Faculty Bibliography

Machine learning (ML) on structural MRI data shows high potential for classifying Alzheimer's disease (AD) progression, but the specific contribution of brain regions, demographics, and proteinopathy remains unclear. Using Alzheimer's Disease Neuroimaging Initiative (ADNI) data, we applied an extreme gradient-boosting algorithm and SHAP (SHapley Additive exPlanations) values to classify cognitively normal (CN) older adults, those with mild cognitive impairment (MCI) and AD dementia patients. Features included structural MRI, CSF status, demographics, and genetic data. Analyses comprised one cross-sectional multi-class classification (CN vs. MCI vs. AD dementia, n = 568) and two longitudinal binary-class classifications (CN-to-MCI converters vs. CN stable, n = 92; MCI-to-AD converters vs. MCI stable, n = 378). All classifications achieved 70-77% accuracy and 61-83% precision. Key features were CSF status, hippocampal volume, entorhinal thickness, and amygdala volume, with a clear dissociation: hippocampal properties contributed to the conversion to MCI, while the entorhinal cortex characterized the conversion to AD dementia. The findings highlight explainable, trajectory-specific insights into AD progression.

Curbs serve as vital borders that delineate safe pedestrian zones from potential vehicular traffic hazards. Curbs also represent a primary spatial hazard during dynamic navigation with significant stumbling potential. Such vulnerabilities are particularly exacerbated for persons with blindness and low vision (PBLV). Accurate visual-based discrimination of curbs is paramount for assistive technologies that aid PBLV with safe navigation in urban environments. Herein, we investigate the efficacy of curb segmentation for foundation models. We introduce the largest curb segmentation dataset to date to benchmark leading foundation models. Our results show that state-of-the-art foundation models face significant challenges in curb segmentation. This is due to their low precision and recall with poor performance distinguishing curbs from curb-like objects or non-curb areas, such as sidewalks. In addition, the best-performing model averaged a 3.70-s inference time, underscoring problems in providing real-time assistance. In response, we propose solutions including filtered bounding box selections to achieve more accurate curb segmentation. Overall, despite the immediate flexibility of foundation models, their application for practical assistive technology applications still requires refinement. This research highlights the critical need for specialized datasets and tailored model training to address navigation challenges for PBLV and underscores implicit weaknesses in foundation models.

BACKGROUND:Neurofibromatosis 1 (NF1) is a rare, neurogenetic disorder predisposing individuals to multiple tumors and other issues requiring expert care and regular health monitoring. We sought to assess U.S. patients' access to specialized NF1 clinics and receipt of evidence-informed health surveillance. METHODS:An online survey was distributed to NF Registry participants in May 2021. Rate of NF1 clinic attendance and self-reported receipt of health surveillance amongst NF Registry participants was estimated using inverse propensity scores. Differences in these outcomes based on participant demographics were assessed using weighted logistic regression and robust linear regression, respectively. RESULTS:322 individuals responded (160 adults, 162 parents; 4.7% overall response rate). We estimate that 51.7% of children and 35.6% of adults attend NF1 clinics. Younger children were more likely to attend an NF1 clinic, as were adults living in urban areas, with a college degree or higher, or with a household income ≥ $130,000 (all ps < 0.05). Completion rates for each individual health surveillance evaluation ranged from 41 to 79% for children and 33-61% for adults. Higher rates of recommended evaluations were reported by both adults and children who attend a specialized NF1 clinic, non-Hispanic White adults, and adults with commercial or Medicare insurance (all ps < 0.05). CONCLUSIONS:Adults with NF1 experience significant sociodemographic disparities in care, and patients of all ages attending NF1 specialty clinics receive more recommended health surveillance. Given the limited access to specialty NF clinics, quality improvement efforts are needed to increase access for underserved adults and improve provision of recommended health surveillance outside specialty clinics.

Leigh syndrome is a progressive infantile neurodegenerative disorder of mitochondrial metabolism that often leads to decompensation in the setting of metabolic stress. It is genetically heterogenous with varied inheritance patterns. One subtype includes NDUFS8-related autosomal recessive Leigh syndrome. This nuclear gene encodes a complex I subunit of the mitochondrial complex chain. Although Leigh syndrome is typically associated with basal ganglia and brainstem involvement, cases of confluent white matter disease have been described with NDUFS8-related disorders. We present the case of a 6-month-old girl with initial imaging suggestive of a leukodystrophy, later found to have a novel homozygous variant in NDUFS8. In conjunction with the clinical course, a diagnosis of Leigh syndrome was made. This case highlights that mitochondrial disorders should be considered on the differential for confluent cerebral white matter disease in early childhood.

More than half of individuals with multiple sclerosis (MS) use cannabis, with up to 20 % at risk of cannabis use disorder (CUD). While some individuals with MS report symptom relief from cannabis use, particularly for pain, sleep, and mood, there is limited support for its evidence-based therapeutic use. In contrast, long-term use has been associated with poorer cognitive and emotional functioning, fatigue, and reduced quality of life. Although reducing or stopping cannabis use has shown benefits, there is a lack of accessible interventions. We recruited nationally for a pilot study of a remotely supervised home-based intervention to reduce cannabis use among women with MS and CUD. The trial response provided critical insights into cannabis use patterns and the significant demand for accessible, effective interventions, highlighting an urgent unmet need within the MS community.

BACKGROUND:Cerebral arteriovenous malformations (AVMs) are complex vascular lesions that pose a risk for hemorrhagic stroke. The number of draining veins has recently emerged as a significant predictor of rupture risk. This multicenter study aimed to evaluate the outcomes in adult AVM patients with single versus multiple draining veins. METHODS:We conducted a retrospective analysis of 735 AVM patients from the Multicenter International Study for Treatment of Brain AVMs (MISTA) database. Patients were categorized into single draining vein (n = 430) and multiple draining veins (n = 305) groups. Logistic and linear regression models were used to assess outcomes, adjusting for baseline characteristics, including age, rupture status, Spetzler-Martin grade, and other relevant factors. RESULTS:After adjustment, no significant differences were observed in complete AVM obliteration at last follow-up between the multiple and single draining veins groups (OR: 1.1; 95 % CI: 0.72-1.93, p = 0.49) after any treatment type. Good functional outcomes at last follow-up (mRS 0-2) were similar between the two groups (OR: 1.00; 95 % CI: 0.48-2.09, p = 0.98), as were retreatment rates (OR: 1.68; 95 % CI: 0.74-3.83, p = 0.21). Ruptured AVMs were more common in the single draining vein group (52.0 % vs. 35.4 %, p < 0.001). Patients in the multiple draining vein group had lower odds of hemorrhagic complications compared to the single vein group (OR: 0.38; 95 % CI: 0.14-1.02, p = 0.05). CONCLUSION/CONCLUSIONS:Single draining vein AVMs were more likely to present with rupture, but no significant differences in obliteration rates, functional outcomes, or retreatment rates were found between the groups after adjustment. These findings suggest that while venous drainage patterns may influence initial presentation, they do not appear to affect overall treatment success or patient prognosis after any treatment type. Further studies are needed to confirm.

STATEMENT OF THE CLINICAL PROBLEM ADDRESSED BY THE CASE/UNASSIGNED:We report an atypical clinical presentation of a rapidly progressive neurologic emergency that required prompt investigation and treatment of impending respiratory failure. We discuss the differential diagnosis, evaluation, emergency management, and treatment options of patients with atypical variants of this disorder. BRIEF DESCRIPTION OF CASE PRESENTATION/UNASSIGNED:A 56-year-old woman with a history of hypothyroidism, anxiety, and depression presented to the emergency department 3 weeks after an upper respiratory and ear infection with cough, pain with sinus palpation, tingling in her fingers bilaterally and right foot, hives, and an episode of blurry vision on awakening. She was discharged home with antibiotics. That evening, she developed rapidly progressing hoarseness and dysphagia and returned to the emergency department. An initial examination and laryngoscopy revealed complete left vocal cord paralysis, consistent with a left cranial nerve X palsy, which prompted a neurologic evaluation. Her examination progressively worsened over the next day requiring mechanical ventilation and ICU admission. SUMMARY OF THE KEY TEACHING POINT IN THE CASE/UNASSIGNED:New-onset bulbar cranial neuropathies should raise concern for neurologic disorders that can be rapidly progressive and result in respiratory failure. Urgent diagnosis and treatment are warranted.

OBJECTIVE/UNASSIGNED:The authors aimed to characterize the long-term psychiatric outcomes and their predictors among survivors of autoimmune encephalitis (AE). METHODS/UNASSIGNED:In this retrospective cohort study, patients diagnosed as having AE between 2008 and 2023 at two academic medical centers (in New York City and Toronto) completed the Mini International Neuropsychiatric Interview 7.0.2 (MINI) and Profile of Mood States (POMS-2) to assess long-term psychiatric outcomes. Clinical characteristics were assessed for potential predictors of psychiatric outcomes. Bivariate analyses and univariate logistic regressions were conducted to assess the relationship between the predictors and the primary outcome. RESULTS/UNASSIGNED:-methyl-d-aspartate (33%), anti-leucine-rich-glioma-inactivated 1 (24%), anti-glutamic acid decarboxylase 65 (14%), and antibody-negative encephalitis (29%). In total, 71% of participants who completed the MINI met criteria for a DSM-5 diagnosis, and 56% were diagnosed as having a mood disorder. Thirteen participants (31%) reported above-average total mood disturbance on the POMS-2. Mann-Whitney U tests revealed that participants diagnosed as having a mood disorder self-reported significantly higher levels of confusion and bewilderment (z=-2.04, p=0.04) and depression and dejection (z=-2.24, p=0.03) and lower levels of vigor and activity (z=-2.62, p=0.01). CONCLUSIONS/UNASSIGNED:AE survivors have a high prevalence of psychiatric comorbid conditions, with most being diagnosed as having a mood disorder and a significant proportion endorsing ongoing mood disturbance. Patients with a psychiatric history may benefit from closer psychiatric follow-up.

Space occupying cerebral edema is a feared complication after large ischemic stroke, occurring in up to 30% of patients with middle cerebral artery (MCA) occlusion and peaking 2-4 days after injury. Little is known about the factors and outcomes associated with peak edema timing, especially after 96 h. We aimed to characterize differences and compare discharge status between patients who experienced maximum midline shift (MLS) or decompressive hemicraniectomy (DHC) in the acute (< 48 h), average (48-96 h), and subacute (> 96 h) groups. We performed a two-center, retrospective study of patients with ≥ 1/2 MCA territory infarct and MLS. We constructed a multivariable model to test the association of subacute peak edema and favorable discharge status, adjusting for various confounders. Of 321 eligible patients, 32%, 36%, and 32% experienced acute, average, and subacute peak edema. Subacute peak edema was significantly associated with higher odds of favorable discharge than acute peak edema (aOR, 2.05; 95% CI, 1.03-4.11). Subacute peak edema after large MCA stroke is associated with better discharge status compared to acute peak edema courses. Understanding how the timing of cerebral edema affects risk of unfavorable discharge has important implications for treatment decisions and prognostication.