Development of a Tremor Detection Algorithm for Use in an Academic Movement Disorders Center.

Tremor, defined as an "involuntary, rhythmic, oscillatory movement of a body part", is a key feature of many neurological conditions including Parkinson's disease and essential tremor. Clinical assessment continues to be performed by visual observation with quantification on clinical scales. Methodologies for objectively quantifying tremor are promising but remain non-standardized across centers. Our center performs full-body behavioral testing with 3D motion capture for clinical and research purposes in patients with Parkinson's disease, essential tremor, and other conditions. The objective of this study was to assess the ability of several candidate processing pipelines to identify the presence or absence of tremor in kinematic data from patients with confirmed movement disorders and compare them to expert ratings from movement disorders specialists. We curated a database of 2272 separate kinematic data recordings from our center, each of which was contemporaneously annotated as tremor present or absent by a movement physician. We compared the ability of six separate processing pipelines to recreate clinician ratings based on F1 score, in addition to accuracy, precision, and recall. The performance across algorithms was generally comparable. The average F1 score was 0.84±0.02 (mean ± SD; range 0.81-0.87). The second highest performing algorithm (cross-validated F1=0.87) was a hybrid that used engineered features adapted from an algorithm in longstanding clinical use with a modern Support Vector Machine classifier. Taken together, our results suggest the potential to update legacy clinical decision support systems to incorporate modern machine learning classifiers to create better-performing tools.

An Explainable Spatial-Temporal Graphical Convolutional Network to Score Freezing of Gait in Parkinsonian Patients.

Freezing of gait (FOG) is a poorly understood heterogeneous gait disorder seen in patients with parkinsonism which contributes to significant morbidity and social isolation. FOG is currently measured with scales that are typically performed by movement disorders specialists (i.e., MDS-UPDRS), or through patient completed questionnaires (N-FOG-Q) both of which are inadequate in addressing the heterogeneous nature of the disorder and are unsuitable for use in clinical trials The purpose of this study was to devise a method to measure FOG objectively, hence improving our ability to identify it and accurately evaluate new therapies. A major innovation of our study is that it is the first study of its kind that uses the largest sample size (>30 h, N = 57) in order to apply explainable, multi-task deep learning models for quantifying FOG over the course of the medication cycle and at varying levels of parkinsonism severity. We trained interpretable deep learning models with multi-task learning to simultaneously score FOG (cross-validated F1 score 97.6%), identify medication state (OFF vs. ON levodopa; cross-validated F1 score 96.8%), and measure total PD severity (MDS-UPDRS-III score prediction error ≤ 2.7 points) using kinematic data of a well-characterized sample of N = 57 patients during levodopa challenge tests. The proposed model was able to explain how kinematic movements are associated with each FOG severity level that were highly consistent with the features, in which movement disorders specialists are trained to identify as characteristics of freezing. Overall, we demonstrate that deep learning models' capability to capture complex movement patterns in kinematic data can automatically and objectively score FOG with high accuracy. These models have the potential to discover novel kinematic biomarkers for FOG that can be used for hypothesis generation and potentially as clinical trial outcome measures.

An explainable spatial-temporal graphical convolutional network to score freezing of gait in parkinsonian patients.

Freezing of gait (FOG) is a poorly understood heterogeneous gait disorder seen in patients with parkinsonism which contributes to significant morbidity and social isolation. FOG is currently measured with scales that are typically performed by movement disorders specialists (ie. MDS-UPDRS), or through patient completed questionnaires (N-FOG-Q) both of which are inadequate in addressing the heterogeneous nature of the disorder and are unsuitable for use in clinical trials The purpose of this study was to devise a method to measure FOG objectively, hence improving our ability to identify it and accurately evaluate new therapies. We trained interpretable deep learning models with multi-task learning to simultaneously score FOG (cross-validated F1 score 97.6%), identify medication state (OFF vs. ON levodopa; cross-validated F1 score 96.8%), and measure total PD severity (MDS-UPDRS-III score prediction error ≤ 2.7 points) using kinematic data of a well-characterized sample of N=57 patients during levodopa challenge tests. The proposed model was able to identify kinematic features associated with each FOG severity level that were highly consistent with the features that movement disorders specialists are trained to identify as characteristic of freezing. In this work, we demonstrate that deep learning models' capability to capture complex movement patterns in kinematic data can automatically and objectively score FOG with high accuracy. These models have the potential to discover novel kinematic biomarkers for FOG that can be used for hypothesis generation and potentially as clinical trial outcome measures.

Subtle microstructural changes of the cerebellum in a knock-in mouse model of DYT1 dystonia.

The dystonias are a group of disorders characterized by involuntary twisting and repetitive movements. DYT1 dystonia is an inherited form of dystonia caused by a mutation in the TOR1A gene, which encodes torsinA. TorsinA is expressed in many regions of the nervous system, and the regions responsible for causing dystonic movements remain uncertain. Most prior studies have focused on the basal ganglia, although there is emerging evidence for abnormalities in the cerebellum too. In the current studies, we examined the cerebellum for structural abnormalities in a knock-in mouse model of DYT1 dystonia. The gross appearance of the cerebellum appeared normal in the mutant mice, but stereological measures revealed the cerebellum to be 5% larger in mutant compared to control mice. There were no changes in the numbers of Purkinje cells, granule cells, or neurons of the deep cerebellar nuclei. However, Golgi histochemical studies revealed Purkinje cells to have thinner dendrites, and fewer and less complex dendritic spines. There also was a higher frequency of heterotopic Purkinje cells displaced into the molecular layer. These results reveal subtle structural changes of the cerebellum that are similar to those reported for the basal ganglia in the DYT1 knock-in mouse model.

Sleep fragmentation and motor restlessness in a Drosophila model of Restless Legs Syndrome.

Restless Legs Syndrome (RLS), first chronicled by Willis in 1672 and described in more detail by Ekbom in 1945, is a prevalent sensorimotor neurological disorder (5%-10% in the population) with a circadian predilection for the evening and night. Characteristic clinical features also include a compelling urge to move during periods of rest, relief with movement, involuntary movements in sleep (viz., periodic leg movements of sleep), and fragmented sleep. Although the pathophysiology of RLS is unknown, dopaminergic neurotransmission and deficits in iron availability modulate expressivity. Genome-wide association studies have identified a polymorphism in an intronic region of the BTBD9 gene on chromosome 6 that confers substantial risk for RLS. Here, we report that loss of the Drosophila homolog CG1826 (dBTBD9) appreciably disrupts sleep with concomitant increases in waking and motor activity. We further show that BTBD9 regulates brain dopamine levels in flies and controls iron homeostasis through the iron regulatory protein-2 in human cell lines. To our knowledge, this represents the first reverse genetic analysis of a "novel" or heretofore poorly understood gene implicated in an exceedingly common and complex sleep disorder and the development of an RLS animal model that closely recapitulates all disease phenotypes.

Loss of enteric dopaminergic neurons and associated changes in colon motility in an MPTP mouse model of Parkinson's disease.

Gastrointestinal (GI) dysfunction is the most common non-motor symptom of Parkinson's disease (PD). Symptoms of GI dysmotility include early satiety and nausea from delayed gastric emptying, bloating from poor small bowel coordination, and constipation and defecatory dysfunction from impaired colonic transit. Understanding the pathophysiology and treatment of these symptoms in PD patients has been hampered by the lack of investigation into GI symptoms and pathology in PD animal models. We report that the prototypical parkinsonian neurotoxin, MPTP (1-methyl 4-phenyl 1,2,3,6-tetrahydropyridine), is a selective dopamine neuron toxin in the enteric nervous system (ENS). When examined 10 days after treatment, there was a 40% reduction of dopamine neurons in the ENS of C57Bl/6 mice administered MPTP (60 mg/kg). There were no differences in the density of cholinergic or nitric oxide neurons. Electrophysiological recording of neural-mediated muscle contraction in isolated colon from MPTP-treated animals confirmed a relaxation defect associated with dopaminergic degeneration. Behaviorally, MPTP induced a transient increase in colon motility, but no changes in gastric emptying or small intestine transit. These results provide the first comprehensive assessment of gastrointestinal pathophysiology in an animal model of PD. They provide insight into the impact of dopaminergic dysfunction on gastrointestinal motility and a benchmark for assessment of other PD model systems.