-weighted signal intensity to T2-weighted signal intensity (T1-w/T2-w) ratio images and the progression of motor symptoms in Parkinson's disease (PD).

MATERIALS AND METHODS

This retrospective study included patients with PD enrolled in the Parkinson's Progression Markers Initiative. The Movement Disorders Society-Unified Parkinson's Disease Rating Scale Part III score ≥ 33 and/or Hoehn and Yahr stage ≥ 3 indicated motor function decline. The Rad-score was constructed using radiomics features extracted from T1-w/T2-w ratio images. The Kaplan-Meier analysis and Cox regression analyses were used to assess the time differences in motor function decline between the high and low Rad-score groups.

RESULTS

A total of 171 patients with PD were divided into training (n = 101, mean age at baseline, 61.6 ± 9.3 years) and testing (n = 70, mean age at baseline, 61.6 ± 10 years). The patients in the high Rad-score group had a shorter time to motor function decline than those in the low Rad-score group in the training dataset (log-rank test, p < 0.001) and testing dataset (log-rank test, p < 0.001). The multivariate Cox regression using the Rad-score and clinical factors revealed a significant association between the Rad-score and motor function decline in the training dataset (HR = 2.368, 95%CI:1.423-3.943, p < 0.001) and testing dataset (HR = 2.931, 95%CI:1.472-5.837, p = 0.002).

CONCLUSION

Rad-scores based on radiomics features derived from T1-w/T2-w ratio images were associated with the progression of motor symptoms in PD.

CLINICAL RELEVANCE STATEMENT

The radiomics score derived from the T1-weighted/T2-weighted ratio images offers a predictive tool for assessing the progression of motor symptom in patients with PD.

KEY POINTS

Radiomics score derived from T1-weighted/T2-weighted ratio images is correlated with the motor symptoms of Parkinson's disease. A high radiomics score correlated with faster motor function decline in patients with Parkinson's disease. The proposed radiomics score offers predictive insight into the progression of motor symptoms of Parkinson's disease.

Sleep disturbances, cognitive decline, and AD biomarkers alterations in early Parkinson's disease

OBJECTIVE

We aimed to investigate whether each type of sleep disturbances (i.e., pRBD, EDS, and insomnia) is specifically associated with faster decline in global cognition and different cognitive domains among de novo PD patients. We also assessed the influence of sleep disturbances on core AD CSF biomarkers alterations and conversion to dementia.

METHODS

Prospectively longitudinal data were obtained from the PPMI cohort. Sleep disturbances and cognition ability were assessed by questionnaires at baseline and follow-up visits. Generalized linear mixed models were utilized to assess the effect of sleep disturbances on cognitive decline and core AD CSF biomarkers change. The associations between sleep disturbances and conversion to dementia were analyzed using Cox regression analysis.

RESULTS

Baseline pRBD was associated with faster decline in global cognition and all cognitive domains, including verbal episodic memory, visuospatial ability, executive function, language, and processing speed. EDS was associated with faster decline in three cognitive domains, including verbal episodic memory, executive function/working memory, and processing speed. Insomnia was associated with faster decline in global cognition and verbal episodic memory. Meanwhile, pRBD and EDS were associated with longitudinal decrease of CSF Aβ42. Baseline pRBD increased the risk of conversion to dementia. The risk of dementia in PD patients with multiple sleep disturbances also increased compared with those without sleep disturbance.

INTERPRETATION

Sleep disturbances (i.e., pRBD, EDS, and insomnia) were associated with cognitive decline in early PD. EDS and pRBD were associated with decrease of CSF Aβ42. Moreover, pRBD was associated with conversion to dementia.

Assessment tools for cognitive performance in Parkinson's disease and its genetic contributors

Background

We have shown that genetic factors associating with motor progression of Parkinson's disease (PD), but their roles in cognitive function is poorly understood. One reason is that while cognitive performance in PD can be evaluated by various cognitive scales, there is no definitive guide indicating which tool performs better.

Methods

Data were obtained from the Parkinson's Progression Markers Initiative, where cognitive performance was assessed using five cognitive screening tools, including Symbol Digit Modalities Test (SDMT), Montreal Cognitive Assessment, Benton Judgment of Line Orientation, Modified Semantic Fluency Test, and Letter Number Sequencing Test, at baseline and subsequent annual follow-up visit for 5 years. Genetic data including ApoE and other PD risk genetic information were also obtained. We used SPSS-receiver operating characteristic and ANOVA repeated measures to evaluate which cognitive assessment is the best reflecting cognitive performance in PD at early stage and over time. Logistic regression analyses were used to determine the genetic associations with the rapidity of cognitive decline in PD.

Results

SDMT performed better in detecting mild cognitive impairment at baseline (AUC = 0.763), and SDMT was the only tool showing a steady cognitive decline during longitudinal observation. Multigenetic factors significantly associated with cognitive impairment at early stage of the disease (AUC = 0.950) with IP6K2 rs12497850 more evident, and a significantly faster decline (AUC = 0.831) within 5 years after motor onset, particularly in those carrying FGF20 rs591323.

Conclusion

SDMT is a preferable cognitive assessment tool for PD and genetic factors synergistically contribute to the cognitive dysfunction in PD.

An improved method for diagnosis of Parkinson's disease using deep learning models enhanced with metaheuristic algorithm

Parkinson's disease (PD) is challenging for clinicians to accurately diagnose in the early stages. Quantitative measures of brain health can be obtained safely and non-invasively using medical imaging techniques like magnetic resonance imaging (MRI) and single photon emission computed tomography (SPECT). For accurate diagnosis of PD, powerful machine learning and deep learning models as well as the effectiveness of medical imaging tools for assessing neurological health are required. This study proposes four deep learning models with a hybrid model for the early detection of PD. For the simulation study, two standard datasets are chosen. Further to improve the performance of the models, grey wolf optimization (GWO) is used to automatically fine-tune the hyperparameters of the models. The GWO-VGG16, GWO-DenseNet, GWO-DenseNet + LSTM, GWO-InceptionV3 and GWO-VGG16 + InceptionV3 are applied to the T1,T2-weighted and SPECT DaTscan datasets. All the models performed well and obtained near or above 99% accuracy. The highest accuracy of 99.94% and AUC of 99.99% is achieved by the hybrid model (GWO-VGG16 + InceptionV3) for T1,T2-weighted dataset and 100% accuracy and 99.92% AUC is recorded for GWO-VGG16 + InceptionV3 models using SPECT DaTscan dataset.

Transcriptional pathobiology and multi-omics predictors for Parkinson's disease

Early diagnosis and biomarker discovery to bolster the therapeutic pipeline for Parkinson's disease (PD) are urgently needed. In this study, we leverage the large-scale whole-blood total RNA-seq dataset from the Accelerating Medicine Partnership in Parkinson's Disease (AMP PD) program to identify PD-associated RNAs, including both known genes and novel circular RNAs (circRNA) and enhancer RNAs (eRNAs). There were 1,111 significant marker RNAs, including 491 genes, 599 eRNAs, and 21 circRNAs, that were first discovered in the PPMI cohort (FDR < 0.05) and confirmed in the PDBP/BioFIND cohorts (nominal p < 0.05). Functional enrichment analysis showed that the PD-associated genes are involved in neutrophil activation and degranulation, as well as the TNF-alpha signaling pathway. We further compare the PD-associated genes in blood with those in post-mortem brain dopamine neurons in our BRAINcode cohort. 44 genes show significant changes with the same direction in both PD brain neurons and PD blood, including neuroinflammation-associated genes IKBIP, CXCR2, and NFKBIB. Finally, we built a novel multi-omics machine learning model to predict PD diagnosis with high performance (AUC = 0.89), which was superior to previous studies and might aid the decision-making for PD diagnosis in clinical practice. In summary, this study delineates a wide spectrum of the known and novel RNAs linked to PD and are detectable in circulating blood cells in a harmonized, large-scale dataset. It provides a generally useful computational framework for further biomarker development and early disease prediction.

Dopa Responsiveness in Parkinson's Disease

BACKGROUND

Dopaminergic responsiveness is a defining feature of Parkinson's disease (PD). However, there is limited information on how this evolves over time.

OBJECTIVES

To examine serial dopaminergic responses, if there are distinct patterns, and which factors predict these.

METHODS

We analyzed data from the Parkinson's Progression Markers Initiative on repeated dopaminergic challenge tests (≥24.5% defined as a definite response). Growth-mixture modeling evaluated for different response patterns and multinomial logistic regression tested for predictors of these clusters.

RESULTS

1525 dopaminergic challenge tests were performed in 336 patients. At enrolment, mean age was 61.2 years (SD 9.6), 66.4% were male and disease duration was 0.5 years (SD 0.5). 1 to 2 years after diagnosis, 48.0% of tests showed a definite response, but this proportion increased with longer disease duration (51.1-74.3%). We identified 3 response groups: "Striking" (n = 29, 8.7%); "Excellent" (n = 110; 32.7%) and "Modest" (n = 197, 58.6%). Significant differences were as follows: striking responders commenced treatment earlier (P = 0.02), were less likely to be on dopamine agonist monotherapy (P = 0.01), and had better cognition (P < 0.01) and activities of daily living (P = 0.01). Excellent responders had higher challenge doses (P = 0.03) and were more likely to be on combination therapy (P < 0.01).

CONCLUSION

Three distinct patterns of the dopaminergic response were observed. As the proportion of PD cases with definite dopa responsiveness increased over time, the initial treatment response may be an unreliable diagnostic aid.

Association analyses of apolipoprotein E genotypes and cognitive performance in patients with Parkinson's disease

BACKGROUND

Cognitive impairment is a common non-motor symptom of Parkinson's disease (PD). The apolipoprotein E (APOE) ε4 genotype increases the risk of Alzheimer's disease (AD). However, the effect of APOEε4 on cognitive function of PD patients remains unclear. In this study, we aimed to understand whether and how carrying APOEε4 affects cognitive performance in patients with early-stage and advanced PD.

METHODS

A total of 119 Chinese early-stage PD patients were recruited. Movement Disorder Society Unified Parkinson's Disease Rating Scale, Hamilton anxiety scale, Hamilton depression scale, non-motor symptoms scale, Mini-mental State Examination, Montreal Cognitive Assessment, and Fazekas scale were evaluated. APOE genotypes were determined by polymerase chain reactions and direct sequencing. Demographic and clinical information of 521 early-stage and 262 advanced PD patients were obtained from Parkinson's Progression Marker Initiative (PPMI).

RESULTS

No significant difference in cognitive performance was found between ApoEε4 carriers and non-carriers in early-stage PD patients from our cohort and PPMI. The cerebrospinal fluid (CSF) Amyloid Beta 42 (Aβ42) level was significantly lower in ApoEε4 carrier than non-carriers in early-stage PD patients from PPMI. In advanced PD patients from PPMI, the BJLOT, HVLT retention and SDMT scores seem to be lower in ApoEε4 carriers without reach the statistical significance.

CONCLUSIONS

APOEε4 carriage does not affect the cognitive performance of early-stage PD patients. However, it may promote the decline of CSF Aβ42 level and the associated amyloidopathy, which is likely to further contribute to the cognitive dysfunction of PD patients in the advanced stage.

Neuroimaging epicenters as potential sites of onset of the neuroanatomical pathology in schizophrenia

Schizophrenia lacks a clear definition at the neuroanatomical level, capturing the sites of origin and progress of this disorder. Using a network-theory approach called epicenter mapping on cross-sectional magnetic resonance imaging from 1124 individuals with schizophrenia, we identified the most likely "source of origin" of the structural pathology. Our results suggest that the Broca's area and adjacent frontoinsular cortex may be the epicenters of neuroanatomical pathophysiology in schizophrenia. These epicenters can predict an individual's response to treatment for psychosis. In addition, cross-diagnostic similarities based on epicenter mapping over of 4000 individuals diagnosed with neurological, neurodevelopmental, or psychiatric disorders appear to be limited. When present, these similarities are restricted to bipolar disorder, major depressive disorder, and obsessive-compulsive disorder. We provide a comprehensive framework linking schizophrenia-specific epicenters to multiple levels of neurobiology, including cognitive processes, neurotransmitter receptors and transporters, and human brain gene expression. Epicenter mapping may be a reliable tool for identifying the potential onset sites of neural pathophysiology in schizophrenia.

Genome-wide determinants of mortality and motor progression in Parkinson's disease

There are 90 independent genome-wide significant genetic risk variants for Parkinson's disease (PD) but currently only five nominated loci for PD progression. The biology of PD progression is likely to be of central importance in defining mechanisms that can be used to develop new treatments. We studied 6766 PD patients, over 15,340 visits with a mean follow-up of between 4.2 and 15.7 years and carried out genome-wide survival studies for time to a motor progression endpoint, defined by reaching Hoehn and Yahr stage 3 or greater, and death (mortality). There was a robust effect of the APOE ε4 allele on mortality in PD. We also identified a locus within the TBXAS1 gene encoding thromboxane A synthase 1 associated with mortality in PD. We also report 4 independent loci associated with motor progression in or near MORN1, ASNS, PDE5A, and XPO1. Only the non-Gaucher disease causing GBA1 PD risk variant E326K, of the known PD risk variants, was associated with mortality in PD. Further work is needed to understand the links between these genomic variants and the underlying disease biology. However, these may represent new candidates for disease modification in PD.

A more objective PD diagnostic model: integrating texture feature markers of cerebellar gray matter and white matter through machine learning

Objective

The purpose of this study is to explore whether machine learning can be used to establish an effective model for the diagnosis of Parkinson's disease (PD) by using texture features extracted from cerebellar gray matter and white matter, so as to identify subtle changes that cannot be observed by the naked eye.

Method

This study involved a data collection period from June 2010 to March 2023, including 374 subjects from two cohorts. The Parkinson's Progression Markers Initiative (PPMI) served as the training set, with control group and PD patients (HC: 102 and PD: 102) from 24 global sites. Our institution's data was utilized as the test set (HC: 91 and PD: 79). Machine learning was employed to establish multiple models for PD diagnosis based on texture features of the cerebellum's gray and white matter. Results underwent evaluation through 5-fold cross-validation analysis, calculating the area under the receiver operating characteristic curve (AUC) for each model. The performance of each model was compared using the Delong test, and the interpretability of the optimized model was further augmented by employing Shapley additive explanations (SHAP).

Results

The AUCs for all pipelines in the validation dataset were compared using FeAture Explorer (FAE) software. Among the models established by Kruskal-Wallis (KW) and logistic regression via Lasso (LRLasso), the AUC was highest using the "one-standard error" rule. 'WM_original_glrlm_GrayLevelNonUniformity' was considered the most stable and predictive feature.

Conclusion

The texture features of cerebellar gray matter and white matter combined with machine learning may have potential value in the diagnosis of Parkinson's disease, in which the heterogeneity of white matter may be a more valuable imaging marker.

Left corticospinal tract could be a biomarker to identify the dual prodromal LRRK2/GBA mutated Parkinson's disease

INTRODUCTION

Prodromal Parkinson's disease (PD) carriers of dual leucine-rich repeat kinase 2 (LRRK2) and glucosylceramidase β (GBA) variants are rare, and their biomarkers are less well developed.

OBJECTIVE

This study aimed to investigate the biomarkers for diagnosing the prodromal phase of LRRK2-GBA-PD (LRRK2-GBA-prodromal).

METHODS

We assessed the clinical and whole-brain white matter microstructural characteristics of 54 prodromal PD carriers of dual LRRK2 (100% M239T) and GBA (95% N409S) variants, along with 76 healthy controls (HCs) from the Parkinson's Progression Markers Initiative (PPMI) cohort.

RESULTS

By analyzing the four values of 100 nodes on 20 fiber bundles, totaling 8000 data points, we identified the smallest p value in the fractional anisotropy (FA) value of the 38th segment of left corticospinal tract (L-CST) with differences between LRRK2-GBA-prodromal and HCs (p = 8.94 × 10-9). The FA value of the 38th node of the L-CST was significantly lower in LRRK2-GBA-prodromal (FA value, 0.65) compared with HCs (FA value, 0.71). The receiver-operating characteristic curve showed a cut-off value of 0.218 for the FA value of L-CST, providing sufficient sensitivity (79.2%) and specificity (72.2%) to distinguish double mutation prodromal PD from the healthy population.

CONCLUSION

L-CST, especially the 38th node, may potentially serve as a biomarker for distinguishing individuals with double mutation prodromal PD from the healthy population.

Epidemiology of Parkinson's Disease: An Update

PURPOSE OF REVIEW

In recent decades, epidemiological understanding of Parkinson disease (PD) has evolved significantly. Major discoveries in genetics and large epidemiological investigations have provided a better understanding of the genetic, behavioral, and environmental factors that play a role in the pathogenesis and progression of PD. In this review, we provide an epidemiological update of PD with a particular focus on advances in the last five years of published literature.

RECENT FINDINGS

We include an overview of PD pathophysiology, followed by a detailed discussion of the known distribution of disease and varied determinants of disease. We describe investigations of risk factors for PD, and provide a critical summary of current knowledge, knowledge gaps, and both clinical and research implications. We emphasize the need to characterize the epidemiology of the disease in diverse populations. Despite increasing understanding of PD epidemiology, recent paradigm shifts in the conceptualization of PD as a biological entity will also impact epidemiological research moving forward and guide further work in this field.

DIMOND: DIffusion Model OptimizatioN with Deep Learning

Diffusion magnetic resonance imaging is an important tool for mapping tissue microstructure and structural connectivity non-invasively in the in vivo human brain. Numerous diffusion signal models are proposed to quantify microstructural properties. Nonetheless, accurate estimation of model parameters is computationally expensive and impeded by image noise. Supervised deep learning-based estimation approaches exhibit efficiency and superior performance but require additional training data and may be not generalizable. A new DIffusion Model OptimizatioN framework using physics-informed and self-supervised Deep learning entitled "DIMOND" is proposed to address this problem. DIMOND employs a neural network to map input image data to model parameters and optimizes the network by minimizing the difference between the input acquired data and synthetic data generated via the diffusion model parametrized by network outputs. DIMOND produces accurate diffusion tensor imaging results and is generalizable across subjects and datasets. Moreover, DIMOND outperforms conventional methods for fitting sophisticated microstructural models including the kurtosis and NODDI model. Importantly, DIMOND reduces NODDI model fitting time from hours to minutes, or seconds by leveraging transfer learning. In summary, the self-supervised manner, high efficacy, and efficiency of DIMOND increase the practical feasibility and adoption of microstructure and connectivity mapping in clinical and neuroscientific applications.

Resting-state networks and their relationship with MoCA performance in PD patients

Although mild cognitive impairment is a common non-motor symptom experienced by individuals with Parkinson's Disease, the changes in intrinsic resting-state networks associated with its onset in Parkinson's remain underexamined. To address the issue, our study sought to examine resting-state network alterations and their association with total performance in the Montreal Cognitive Assessment and its cognitive domains in Parkinson's by means of functional magnetic resonance imaging of 29 Parkinson's patients with normal cognition, 25 Parkinson's patients with mild cognitive impairment, and 13 healthy controls. To contrast the Parkinson's groups with each other and the controls, the images were used to estimate the Z-score coefficient between the regions of interest from the default mode network, the salience network and the central executive network. Our first finding was that default mode and salience network connectivity decreased significantly in Parkinson's patients regardless of their cognitive status. Additionally, default mode network nodes had a negative and salience network nodes a positive correlation with the global assessment in Parkinson's with normal cognition; this inverse relationship of both networks to total score was not found in the group with cognitive impairment. Finally, a positive correlation was found between executive scores and anterior and posterior cortical network connectivity and, in the group with cognitive impairment, between language scores and salience network connectivity. Our results suggest that specific resting-state networks of Parkinson's patients with cognitive impairment differ from those of Parkinson's patients with normal cognition, supporting the evidence that cognitive impairment in Parkinson's Disease displays a differentiated neurodegenerative pattern.

The Association of Antidepressant Use and Impulse Control Disorder in Parkinson's Disease

OBJECTIVES

To examine whether initiation of an antidepressant is associated with the development of impulse control disorder (ICD) in patients with Parkinson's disease (PD).

DESIGN

We performed a retrospective analysis utilizing data from the Parkinson's Progression Markers Initiative (PPMI). Two-sample Mann-Whitney tests were used for comparison of continuous variables and Pearson χ2 tests were used for categorical variables. Kaplan-Meier survival analysis and cox proportional hazards regression analysis was used to assess the hazard of ICD with antidepressant exposure.

SETTING

The PPMI is a multicenter observational study of early PD with 52 sites throughout North America, Europe, and Africa.

PARTICIPANTS

Participants in the current study were those in the PPMI PD cohort with a primary diagnosis of idiopathic PD.

MEASUREMENTS

The presence of ICD was captured using the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease (QUIP). Antidepressant use was defined based on medication logs for each participant. Depressive symptoms were captured using the Geriatric Depression Scale (GDS).

RESULTS

A total of 1,045 individuals were included in the final analysis. There was a significant increase in the probability of ICD in those exposed to serotonergic antidepressants compared to those not exposed (Log-rank p <0.001). Serotonergic antidepressant use was associated with a hazard ratio for ICD of 1.4 (95% CI 1.0-1.8, z-value 2.1, p = 0.04) after adjusting for dopamine agonist use, depression, bupropion use, MAOI-B use, amantadine use, LEDD, disease duration, sex, and age.

CONCLUSIONS

Serotonergic antidepressant use appears to be temporally associated with ICD in patients with PD.

Longitudinal prevalence of neurogenic orthostatic hypotension in the idiopathic Parkinson Progression Marker Initiative (PPMI) cohort

BACKGROUND

Reported orthostatic hypotension (OH) prevalence in Parkinson's disease (PD) varies widely, with few studies evaluating specifically neurogenic-OH (nOH). The ratio of orthostatic heart rate (HR) to systolic blood pressure (SBP) change (Δ) is a valid screening method to stratify nOH/non-nOH but has had minimal epidemiologic application.

OBJECTIVE

To estimate the prevalence of nOH and non-nOH in the PPMI using the ΔHR/ΔSBP ratio and examine associations between nOH and various motor and non-motor measures.

METHODS

Longitudinal orthostatic vitals and motor and non-motor measures were extracted (baseline-month 48). Patients were consensus criteria classified as OH+/-, with ΔHR/ΔSBP sub-classification to nOH (ΔHR/ΔSBP < 0.5) or non-nOH (ratio ≥ 0.5). Prevalence was determined across visits. Independent linear mixed models tested associations between nOH/non-nOH and clinical variables.

RESULTS

Of N = 907 PD with baseline orthostatic vitals, 3.9 % and 1.8 % exhibited nOH and non-nOH, respectively. Prevalence of nOH/non-nOH increased yearly (P = 0.012, chi-square), though with modest magnitude (baseline: 5.6 % [95 % CI: 4.3-7.3 %]; month 48: 8.6 % [6.4-11.5 %]). nOH patients were older than PD with no OH and nOH was associated with greater impairment of motor and independent functioning than non-nOH/OH- groups. Cognitive function and typical OH symptoms were worse in PD + OH, generally.

CONCLUSIONS

nOH prevalence was greater than non-nOH in the PPMI early PD cohort, with modest prevalence increase over time. Our findings are consistent with prior studies of large cohorts that evaluated nOH, specifically. Those with early PD and nOH were likelier to be older and suffer from greater motor and functional impairment, but OH presence was generally associated with more cognitive impairment.

Motor Complications in Parkinson's Disease: Results from 3343 Patients Followed for up to 12 Years

BACKGROUND

Motor complications are well recognized in Parkinson's disease (PD), but their reported prevalence varies and functional impact has not been well studied.

OBJECTIVES

To quantify the presence, severity, impact and associated factors for motor complications in PD.

METHODS

Analysis of three large prospective cohort studies of recent-onset PD patients followed for up to 12 years. The MDS-UPDRS part 4 assessed motor complications and multivariable logistic regression tested for associations. Genetic risk score (GRS) for Parkinson's was calculated from 79 single nucleotide polymorphisms.

RESULTS

3343 cases were included (64.7% male). Off periods affected 35.0% (95% CI 33.0, 37.0) at 4-6 years and 59.0% (55.6, 62.3) at 8-10 years. Dyskinesia affected 18.5% (95% CI 16.9, 20.2) at 4-6 years and 42.1% (38.7, 45.5) at 8-10 years. Dystonia affected 13.4% (12.1, 14.9) at 4-6 years and 22.8% (20.1, 25.9) at 8-10 years. Off periods consistently caused greater functional impact than dyskinesia. Motor complications were more common among those with higher drug doses, younger age at diagnosis, female gender, and greater dopaminergic responsiveness (in challenge tests), with associations emerging 2-4 years post-diagnosis. Higher Parkinson's GRS was associated with early dyskinesia (0.026 ≤ P ≤ 0.050 from 2 to 6 years).

CONCLUSIONS

Off periods are more common and cause greater functional impairment than dyskinesia. We confirm previously reported associations between motor complications with several demographic and medication factors. Greater dopaminergic responsiveness and a higher genetic risk score are two novel and significant independent risk factors for the development of motor complications.

Assistive tools for classifying neurological disorders using fMRI and deep learning: A guide and example

BACKGROUND

Deep-learning (DL) methods are rapidly changing the way researchers classify neurological disorders. For example, combining functional magnetic resonance imaging (fMRI) and DL has helped researchers identify functional biomarkers of neurological disorders (e.g., brain activation and connectivity) and pilot innovative diagnostic models. However, the knowledge required to perform DL analyses is often domain-specific and is not widely taught in the brain sciences (e.g., psychology, neuroscience, and cognitive science). Conversely, neurological diagnoses and neuroimaging training (e.g., fMRI) are largely restricted to the brain and medical sciences. In turn, these disciplinary knowledge barriers and distinct specializations can act as hurdles that prevent the combination of fMRI and DL pipelines. The complexity of fMRI and DL methods also hinders their clinical adoption and generalization to real-world diagnoses. For example, most current models are not designed for clinical settings or use by nonspecialized populations such as students, clinicians, and healthcare workers. Accordingly, there is a growing area of assistive tools (e.g., software and programming packages) that aim to streamline and increase the accessibility of fMRI and DL pipelines for the diagnoses of neurological disorders.

OBJECTIVES AND METHODS

In this study, we present an introductory guide to some popular DL and fMRI assistive tools. We also create an example autism spectrum disorder (ASD) classification model using assistive tools (e.g., Optuna, GIFT, and the ABIDE preprocessed repository), fMRI, and a convolutional neural network.

RESULTS

In turn, we provide researchers with a guide to assistive tools and give an example of a streamlined fMRI and DL pipeline.

CONCLUSIONS

We are confident that this study can help more researchers enter the field and create accessible fMRI and deep-learning diagnostic models for neurological disorders.

Learning Phenotypic Associations for Parkinson's Disease with Longitudinal Clinical Records

Parkinson's disease (PD) is associated with multiple clinical motor and non-motor manifestations. Understanding of PD etiologies has been informed by a growing number of genetic mutations and various fluid-based and brain imaging biomarkers. However, the mechanisms underlying its varied phenotypic features remain elusive. The present work introduces a data-driven approach for generating phenotypic association graphs for PD cohorts. Data collected by the Parkinson's Progression Markers Initiative (PPMI), the Parkinson's Disease Biomarkers Program (PDBP), and the Fox Investigation for New Discovery of Biomarkers (BioFIND) were analyzed by this approach to identify heterogeneous and longitudinal phenotypic associations that may provide insight into the pathology of this complex disease. Findings based on the phenotypic association graphs could improve understanding of longitudinal PD pathologies and how these relate to patient symptomology.

Electrophysiological sweet spot mapping in deep brain stimulation for Parkinson's disease patients

BACKGROUND

Subthalamic deep brain stimulation (STN-DBS) is a well-established therapy to treat Parkinson's disease (PD). However, the STN-DBS sub-target remains debated. Recently, a white matter tract termed the hyperdirect pathway (HDP), directly connecting the motor cortex to STN, has gained interest as HDP stimulation is hypothesized to drive DBS therapeutic effects. Previously, we have investigated EEG-based evoked potentials (EPs) to better understand the neuroanatomical origins of the DBS clinical effect. We found a 3-ms peak (P3) relating to clinical benefit, and a 10-ms peak (P10) suggesting nigral side effects. Here, we aimed to investigate the neuroanatomical origins of DBS EPs using probabilistic mapping.

METHODS

EPs were recorded using EEG whilst low-frequency stimulation was delivered at all DBS-contacts individually. Next, EPs were mapped onto the patients' individual space and then transformed to MNI standard space. Using voxel-wise and fiber-wise probabilistic mapping, we determined hotspots/hottracts and coldspots/coldtracts for P3 and P10. Topography analysis was also performed to determine the spatial distribution of the DBS EPs.

RESULTS

In all 13 patients (18 hemispheres), voxel- and fiber-wise probabilistic mapping resulted in a P3-hotspot/hottract centered on the posterodorsomedial STN border indicative of HDP stimulation, while the P10-hotspot/hottract covered large parts of the substantia nigra.

CONCLUSION

This study investigated EP-based probabilistic mapping in PD patients during STN-DBS, revealing a P3-hotspot/hottract in line with HDP stimulation and P10-hotspot/hottract related to nigral stimulation. Results from this study provide key evidence for an electrophysiological measure of HDP and nigral stimulation.

Increased CSF DOPA Decarboxylase Correlates with Lower DaT-SPECT Binding: Analyses in Biopark and PPMI Cohorts

BACKGROUND

Recent studies identified increased cerebrospinal fluid (CSF) DOPA decarboxylase (DDC) as a promising biomarker for parkinsonian disorders, suggesting a compensation to dying dopaminergic neurons. A correlation with 123I-FP-CIT-SPECT (DaT-SPECT) imaging could shed light on this link.

OBJECTIVE

The objective is to assess the relationship between CSF DDC levels and DaT-SPECT binding values.

METHODS

A total of 51 and 72 Parkinson's disease (PD) subjects with available DaT-SPECT and CSF DDC levels were selected from the PPMI and Biopark cohorts, respectively. DDC levels were analyzed using proximity extension assay and correlated with DaT-SPECT striatal binding ratios (SBR). All analyses were corrected for age and sex.

RESULTS

CSF DDC levels in PD patients correlated negatively with DaT-SPECT SBR in both putamen and caudate nucleus. Additionally, SBR decreased with increased DDC levels over time in PD patients.

CONCLUSION

CSF DDC levels negatively correlate with DaT-SPECT SBR in levodopa-treated PD. © 2024 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Association of body mass index with rapid eye movement sleep behavior disorder in Parkinson's disease

Background

The association between body mass index (BMI) and rapid eye-movement (REM) sleep-related behavioral disorder (RBD) in Parkinson's disease (PD) remains unknown. Our study was to investigate the association of BMI with RBD in PD patients.

Methods

In this cross-sectional study, a total of 1,115 PD participants were enrolled from Parkinson's Progression Markers Initiative (PPMI) database. BMI was calculated as weight divided by height squared. RBD was defined as the RBD questionnaire (RBDSQ) score with the cutoff of 5 or more assessed. Univariable and multivariable logistic regression models were performed to examine the associations between BMI and the prevalence of RBD. Non-linear correlations were explored with use of restricted cubic spline (RCS) analysis. And the inflection point was determined by the two-line piecewise linear models.

Results

We identified 426 (38.2%) RBD. The proportion of underweight, normal, overweight and obese was 2.61, 36.59, 40.36, and 20.44%, respectively. In the multivariate logistic regression model with full adjustment for confounding variables, obese individuals had an odds ratio of 1.77 (95% confidence interval: 1.21 to 2.59) with RBD compared with those of normal weight. In the RCS models with three knots, BMI showed a non-linear association with RBD. The turning points of BMI estimated from piecewise linear models were of 28.16 kg/m2, 28.10 kg/m2, and 28.23 kg/m2 derived from univariable and multivariable adjusted logistic regression models. The effect modification by depression on the association between BMI and RBD in PD was also found in this study. Furthermore, the sensitivity analyses linked with cognition, education, and ethnic groups indicated the robustness of our results.

Conclusion

The current study found a significant dose-response association between BMI and RBD with a depression-based difference in the impact of BMI on RBD in PD patients.

Fine hippocampal morphology analysis with a multi-dataset cross-sectional study on 2911 subjects

CA1 subfield and subiculum of the hippocampus contain a series of dentate bulges, which are also called hippocampus dentation (HD). There have been several studies demonstrating an association between HD and brain disorders. Such as the number of hippocampal dentation correlates with temporal lobe epilepsy. And epileptic hippocampus have a lower number of dentation compared to contralateral hippocampus. However, most studies rely on subjective assessment by manual searching and counting in HD areas, which is time-consuming and labor-intensive to process large amounts of samples. And to date, only one objective method for quantifying HD has been proposed. Therefore, to fill this gap, we developed an automated and objective method to quantify HD and explore its relationship with neurodegenerative diseases. In this work, we performed a fine-scale morphological characterization of HD in 2911 subjects from four different cohorts of ADNI, PPMI, HCP, and IXI to quantify and explore differences between them in MR T1w images. The results showed that the degree of right hippocampal dentation are lower in patients with Alzheimer's disease than samples in mild cognitive impairment or cognitively normal, whereas this change is not significant in Parkinson's disease progression. The innovation of this paper that we propose a quantitative, robust, and fully automated method. These methodological innovation and corresponding results delineated above constitute the significance and novelty of our study. What's more, the proposed method breaks through the limitations of manual labeling and is the first to quantitatively measure and compare HD in four different brain populations including thousands of subjects. These findings revealed new morphological patterns in the hippocampal dentation, which can help with subsequent fine-scale hippocampal morphology research.

Olfactory Dysfunction and Long-Term Trajectories of Sleep Disorders among early Parkinson's Disease: Findings from a Longitudinal Cohort

BACKGROUND

Previous studies have suggested a connection between impaired olfactory function and an increased risk of rapid eye movement sleep behavior disorder (RBD) in individuals diagnosed with Parkinson's disease (PD). However, there is a gap in knowledge regarding the potential impact of olfactory dysfunction on the long-term patterns of sleep disorders among early PD patients.

METHODS

Data from the Parkinson's Progression Markers Initiative program included 589 participants with assessments of sleep disorders using the Epworth Sleepiness Scale (ESS) and RBD Screening Questionnaire (RBDSQ). Olfactory dysfunction at baseline was measured using the University of Pennsylvania Smell Identification Test. Trajectories of sleep disorders over a 5-year follow-up were identified using group-based trajectory modeling, and the relationship between olfactory dysfunction and sleep disorder trajectories was examined through binomial logistic regression.

RESULTS

Two distinct trajectories of sleep disorders over the 5-year follow-up period were identified, characterized by maintaining a low or high ESS score and a low or high RBDSQ score. An inversion association was observed between olfactory function measures and trajectories of excessive daytime sleepiness (odds ratio [OR] = 0.97, 95% confidence interval [CI] 0.95, 1.00, p = 0.038), after controlling for potential covariates. Similarly, olfactory function showed a significant association with lower trajectories of probable RBD (OR = 0.96, 95% CI 0.94, 0.98, p = 0.001) among early PD individuals. Consistent findings were replicated across alternative analytical models.

CONCLUSIONS

Our findings indicated that olfactory dysfunction was associated with unfavorable long-term trajectories of sleep disorders among early PD.

MRIO: the Magnetic Resonance Imaging Acquisition and Analysis Ontology

Magnetic resonance imaging of the brain is a useful tool in both the clinic and research settings, aiding in the diagnosis and treatments of neurological disease and expanding our knowledge of the brain. However, there are many challenges inherent in managing and analyzing MRI data, due in large part to the heterogeneity of data acquisition. To address this, we have developed MRIO, the Magnetic Resonance Imaging Acquisition and Analysis Ontology. MRIO provides well-reasoned classes and logical axioms for the acquisition of several MRI acquisition types and well-known, peer-reviewed analysis software, facilitating the use of MRI data. These classes provide a common language for the neuroimaging research process and help standardize the organization and analysis of MRI data for reproducible datasets. We also provide queries for automated assignment of analyses for given MRI types. MRIO aids researchers in managing neuroimaging studies by helping organize and annotate MRI data and integrating with existing standards such as Digital Imaging and Communications in Medicine and the Brain Imaging Data Structure, enhancing reproducibility and interoperability. MRIO was constructed according to Open Biomedical Ontologies Foundry principles and has contributed several classes to the Ontology for Biomedical Investigations to help bridge neuroimaging data to other domains. MRIO addresses the need for a "common language" for MRI that can help manage the neuroimaging research, by enabling researchers to identify appropriate analyses for sets of scans and facilitating data organization and reporting.

Parkinson's Disease Recognition Using Decorrelated Convolutional Neural Networks: Addressing Imbalance and Scanner Bias in rs-fMRI Data

Parkinson's disease (PD) is a neurodegenerative and progressive disease that impacts the nerve cells in the brain and varies from person to person. The exact cause of PD is still unknown, and the diagnosis of PD does not include a specific objective test with certainty. Although deep learning has made great progress in medical neuroimaging analysis, these methods are very susceptible to biases present in neuroimaging datasets. An innovative decorrelated deep learning technique is introduced to mitigate class bias and scanner bias while simultaneously focusing on finding distinguishing characteristics in resting-state functional MRI (rs-fMRI) data, which assists in recognizing PD with good accuracy. The decorrelation function reduces the nonlinear correlation between features and bias in order to learn bias-invariant features. The publicly available Parkinson's Progression Markers Initiative (PPMI) dataset, referred to as a single-scanner imbalanced dataset in this study, was used to validate our method. The imbalanced dataset problem affects the performance of the deep learning framework by overfitting to the majority class. To resolve this problem, we propose a new decorrelated convolutional neural network (DcCNN) framework by applying decorrelation-based optimization to convolutional neural networks (CNNs). An analysis of evaluation metrics comparisons shows that integrating the decorrelation function boosts the performance of PD recognition by removing class bias. Specifically, our DcCNN models perform significantly better than existing traditional approaches to tackle the imbalance problem. Finally, the same framework can be extended to create scanner-invariant features without significantly impacting the performance of a model. The obtained dataset is a multiscanner dataset, which leads to scanner bias due to the differences in acquisition protocols and scanners. The multiscanner dataset is a combination of two publicly available datasets, namely, PPMI and FTLDNI-the frontotemporal lobar degeneration neuroimaging initiative (NIFD) dataset. The results of t-distributed stochastic neighbor embedding (t-SNE) and scanner classification accuracy of our proposed feature extraction-DcCNN (FE-DcCNN) model validated the effective removal of scanner bias. Our method achieves an average accuracy of 77.80% on a multiscanner dataset for differentiating PD from a healthy control, which is superior to the DcCNN model trained on a single-scanner imbalanced dataset.

Machine learning for the detection and diagnosis of cognitive impairment in Parkinson's Disease: A systematic review

BACKGROUND

Parkinson's Disease is the second most common neurological disease in over 60s. Cognitive impairment is a major clinical symptom, with risk of severe dysfunction up to 20 years post-diagnosis. Processes for detection and diagnosis of cognitive impairments are not sufficient to predict decline at an early stage for significant impact. Ageing populations, neurologist shortages and subjective interpretations reduce the effectiveness of decisions and diagnoses. Researchers are now utilising machine learning for detection and diagnosis of cognitive impairment based on symptom presentation and clinical investigation. This work aims to provide an overview of published studies applying machine learning to detecting and diagnosing cognitive impairment, evaluate the feasibility of implemented methods, their impacts, and provide suitable recommendations for methods, modalities and outcomes.

METHODS

To provide an overview of the machine learning techniques, data sources and modalities used for detection and diagnosis of cognitive impairment in Parkinson's Disease, we conducted a review of studies published on the PubMed, IEEE Xplore, Scopus and ScienceDirect databases. 70 studies were included in this review, with the most relevant information extracted from each. From each study, strategy, modalities, sources, methods and outcomes were extracted.

RESULTS

Literatures demonstrate that machine learning techniques have potential to provide considerable insight into investigation of cognitive impairment in Parkinson's Disease. Our review demonstrates the versatility of machine learning in analysing a wide range of different modalities for the detection and diagnosis of cognitive impairment in Parkinson's Disease, including imaging, EEG, speech and more, yielding notable diagnostic accuracy.

CONCLUSIONS

Machine learning based interventions have the potential to glean meaningful insight from data, and may offer non-invasive means of enhancing cognitive impairment assessment, providing clear and formidable potential for implementation of machine learning into clinical practice.

Measuring Physical Functioning Using Wearable Sensors in Parkinson Disease and Chronic Obstructive Pulmonary Disease (the Accuracy of Digital Assessment of Performance Trial Study): Protocol for a Prospective Observational Study

BACKGROUND

Physical capacity and physical activity are important aspects of physical functioning and quality of life in people with a chronic disease such as Parkinson disease (PD) or chronic obstructive pulmonary disease (COPD). Both physical capacity and physical activity are currently measured in the clinic using standardized questionnaires and tests, such as the 6-minute walk test (6MWT) and the Timed Up and Go test (TUG). However, relying only on in-clinic tests is suboptimal since they offer limited information on how a person functions in daily life and how functioning fluctuates throughout the day. Wearable sensor technology may offer a solution that enables us to better understand true physical functioning in daily life.

OBJECTIVE

We aim to study whether device-assisted versions of 6MWT and TUG, such that the tests can be performed independently at home using a smartwatch, is a valid and reliable way to measure the performance compared to a supervised, in-clinic test.

METHODS

This is a decentralized, prospective, observational study including 100 people with PD and 100 with COPD. The inclusion criteria are broad: age ≥18 years, able to walk independently, and no co-occurrence of PD and COPD. Participants are followed for 15 weeks with 4 in-clinic visits, once every 5 weeks. Outcomes include several walking tests, cognitive tests, and disease-specific questionnaires accompanied by data collection using wearable devices (the Verily Study Watch and Modus StepWatch). Additionally, during the last 10 weeks of this study, participants will follow an aerobic exercise training program aiming to increase physical capacity, creating the opportunity to study the responsiveness of the remote 6MWT.

RESULTS

In total, 89 people with PD and 65 people with COPD were included in this study. Data analysis will start in April 2024.

CONCLUSIONS

The results of this study will provide information on the measurement properties of the device-assisted 6MWT and TUG in the clinic and at home. When reliable and valid, this can contribute to a better understanding of a person's physical capacity in real life, which makes it possible to personalize treatment options.

TRIAL REGISTRATION

ClinicalTrials.gov NCT05756075; https://clinicaltrials.gov/study/NCT05756075.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/55452.

Genetic and phenotypic characterization of Parkinson's disease at the clinic-wide level

Observational studies in Parkinson's disease (PD) deeply characterize relatively small numbers of participants. The Molecular Integration in Neurological Diagnosis Initiative seeks to characterize molecular and clinical features of every PD patient at the University of Pennsylvania (UPenn). The objectives of this study are to determine the feasibility of genetic characterization in PD and assess clinical features by sex and GBA1/LRRK2 status on a clinic-wide scale. All PD patients with clinical visits at the UPenn PD Center between 9/2018 and 12/2022 were eligible. Blood or saliva were collected, and a clinical questionnaire administered. Genotyping at 14 GBA1 and 8 LRRK2 variants was performed. PD symptoms were compared by sex and gene groups. 2063 patients were approached and 1,689 (82%) were enrolled, with 374 (18%) declining to participate. 608 (36%) females were enrolled, 159 (9%) carried a GBA1 variant, and 44 (3%) carried a LRRK2 variant. Compared with males, females across gene groups more frequently reported dystonia (53% vs 46%, p = 0.01) and anxiety (64% vs 55%, p < 0.01), but less frequently reported cognitive impairment (10% vs 49%, p < 0.01) and vivid dreaming (53% vs 60%, p = 0.01). GBA1 variant carriers more frequently reported anxiety (67% vs 57%, p = 0.04) and depression (62% vs 46%, p < 0.01) than non-carriers; LRRK2 variant carriers did not differ from non-carriers. We report feasibility for near-clinic-wide enrollment and characterization of individuals with PD during clinical visits at a high-volume academic center. Clinical symptoms differ by sex and GBA1, but not LRRK2, status.

The Relationship Between Autonomic Dysfunction and Mood Symptoms in De Novo Parkinson's Disease Patients Over Time

BACKGROUND

Autonomic dysfunction is prevalent in Parkinson's disease (PD) and can worsen quality of life. We examined: (a) whether specific autonomic symptoms were more strongly associated with anxiety or depression in PD and (b) whether overall autonomic dysfunction predicted mood trajectories over a 5-year period.

METHODS

Newly diagnosed individuals with PD (N = 414) from the Parkinson's Progression Markers Initiative completed self-report measures of depression, anxiety, and autonomic symptoms annually. Cross-sectional linear regressions examined relationships between specific autonomic subdomains (gastrointestinal, cardiovascular, thermoregulatory, etc.) and mood. Multilevel modeling examined longitudinal relationships with total autonomic load.

RESULTS

Gastrointestinal symptoms were associated with both higher anxiety (b = 1.04, 95% CI [.55, 1.53], P < .001) and depression (b = .24, 95% CI [.11, .37], P = .012), as were thermoregulatory symptoms (anxiety: b = 1.06, 95% CI [.46, 1.65], P = .004; depression: b = .25, 95% CI [.09, .42], P = .013), while cardiovascular (b = .36, 95% CI [.10, .62], P = .012) and urinary symptoms (b = .10, 95% CI [.01, .20], P = .037) were associated only with depression. Longitudinally, higher total autonomic load was associated with increases in both depression (b = .01, 95% CI [.00, .02], P = .015) and anxiety (b = .04, 95% CI [.01, .06], P < .001) over time, as well as occasion-to-occasion fluctuations (depression: b = .08, 95% CI [.05, .10], P < .001; anxiety: b = .24, 95% CI [.15, .32], P < .001).

CONCLUSION

Findings suggest autonomic dysfunction, particularly gastrointestinal and thermoregulatory symptoms, may be an indicator for elevated anxiety/depression and a potential treatment target early on in PD.

Associations of sleep disorders with serum neurofilament light chain levels in Parkinson's disease

BACKGROUND

Sleep disorders are a prevalent non-motor symptom of Parkinson's disease (PD), although reliable biological markers are presently lacking.

OBJECTIVES

To explore the associations between sleep disorders and serum neurofilament light chain (NfL) levels in individuals with prodromal and early PD.

METHODS

The study contained 1113 participants, including 585 early PD individuals, 353 prodromal PD individuals, and 175 healthy controls (HCs). The correlations between sleep disorders (including rapid eye movement sleep behavior disorder (RBD) and excessive daytime sleepiness (EDS)) and serum NfL levels were researched using multiple linear regression models and linear mixed-effects models. We further investigated the correlations between the rates of changes in daytime sleepiness and serum NfL levels using multiple linear regression models.

RESULTS

In baseline analysis, early and prodromal PD individuals who manifested specific behaviors of RBD showed significantly higher levels of serum NfL. Specifically, early PD individuals who experienced nocturnal dream behaviors (β = 0.033; P = 0.042) and movements of arms or legs during sleep (β = 0.027; P = 0.049) showed significantly higher serum NfL levels. For prodromal PD individuals, serum NfL levels were significantly higher in individuals suffering from disturbed sleep (β = 0.038; P = 0.026). Our longitudinal findings support these baseline associations. Serum NfL levels showed an upward trend in early PD individuals who had a higher total RBDSQ score (β = 0.002; P = 0.011) or who were considered as probable RBD (β = 0.012; P = 0.009) or who exhibited behaviors on several sub-items of the RBDSQ. In addition, early PD individuals who had a high total ESS score (β = 0.001; P = 0.012) or who were regarded to have EDS (β = 0.013; P = 0.007) or who exhibited daytime sleepiness in several conditions had a trend toward higher serum NfL levels.

CONCLUSION

Sleep disorders correlate with higher serum NfL, suggesting a link to PD neuronal damage. Early identification of sleep disorders and NfL monitoring are pivotal in detecting at-risk PD patients promptly, allowing for timely intervention. Regular monitoring of NfL levels holds promise for tracking both sleep disorders and disease progression, potentially emerging as a biomarker for evaluating treatment outcomes.

Longitudinal Mediating Effect of Depression on the Relationship between Excessive Daytime Sleepiness and Activities of Daily Living in Parkinson's Disease

OBJECTIVES

Whether depression affects activities of daily living (ADLs) in patients with Parkinson's disease (PD) via excessive daytime sleepiness (EDS) remains unclear; moreover, few longitudinal studies have been conducted.

METHODS

We recruited 421 patients from the Parkinson's Progression Markers Initiative. We constructed a latent growth mediation model to explore the longitudinal mediating effect of depression on the relationship between EDS and ADLs.

RESULTS

EDS (p < .001) and depression scores (p < .001) both increased, and ADL scores (p < .001) decreased. Moreover, EDS was positively correlated with depression, whereas an increase in EDS significantly reduced ADLs. The initial value (95% confidence interval [CI]: 0.026, 0.154) and the rate of change (95% CI: 0.138, 0.514) of self-reported depression measured using the Geriatric Depression Scale(GDS) partially mediated the association between EDS and ADL score.

CONCLUSIONS

The indirect effect of the longitudinal changes of depression on the relationship between EDS and ADLs highlights the importance of depression changes in PD patients with EDS.

CLINICAL IMPLICATIONS

Depression should be considered a mediator by clinicians; preventing the worsening of depression is essential for improving ADLs in patients with PD, especially those with EDS.

Longitudinal study investigating the influence of COMT gene polymorphism on cortical thickness changes in Parkinson's disease over four years

Parkinson's disease (PD) is a progressive neurodegenerative disorder affecting over 3% of those over 65. It's caused by reduced dopaminergic neurons and Lewy bodies, leading to motor and non-motor symptoms. The relationship between COMT gene polymorphisms and PD is complex and not fully elucidated. Some studies have reported associations between certain COMT gene variants and PD risk, while others have not found significant associations. This study investigates how COMT gene variations impact cortical thickness changes in PD patients over time, aiming to link genetic factors, especially COMT gene variations, with PD progression. This study analyzed data from 44 PD patients with complete 4-year imaging follow-up from the Parkinson Progression Marker Initiative (PPMI) database. Magnetic resonance imaging (MRI) scans were acquired using consistent methods across 9 different MRI scanners. COMT single-nucleotide polymorphisms (SNPs) were assessed based on whole genome sequencing data. Longitudinal image analysis was conducted using FreeSurfer's processing pipeline. Linear mixed-effect models were employed to examine the interaction effect of genetic variations and time on cortical thickness, while controlling for covariates and subject-specific variations. The rs165599 SNP stands out as a potential contributor to alterations in cortical thickness, showing a significant reduction in overall mean cortical thickness in both hemispheres in homozygotes (Left: P = 0.023, Right: P = 0.028). The supramarginal, precentral, and superior frontal regions demonstrated significant bilateral alterations linked to rs165599. Our findings suggest that the rs165599 variant leads to earlier manifestation of cortical thinning during the course of the disease. However, it does not result in more severe cortical thinning outcomes over time. There is a need for larger cohorts and control groups to validate these findings and consider genetic variant interactions and clinical features to elucidate the specific mechanisms underlying COMT-related neurodegenerative processes in PD.

Novel Variants Linked to the Prodromal Stage of Parkinson's Disease (PD) Patients

BACKGROUND AND OBJECTIVE

The symptoms of most neurodegenerative diseases, including Parkinson's disease (PD), usually do not occur until substantial neuronal loss occurs. This makes the process of early diagnosis very challenging. Hence, this research used variant call format (VCF) analysis to detect variants and novel genes that could be used as prognostic indicators in the early diagnosis of prodromal PD.

MATERIALS AND METHODS

Data were obtained from the Parkinson's Progression Markers Initiative (PPMI), and we analyzed prodromal patients with gVCF data collected in the 2021 cohort. A total of 304 participants were included, including 100 healthy controls, 146 prodromal genetic individuals, 21 prodromal hyposmia individuals, and 37 prodromal individuals with RBD. A pipeline was developed to process the samples from gVCF to reach variant annotation and pathway and disease association analysis.

RESULTS

Novel variant percentages were detected in the analyzed prodromal subgroups. The prodromal subgroup analysis revealed novel variations of 1.0%, 1.2%, 0.6%, 0.3%, 0.5%, and 0.4% for the genetic male, genetic female, hyposmia male, hyposmia female, RBD male, and RBD female groups, respectively. Interestingly, 12 potentially novel loci (MTF2, PIK3CA, ADD1, SYBU, IRS2, USP8, PIGL, FASN, MYLK2, USP25, EP300, and PPP6R2) that were recently detected in PD patients were detected in the prodromal stage of PD.

CONCLUSIONS

Genetic biomarkers are crucial for the early detection of Parkinson's disease and its prodromal stage. The novel PD genes detected in prodromal patients could aid in the use of gene biomarkers for early diagnosis of the prodromal stage without relying only on phenotypic traits.

Short fiber bundle filtering and test-retest reproducibility of the Superficial White Matter

In recent years, there has been a growing interest in studying the Superficial White Matter (SWM). The SWM consists of short association fibers connecting near giry of the cortex, with a complex organization due to their close relationship with the cortical folding patterns. Therefore, their segmentation from dMRI tractography datasets requires dedicated methodologies to identify the main fiber bundle shape and deal with spurious fibers. This paper presents an enhanced short fiber bundle segmentation based on a SWM bundle atlas and the filtering of noisy fibers. The method was tuned and evaluated over HCP test-retest probabilistic tractography datasets (44 subjects). We propose four fiber bundle filters to remove spurious fibers. Furthermore, we include the identification of the main fiber fascicle to obtain well-defined fiber bundles. First, we identified four main bundle shapes in the SWM atlas, and performed a filter tuning in a subset of 28 subjects. The filter based on the Convex Hull provided the highest similarity between corresponding test-retest fiber bundles. Subsequently, we applied the best filter in the 16 remaining subjects for all atlas bundles, showing that filtered fiber bundles significantly improve test-retest reproducibility indices when removing between ten and twenty percent of the fibers. Additionally, we applied the bundle segmentation with and without filtering to the ABIDE-II database. The fiber bundle filtering allowed us to obtain a higher number of bundles with significant differences in fractional anisotropy, mean diffusivity, and radial diffusivity of Autism Spectrum Disorder patients relative to controls.

Lysosomal genes contribute to Parkinson's disease near agriculture with high intensity pesticide use

Parkinson's disease (PD), the second most common neurodegenerative disorder, develops sporadically, likely through a combination of polygenic and environmental factors. Previous studies associate pesticide exposure and genes involved in lysosomal function with PD risk. We evaluated the frequency of variants in lysosomal function genes among patients from the Parkinson's, Environment, and Genes (PEG) study with ambient pesticide exposure from agricultural sources. 757 PD patients, primarily of White European/non-Hispanic ancestry (75%), were screened for variants in 85 genes using a custom amplicon panel. Variant enrichment was calculated against the Genome Aggregation Database (gnomAD). Enriched exonic variants were prioritized by exposure to a cluster of pesticides used on cotton and severity of disease progression in a subset of 386 patients subdivided by race/ethnicity. Gene enrichment analysis identified 36 variants in 26 genes in PEG PD patients. Twelve of the identified genes (12/26, 46%) had multiple enriched variants and/or a single enriched variant present in multiple individuals, representing 61% (22/36) of the observed variation in the cohort. The majority of enriched variants (26/36, 72%) were found in genes contributing to lysosomal function, particularly autophagy, and were bioinformatically deemed functionally deleterious (31/36, 86%). We conclude that, in this study, variants in genes associated with lysosomal function, notably autophagy, were enriched in PD patients exposed to agricultural pesticides suggesting that altered lysosomal function may generate an underlying susceptibility for developing PD with pesticide exposure. Further study of gene-environment interactions targeting lysosomal function may improve understanding of PD risk in individuals exposed to pesticides.

Cortical and subcortical functional connectivity and cognitive impairment in Parkinson's disease

Parkinson's disease (PD) is a neurodegenerative disease with cognitive as well as motor impairments. While much is known about the brain networks leading to motor impairments in PD, less is known about the brain networks contributing to cognitive impairments. Here, we leveraged resting-state functional magnetic resonance imaging (rs-fMRI) data from the Parkinson's Progression Marker Initiative (PPMI) to examine network dysfunction in PD patients with cognitive impairment. We focus on canonical cortical networks linked to cognition, including the salience network (SAL), frontoparietal network (FPN), and default mode network (DMN), as well as a subcortical basal ganglia network (BGN). We used the Montreal Cognitive Assessment (MoCA) as a continuous index of coarse cognitive function in PD. In 82 PD patients, we found that lower MoCA scores were linked with lower intra-network connectivity of the FPN. We also found that lower MoCA scores were linked with lower inter-network connectivity between the SAL and the BGN, the SAL and the DMN, as well as the FPN and the DMN. These data elucidate the relationship of cortical and subcortical functional connectivity with cognitive impairments in PD.

Impulse control behaviors and apathy commonly co-occur in de novo Parkinson's disease and predict the incidence of levodopa-induced dyskinesia

OBJECTIVE

Impulse control behaviors (ICBs) and apathy are believed to represent opposite motivational expressions of the same behavioral spectrum involving hypo- and hyperdopaminergic status, but this has been recently debated. Our study aims to estimate the co-occurrence of ICBs and apathy in early Parkinson's disease (PD) and to determine whether this complex neuropsychiatric condition is an important marker of PD prognoses.

METHODS

Neuropsychiatric symptoms, clinical data, neuroimaging results, and demographic data from de novo PD patients were obtained from the Parkinson's Progression Markers Initiative, a prospective, multicenter, observational cohort. The clinical characteristics of ICBs co-occurring with apathy and their prevalence were analyzed. We compared the prognoses of the different groups during the 8-year follow-up. Multivariate Cox regression analysis was conducted to predict the development of levodopa-induced dyskinesia (LID) using baseline neuropsychiatric symptoms.

RESULTS

A total of 422 PD patients and 195 healthy controls (HCs) were included. In brief, 87 (20.6 %) de novo PD patients and 37 (19.0 %) HCs had ICBs at baseline. Among them, 23 (26.4 %) de novo PD patients and 3 (8.1 %) HCs had clinical symptoms of both ICBs and apathy. The ICBs and apathy group had more severe non-motor symptoms than the isolated ICBs group. Cox regression analysis demonstrated that the co-occurrence of ICBs and apathy was a risk factor for LID development (HR 2.229, 95 % CI 1.209 to 4.110, p = 0.010).

CONCLUSIONS

Co-occurrence of ICBs and apathy is common in patients with early PD and may help to identify the risk of LID development.

Using a deep generation network reveals neuroanatomical specificity in hemispheres

Asymmetry is an important property of brain organization, but its nature is still poorly understood. Capturing the neuroanatomical components specific to each hemisphere facilitates the understanding of the establishment of brain asymmetry. Since deep generative networks (DGNs) have powerful inference and recovery capabilities, we use one hemisphere to predict the opposite hemisphere by training the DGNs, which automatically fit the built-in dependencies between the left and right hemispheres. After training, the reconstructed images approximate the homologous components in the hemisphere. We use the difference between the actual and reconstructed hemispheres to measure hemisphere-specific components due to asymmetric expression of environmental and genetic factors. The results show that our model is biologically plausible and that our proposed metric of hemispheric specialization is reliable, representing a wide range of individual variation. Together, this work provides promising tools for exploring brain asymmetry and new insights into self-supervised DGNs for representing the brain.

A decision support system based on recurrent neural networks to predict medication dosage for patients with Parkinson's disease

Using deep learning has demonstrated significant potential in making informed decisions based on clinical evidence. In this study, we deal with optimizing medication and quantitatively present the role of deep learning in predicting the medication dosage for patients with Parkinson's disease (PD). The proposed method is based on recurrent neural networks (RNNs) and tries to predict the dosage of five critical medication types for PD, including levodopa, dopamine agonists, monoamine oxidase-B inhibitors, catechol-O-methyltransferase inhibitors, and amantadine. Recurrent neural networks have memory blocks that retain crucial information from previous patient visits. This feature is helpful for patients with PD, as the neurologist can refer to the patient's previous state and the prescribed medication to make informed decisions. We employed data from the Parkinson's Progression Markers Initiative. The dataset included information on the Unified Parkinson's Disease Rating Scale, Activities of Daily Living, Hoehn and Yahr scale, demographic details, and medication use logs for each patient. We evaluated several models, such as multi-layer perceptron (MLP), Simple-RNN, long short-term memory (LSTM), and gated recurrent units (GRU). Our analysis found that recurrent neural networks (LSTM and GRU) performed the best. More specifically, when using LSTM, we were able to predict levodopa and dopamine agonist dosage with a mean squared error of 0.009 and 0.003, mean absolute error of 0.062 and 0.030, root mean square error of 0.099 and 0.053, and R-squared of 0.514 and 0.711, respectively.

Corpus callosum and cerebellum participate in semantic dysfunction of Parkinson's disease: a diffusion tensor imaging-based cross-sectional study

Language dysfunction is common in Parkinson's disease (PD) patients, among which, the decline of semantic fluency is usually observed. This study aims to explore the relationship between white matter (WM) alterations and semantic fluency changes in PD patients. 127 PD patients from the Parkinson's Progression Markers Initiative cohort who received diffusion tensor imaging scanning, clinical assessment and semantic fluency test (SFT) were included. Tract-based special statistics, automated fiber quantification, graph-theoretical and network-based analyses were performed to analyze the correlation between WM structural changes, brain network features and semantic fluency in PD patients. Fractional anisotropy of corpus callosum, anterior thalamic radiation, inferior front-occipital fasciculus, and uncinate fasciculus, were positively correlated with SFT scores, while a negative correlation was identified between radial diffusion of the corpus callosum, inferior longitudinal fasciculus, and SFT scores. Automatic fiber quantification identified similar alterations with more details in these WM tracts. Brain network analysis positively correlated SFT scores with nodal efficiency of cerebellar lobule VIII, and nodal local efficiency of cerebellar lobule X. WM integrity and myelin integrity in the corpus callosum and several other language-related WM tracts may influence the semantic function in PD patients. Damage to the cerebellum lobule VIII and lobule X may also be involved in semantic dysfunction in PD patients.

Predictors of the Rapid Progression in Prodromal Parkinson's Disease: A Longitudinal Follow-Up Study

INTRODUCTION

Parkinson's disease (PD) is characterized by a prodromal phase preceding the onset of classic motor symptoms. The duration and clinical manifestations of prodromal PD vary widely, indicating underlying heterogeneity within this stage. This discrepancy prompts the question of whether specific factors contribute to the divergent rates of progression in prodromal PD.

METHODS

This study included prodromal PD patients from the Parkinson's progression marker initiative. They were followed up to assess the disease progression. The data collected during the follow-up period were analyzed to identify potential predictors of rapid disease progression in prodromal PD.

RESULTS

In this study, 61 individuals with prodromal PD were enrolled. Among them, 43 patients presented with both RBD and hyposmia, 17 had hyposmia alone, and 1 had RBD alone at baseline. 13 (21.3%) prodromal PD participants exhibited rapid disease progression, with two of these cases advancing to non-neurological diseases. Significant differences were observed between the rapid progression group and no rapid progression group in terms of MDS-UPDRS II score and UPSIT score. Longitudinal analysis showed a significant increase in the MDS-UPDRS III score and MDS-UPDRS total score in the rapid progression group. Regression analyses identified the MDS-UPDRS II score and UPSIT score as predictors of rapid disease progression in prodromal PD.

CONCLUSION

Our study findings suggest that the MDS-UPDRS II score and UPSIT score may serve as clinical markers associated with rapid disease progression. Further research and development of precise biomarkers and advanced assessment methods are needed to enhance our understanding of prodromal PD and its progression patterns.

Automated identification of uncertain cases in deep learning-based classification of dopamine transporter SPECT to improve clinical utility and acceptance

PURPOSE

Deep convolutional neural networks (CNN) are promising for automatic classification of dopamine transporter (DAT)-SPECT images. Reporting the certainty of CNN-based decisions is highly desired to flag cases that might be misclassified and, therefore, require particularly careful inspection by the user. The aim of the current study was to design and validate a CNN-based system for the identification of uncertain cases.

METHODS

A network ensemble (NE) combining five CNNs was trained for binary classification of [123I]FP-CIT DAT-SPECT images as "normal" or "neurodegeneration-typical reduction" with high accuracy (NE for classification, NEfC). An uncertainty detection module (UDM) was obtained by combining two additional NE, one trained for detection of "reduced" DAT-SPECT with high sensitivity, the other with high specificity. A case was considered "uncertain" if the "high sensitivity" NE and the "high specificity" NE disagreed. An internal "development" dataset of 1740 clinical DAT-SPECT images was used for training (n = 1250) and testing (n = 490). Two independent datasets with different image characteristics were used for testing only (n = 640, 645). Three established approaches for uncertainty detection were used for comparison (sigmoid, dropout, model averaging).

RESULTS

In the test data from the development dataset, the NEfC achieved 98.0% accuracy. 4.3% of all test cases were flagged as "uncertain" by the UDM: 2.5% of the correctly classified cases and 90% of the misclassified cases. NEfC accuracy among "certain" cases was 99.8%. The three comparison methods were less effective in labelling misclassified cases as "uncertain" (40-80%). These findings were confirmed in both additional test datasets.

CONCLUSION

The UDM allows reliable identification of uncertain [123I]FP-CIT SPECT with high risk of misclassification. We recommend that automatic classification of [123I]FP-CIT SPECT images is combined with an UDM to improve clinical utility and acceptance. The proposed UDM method ("high sensitivity versus high specificity") might be useful also for DAT imaging with other ligands and for other binary classification tasks.