Faculty Bibliography

Automatic action identification from video and kinematic data is an important machine learning problem with applications ranging from robotics to smart health. Most existing works focus on identifying coarse actions such as running, climbing, or cutting vegetables, which have relatively long durations and a complex series of motions. This is an important limitation for applications that require identification of more elemental motions at high temporal resolution. For example, in the rehabilitation of arm impairment after stroke, quantifying the training dose (number of repetitions) requires differentiating motions with sub-second durations. Our goal is to bridge this gap. To this end, we introduce a large-scale, multimodal dataset, StrokeRehab, as a new action-recognition benchmark that includes elemental short-duration actions labeled at a high temporal resolution. StrokeRehab consists of high-quality inertial measurement unit sensor and video data of 51 stroke-impaired patients and 20 healthy subjects performing activities of daily living like feeding, brushing teeth, etc. Because it contains data from both healthy and impaired individuals, StrokeRehab can be used to study the influence of distribution shift in action-recognition tasks. When evaluated on StrokeRehab, current state-of-the-art models for action segmentation produce noisy predictions, which reduces their accuracy in identifying the corresponding sequence of actions. To address this, we propose a novel approach for high-resolution action identification, inspired by speech-recognition techniques, which is based on a sequence-to-sequence model that directly predicts the sequence of actions. This approach outperforms current state-of-the-art methods on StrokeRehab, as well as on the standard benchmark datasets 50Salads, Breakfast, and Jigsaws.

OBJECTIVE:One of the clinical hallmarks of tuberous sclerosis complex (TSC) is radiologically identified cortical tubers, which are present in most patients. Intractable epilepsy may require surgery, often involving invasive diagnostic procedures such as intracranial electroencephalography (EEG). Identifying the location of the dominant tuber responsible for generating epileptic activities is a critical issue. However, the link between cortical tubers and epileptogenesis is poorly understood. Given this, we hypothesized that tuber voxel intensity may be an indicator of the dominant epileptogenic tuber. Also, via tuber segmentation based on deep learning, we explored whether an automatic quantification of the tuber burden is feasible. METHODS:We annotated tubers from structural magnetic resonance images across 29 TSC subjects, summarized tuber statistics in eight brain lobes, and determined suspected epileptogenic lobes from the same group using EEG monitoring data. Then, logistic regression analyses were performed to demonstrate the linkage between the statistics of cortical tuber and the epileptogenic zones. Furthermore, we tested the ability of a neural network to identify and quantify tuber burden. RESULTS:Logistic regression analyses showed that the volume and count of tubers per lobe, not the mean or variance of tuber voxel intensity, were positively correlated with electrophysiological data. In 47.6% of subjects, the lobe with the largest tuber volume concurred with the epileptic brain activity. A neural network model on the test dataset showed a sensitivity of .83 for localizing individual tubers. The predicted masks from the model correlated highly with the neurologist labels, and thus may be a useful tool for determining tuber burden and searching for the epileptogenic zone. SIGNIFICANCE:We have proven the feasibility of an automatic segmentation of tubers and a derivation of tuber burden across brain lobes. Our method may provide crucial insights regarding the treatment and outcome of TSC patients.

Post-zygotically acquired genetic variants, or somatic variants, that arise during cortical development have emerged as important causes of focal epilepsies, particularly those due to malformations of cortical development. Pathogenic somatic variants have been identified in many genes within the PI3K-AKT-mTOR-signalling pathway in individuals with hemimegalencephaly and focal cortical dysplasia (type II), and more recently in SLC35A2 in individuals with focal cortical dysplasia (type I) or non-dysplastic epileptic cortex. Given the expanding role of somatic variants across different brain malformations, we sought to delineate the landscape of somatic variants in a large cohort of patients who underwent epilepsy surgery with hemimegalencephaly or focal cortical dysplasia. We evaluated samples from 123 children with hemimegalencephaly (n = 16), focal cortical dysplasia type I and related phenotypes (n = 48), focal cortical dysplasia type II (n = 44), or focal cortical dysplasia type III (n = 15). We performed high-depth exome sequencing in brain tissue-derived DNA from each case and identified somatic single nucleotide, indel and large copy number variants. In 75% of individuals with hemimegalencephaly and 29% with focal cortical dysplasia type II, we identified pathogenic variants in PI3K-AKT-mTOR pathway genes. Four of 48 cases with focal cortical dysplasia type I (8%) had a likely pathogenic variant in SLC35A2. While no other gene had multiple disease-causing somatic variants across the focal cortical dysplasia type I cohort, four individuals in this group had a single pathogenic or likely pathogenic somatic variant in CASK, KRAS, NF1 and NIPBL, genes previously associated with neurodevelopmental disorders. No rare pathogenic or likely pathogenic somatic variants in any neurological disease genes like those identified in the focal cortical dysplasia type I cohort were found in 63 neurologically normal controls (P = 0.017), suggesting a role for these novel variants. We also identified a somatic loss-of-function variant in the known epilepsy gene, PCDH19, present in a small number of alleles in the dysplastic tissue from a female patient with focal cortical dysplasia IIIa with hippocampal sclerosis. In contrast to focal cortical dysplasia type II, neither focal cortical dysplasia type I nor III had somatic variants in genes that converge on a unifying biological pathway, suggesting greater genetic heterogeneity compared to type II. Importantly, we demonstrate that focal cortical dysplasia types I, II and III are associated with somatic gene variants across a broad range of genes, many associated with epilepsy in clinical syndromes caused by germline variants, as well as including some not previously associated with radiographically evident cortical brain malformations.

OBJECTIVE:This study aimed to assess the clinical outcome and outcome predictive factors in pediatric epilepsy patients evaluated with stereo-electroencephalography (SEEG). METHODS:Thirty-eight patients who underwent SEEG implantation at the Pediatric Epilepsy Center in New York Presbyterian Hospital between June 2014 and December 2019 were enrolled for retrospective chart review. Postoperative seizure outcomes were evaluated in patients with at least 12-months follow up. Meta-analysis was conducted via electronic literature search of data reported from 2000 to 2020 to evaluate significant surgical outcome predictors for SEEG evaluation in the pediatric population. RESULTS:In the current case series of 25 postsurgical patients with long-term follow up, 16 patients (64.0%) were seizure free. An additional 7 patients (28.0%) showed significant seizure improvement and 2 patients (8.0%) showed no change in seizure activity. Patients with nonlesional magnetic resonance imaging (MRI) achieved seizure freedom in 50% (5/10) of cases. By comparison, 73% (11/15) of patients with lesional MRI achieved seizure freedom. Out of 12 studies, 158 pediatric patients were identified for inclusion in a meta-analysis of the effectiveness of SEEG. Seizure freedom was reported 54.4% (n = 86/158) of patients at last follow up. Among patients with nonlesional MRI, 45% (n = 24) achieved seizure freedom compared with patients with lesional MRI findings (61.2%, n:= 60) (p = 0.02). The risk for seizure recurrence was 2.15 times higher [95% confidence interval [CI] 1.06-4.37, p = 0.033] in patients diagnosed with nonlesional focal epilepsy compared to those with lesional epilepsy [ 1.49 (95% CI 1.06-2.114, p = 0.021]. CONCLUSION:Evaluation by SEEG implantation in pediatric epilepsy is effective in localizing the epileptogenic zone with favorable outcome. Presence of a non-lesional brain MRI was associated with lower chances of seizure freedom. Further research is warranted to improve the efficacy of SEEG in localizing the epileptogenic zone in pediatric patients with non-lesional brain MRI.

BACKGROUND:Safe sampling of central nervous system tumor tissue for diagnostic purposes may be difficult if not impossible, especially in pediatric patients, and an unmet need exists to develop less invasive diagnostic tests. METHODS:We report our clinical experience with minimally invasive molecular diagnostics using a clinically validated assay for sequencing of cerebrospinal fluid (CSF) cell-free DNA (cfDNA). All CSF samples were collected as part of clinical care, and results reported to both clinicians and patients/families. RESULTS:We analyzed 64 CSF samples from 45 pediatric, adolescent and young adult (AYA) patients (pediatric = 25; AYA = 20) with primary and recurrent brain tumors across 12 histopathological subtypes including high-grade glioma (n = 10), medulloblastoma (n = 10), pineoblastoma (n = 5), low-grade glioma (n = 4), diffuse leptomeningeal glioneuronal tumor (DLGNT) (n = 4), retinoblastoma (n = 4), ependymoma (n = 3), and other (n = 5). Somatic alterations were detected in 30/64 samples (46.9%) and in at least one sample per unique patient in 21/45 patients (46.6%). CSF cfDNA positivity was strongly associated with the presence of disseminated disease at the time of collection (81.5% of samples from patients with disseminated disease were positive). No association was seen between CSF cfDNA positivity and the timing of CSF collection during the patient's disease course. CONCLUSIONS:We identified three general categories where CSF cfDNA testing provided additional relevant diagnostic, prognostic, and/or therapeutic information, impacting clinical assessment and decision making: (1) diagnosis and/or identification of actionable alterations; (2) monitor response to therapy; and (3) tracking tumor evolution. Our findings support broader implementation of clinical CSF cfDNA testing in this population to improve care.

The standard neurology clinical experience in medical school focuses primarily on bedside patient encounters; however, the limitations of the clinical environment due to the current COVID-19 pandemic have accelerated the need for virtual curriculum development. To provide guidance to Neurology clerkship directors during this unprecedented time, the American Academy of Neurology (AAN) Undergraduate Education Subcommittee (UES) formed a workgroup to develop an outline for a virtual curriculum, provide recommendations, and describe models of integrating virtual curricula into the neurology clerkship. In this overview, we discuss different methods of virtual instruction, hybrid models of clerkship training and the challenges to its implementation, professionalism issues, and modification of feedback and assessment techniques specific to the virtual learning environment. We also offer suggestions for implementation of a hybrid virtual curriculum into the neurology clerkship. The virtual curriculum is intended to supplement the core neurology in-person clinical experience and should not be used for shortening or replacing the required neurology clinical clerkship.

BACKGROUND AND OBJECTIVES/OBJECTIVE:To report the tolerability and efficacy of olaparib with temozolomide (TMZ) for glioma. METHODS:Single-center retrospective series of patients with glioma treated with olaparib/TMZ from September 2018 to December 2021. RESULTS:-wildtype gliomas (0/3). Progression-free survival was 7.8, 1.3, and 2.0 months, respectively. DISCUSSION/CONCLUSIONS:-mutant grade 2-3 gliomas with encouraging progression-free survival and manageable toxicity. This supports a prospective trial of olaparib/TMZ for this population. CLASSIFICATION OF EVIDENCE/METHODS:This case series provides Class IV evidence that treatment with olaparib/TMZ may result in radiographic response in patients with glioma.

BACKGROUND:High-dose methotrexate (HD-MTX) has broad use in the treatment of central nervous system (CNS) malignancies but confers significant toxicity without inpatient hydration and monitoring. Glucarpidase is a bacterial recombinant enzyme dosed at 50 units (u)/kg, resulting in rapid systemic MTX clearance. The aim of this study was to demonstrate feasibility of low-dose glucarpidase to facilitate MTX clearance in patients with CNS lymphoma (CNSL). METHODS:and glucarpidase 2000 or 1000u 24 h later. Treatments repeated every 2 weeks up to 8 cycles. RESULTS:Fifty-five treatments were administered. Glucarpidase 2000u yielded > 95% reduction in plasma MTX within 15 min following 33/34 doses (97.1%) and glucarpidase 1000u yielded > 95% reduction following 15/20 doses (75%). Anti-glucarpidase antibodies developed in 4 patients and were associated with MTX rebound. In CSF, glucarpidase was not detected and MTX levels remained cytotoxic after 1 (3299.5 nmol/L, n = 8) and 6 h (1254.7 nmol/L, n = 7). Treatment was safe and well-tolerated. Radiographic responses in 6 of 8 patients (75%) were as expected following MTX-based therapy. CONCLUSIONS:This study demonstrates feasibility of planned-use low-dose glucarpidase for MTX clearance and supports the hypothesis that glucarpidase does not impact MTX efficacy in the CNS. CLINICAL TRIAL REGISTRATION/BACKGROUND:NCT03684980 (Registration date 26/09/2018).

INTRODUCTION/UNASSIGNED:Older adults constitute a large and growing proportion of the population and have unique care needs in the emergency department (ED) setting. The geriatric ED accreditation program aims to improve emergency care provided to older adults by standardizing care provided across accredited geriatric EDs (GED) and through implementation of geriatric-specific care processes. The purpose of this study was to evaluate select care processes at accredited level 1 and level 2 GEDs. METHODS/UNASSIGNED:selected five GED care processes for analysis: initiatives related to delirium, screening for dementia, assessment of function and functional decline, geriatric falls, and minimizing medication-related adverse events. For all protocols, a trained research assistant abstracted information on the tool used or care process, which patients received the interventions, and staff members were involved in the care process; additional information was abstracted specific to individual care processes. RESULTS/UNASSIGNED:A total of 35 level 1 and 2 GEDs were included in this analysis. Among care processes studied, geriatric falls were the most common (31 GEDs, 89%) followed by geriatric pain management (25 GEDs, 71%), minimizing the use of potentially inappropriate medications (24 EDs, 69%), delirium (22 GEDs, 63%), medication reconciliation (21 GEDs, 60%), functional assessment (20 GEDs, 57%), and dementia screening (17 GEDs, 49%). For protocols related to delirium, dementia, function, and geriatric falls, sites used an array of different screening tools and there was heterogeneity in who performed the screening and which patients were assessed. Medication reconciliation protocols leveraged pharmacists, pharmacy technicians and/or nurses. Protocols on avoiding potentially inappropriate medication administration generally focused on ED administration of medications and used the BEERs criteria, and few sites indicated whether pain medications protocols had dosing modifications for age and/or renal function. CONCLUSION/UNASSIGNED:This study provides a snapshot of care processes implemented in level 1 and level 2 accredited GEDs and demonstrates significant heterogeny in how these care processes are implemented.