Machine-learning assisted swallowing assessment: a deep learning-based quality improvement tool to screen for post-stroke dysphagia.

Introduction: Post-stroke dysphagia is common and associated with significant morbidity and mortality, rendering bedside screening of significant clinical importance. Using voice as a biomarker coupled with deep learning has the potential to improve patient access to screening and mitigate the subjectivity associated with detecting voice change, a component of several validated screening protocols. Methods: In this single-center study, we developed a proof-of-concept model for automated dysphagia screening and evaluated the performance of this model on training and testing cohorts. Patients were admitted to a comprehensive stroke center, where primary English speakers could follow commands without significant aphasia and participated on a rolling basis. The primary outcome was classification either as a pass or fail equivalent using a dysphagia screening test as a label. Voice data was recorded from patients who spoke a standardized set of vowels, words, and sentences from the National Institute of Health Stroke Scale. Seventy patients were recruited and 68 were included in the analysis, with 40 in training and 28 in testing cohorts, respectively. Speech from patients was segmented into 1,579 audio clips, from which 6,655 Mel-spectrogram images were computed and used as inputs for deep-learning models (DenseNet and ConvNext, separately and together). Clip-level and participant-level swallowing status predictions were obtained through a voting method. Results: The models demonstrated clip-level dysphagia screening sensitivity of 71% and specificity of 77% (F1 = 0.73, AUC = 0.80 [95% CI: 0.78-0.82]). At the participant level, the sensitivity and specificity were 89 and 79%, respectively (F1 = 0.81, AUC = 0.91 [95% CI: 0.77-1.05]). Discussion: This study is the first to demonstrate the feasibility of applying deep learning to classify vocalizations to detect post-stroke dysphagia. Our findings suggest potential for enhancing dysphagia screening in clinical settings. https://github.com/UofTNeurology/masa-open-source.

Calcium pyrophosphate dihydrate crystal deposition disease and retro-odontoid pseudotumor rupture managed via posterior occipital cervical instrumented fusion: illustrative case.

BACKGROUND: Craniocervical junction and subaxial cervical spinal manifestations of calcium pyrophosphate deposition disease are rarely encountered. The authors presented a severe case of retro-odontoid pseudotumor rupture causing rapid quadriparesis and an acute comatose state with subsequent radiographic and clinical improvement after posterior occipital cervical fusion. OBSERVATIONS: The authors surveyed the literature and outlined multiple described operative management strategies for compressive cervical and craniocervical junction calcium pyrophosphate deposition disease manifestations ranging from neck pain to paresthesia, weakness, myelopathy, quadriparesis, and cranial neuropathies. In this report, radiographic features of cervical and craniocervical junction calcium pyrophosphate deposition disease were explored. Several previously described surgical strategies were compiled, including patient characteristics and outcomes. LESSONS: With this case report, the authors presented for the first time an isolated posterior occipital cervical fusion for treatment of a compressive retro-odontoid pseudotumor with rupture into the brainstem. They demonstrated rapid clinical and radiographic resolution after stabilization of cranial cervical junction only 12 weeks postsurgery.

Are smartphones and machine learning enough to diagnose tremor?

BACKGROUND: Patients with essential tremor (ET), Parkinson's disease (PD) and dystonic tremor (DT) can be difficult to classify and often share similar characteristics. OBJECTIVES: To use ubiquitous smartphone accelerometers with and without clinical features to automate tremor classification using supervised machine learning, and to use unsupervised learning to evaluate if natural clusterings of patients correspond to assigned clinical diagnoses. METHODS: A supervised machine learning classifier was trained to classify 78 tremor patients using leave-one-out cross-validation to estimate performance on unseen accelerometer data. An independent cohort of 27 patients were also studied. Next, we focused on a subset of 48 patients with both smartphone-based tremor measurements and detailed clinical assessment metrics and compared two separate machine learning classifiers trained on these data. RESULTS: The classifier yielded a total accuracy of 74.4% and F1-score of 0.74 for a trinary classification with an area under the curve of 0.904, average F1-score of 0.94, specificity of 97% and sensitivity of 84% in classifying PD from ET or DT. The algorithm classified ET from non-ET with 88% accuracy, but only classified DT from non-DT with 29% accuracy. A poorer performance was found in the independent cohort. Classifiers trained on accelerometer and clinical data respectively obtained similar results. CONCLUSIONS: Machine learning classifiers achieved a high accuracy of PD, however moderate accuracy of ET, and poor accuracy of DT classification. This underscores the difficulty of using AI to classify some tremors due to lack of specificity in clinical and neuropathological features, reinforcing that they may represent overlapping syndromes.

Origin of heterogeneous spiking patterns from continuously distributed ion channel densities: a computational study in spinal dorsal horn neurons.

KEY POINTS: Distinct spiking patterns may arise from qualitative differences in ion channel expression (i.e. when different neurons express distinct ion channels) and/or when quantitative differences in expression levels qualitatively alter the spike generation process. We hypothesized that spiking patterns in neurons of the superficial dorsal horn (SDH) of spinal cord reflect both mechanisms. We reproduced SDH neuron spiking patterns by varying densities of KV 1- and A-type potassium conductances. Plotting the spiking patterns that emerge from different density combinations revealed spiking-pattern regions separated by boundaries (bifurcations). This map suggests that certain spiking pattern combinations occur when the distribution of potassium channel densities straddle boundaries, whereas other spiking patterns reflect distinct patterns of ion channel expression. The former mechanism may explain why certain spiking patterns co-occur in genetically identified neuron types. We also present algorithms to predict spiking pattern proportions from ion channel density distributions, and vice versa. ABSTRACT: Neurons are often classified by spiking pattern. Yet, some neurons exhibit distinct patterns under subtly different test conditions, which suggests that they operate near an abrupt transition, or bifurcation. A set of such neurons may exhibit heterogeneous spiking patterns not because of qualitative differences in which ion channels they express, but rather because quantitative differences in expression levels cause neurons to operate on opposite sides of a bifurcation. Neurons in the spinal dorsal horn, for example, respond to somatic current injection with patterns that include tonic, single, gap, delayed and reluctant spiking. It is unclear whether these patterns reflect five cell populations (defined by distinct ion channel expression patterns), heterogeneity within a single population, or some combination thereof. We reproduced all five spiking patterns in a computational model by varying the densities of a low-threshold (KV 1-type) potassium conductance and an inactivating (A-type) potassium conductance and found that single, gap, delayed and reluctant spiking arise when the joint probability distribution of those channel densities spans two intersecting bifurcations that divide the parameter space into quadrants, each associated with a different spiking pattern. Tonic spiking likely arises from a separate distribution of potassium channel densities. These results argue in favour of two cell populations, one characterized by tonic spiking and the other by heterogeneous spiking patterns. We present algorithms to predict spiking pattern proportions based on ion channel density distributions and, conversely, to estimate ion channel density distributions based on spiking pattern proportions. The implications for classifying cells based on spiking pattern are discussed.

Characterizing Orthostatic Tremor Using a Smartphone Application.

BACKGROUND: Orthostatic tremor is one of the few tremor conditions requiring an electromyogram for definitive diagnosis since leg tremor might not be visible to the naked eye. PHENOMENOLOGY SHOWN: An iOS application (iSeismometer, ObjectGraph LLC, New York) using an Apple iPhone 5 (Cupertino, CA, USA) inserted into the patient's sock detected a tremor with a frequency of 16.4 Hz on both legs. EDUCATIONAL VALUE: The rapid and straightforward accelerometer-based recordings accomplished in this patient demonstrate the ease with which quantitative analysis of orthostatic tremor can be conducted and, importantly, demonstrates the potential application of this approach in the assessment of any lower limb tremor.