Human vs. Machine Learning Based Detection of Facial Weakness Using Video Analysis.

Background: Current EMS stroke screening tools facilitate early detection and triage, but the tools' accuracy and reliability are limited and highly variable. An automated stroke screening tool could improve stroke outcomes by facilitating more accurate prehospital diagnosis and delivery. We hypothesize that a machine learning algorithm using video analysis can detect common signs of stroke. As a proof-of-concept study, we trained a computer algorithm to detect presence and laterality of facial weakness in publically available videos with comparable accuracy, sensitivity, and specificity to paramedics. Methods and Results: We curated videos of people with unilateral facial weakness (n = 93) and with a normal smile (n = 96) from publicly available web-based sources. Three board certified vascular neurologists categorized the videos according to the presence or absence of weakness and laterality. Three paramedics independently analyzed each video with a mean accuracy, sensitivity and specificity of 92.6% [95% CI 90.1-94.7%], 87.8% [95% CI 83.9-91.7%] and 99.3% [95% CI 98.2-100%]. Using a 5-fold cross validation scheme, we trained a computer vision algorithm to analyze the same videos producing an accuracy, sensitivity and specificity of 88.9% [95% CI 83.5-93%], 90.3% [95% CI 82.4-95.5%] and 87.5 [95% CI 79.2-93.4%]. Conclusions: These preliminary results suggest that a machine learning algorithm using computer vision analysis can detect unilateral facial weakness in pre-recorded videos with an accuracy and sensitivity comparable to trained paramedics. Further research is warranted to pursue the concept of augmented facial weakness detection and external validation of this algorithm in independent data sets and prospective patient encounters.

Video-Based Facial Weakness Analysis.

OBJECTIVE: Facial weakness is a common sign of neurological diseases such as Bell's palsy and stroke. However, recognizing facial weakness still remains as a challenge, because it requires experience and neurological training. METHODS: We propose a framework for facial weakness detection, which models the temporal dynamics of both shape and appearance-based features of each target frame through a bi-directional long short-term memory network (Bi-LSTM). The system is evaluated on a "in-the-wild"video dataset that is verified by three board-certified neurologists. In addition, three emergency medical services (EMS) personnel and three upper level residents rated the dataset. We compare the evaluation of the proposed algorithm with other comparison methods as well as the human raters. RESULTS: Experimental evaluation demonstrates that: (1) the proposed algorithm achieves the accuracy, sensitivity, and specificity of 94.3%, 91.4%, and 95.7%, which outperforms other comparison methods and achieves the equal performance to paramedics; (2) the framework can provide visualizable and interpretable results that increases model transparency and interpretability; (3) a prototype is implemented as a proof-of-concept showcase to show the feasibility of an inexpensive solution for facial weakness detection. CONCLUSION: The experiment results suggest that the proposed framework can identify facial weakness effectively. SIGNIFICANCE: We provide a proof-of-concept study, showing that such technology could be used by non-neurologists to more readily identify facial weakness in the field, leading to increasing coverage and earlier treatment.

Inflammatory Manifestations of Systemic Diseases in the Central Nervous System.

PURPOSE OF REVIEW: This review presents the current recommended therapeutic interventions for inflammatory disease in the central nervous system (CNS) secondary to systemic diseases of immune dysregulation. Treatment recommendations for CNS inflammation associated with rheumatologic conditions, immune-related adverse effects from immune checkpoint inhibitors (ICIs), and demyelinating disease from tumor necrosis factor-α (anti-TNFs) are explored. Additional therapeutic options for inflammation related to postviral syndromes and genetic immunodeficiencies are also discussed. RECENT FINDINGS: In addition to treatment of mild, moderate, and severe CNS rheumatologic disease as guided by the European League Against Rheumatism (EULAR), early consideration of rituximab for severe IgG4-related disease and induction with anti-TNF therapy for severe neurosarcoidosis should be considered. Although often not first line, treatment options for CNS inflammatory diseases based on disease mechanism are emerging, including tocilizumab for Behcet's disease, natalizumab for ICI associated autoimmune encephalitis, and abatacept for treatment of infiltrative disease secondary to CTLA-4 deficiency. Hematopoietic stem cell treatments represent highly efficacious but risky options for autoimmunity related to genetic immunodeficiency. SUMMARY: While early high dose steroids remains first line therapy for most CNS inflammatory conditions, a rapidly expanding arsenal of immune targeted therapies offers clinicians tailored disease specific options for treatment.

Dual Antiplatelet Therapy Is Associated With Coagulopathy Detectable by Thrombelastography in Acute Stroke.

BACKGROUND: Thrombelastography (TEG) provides a global, dynamic measure of coagulation. We examined the effect of antiplatelet (AP) medications on coagulation in patients with acute stroke as measured by TEG. METHODS: We reviewed prospectively collected data on patients presenting with acute ischemic stroke (AIS) and spontaneous intracerebral hemorrhage (ICH) between 2009 and 2014. Patient demographics and baseline TEG values were compared among 4 different drug use groups: aspirin only, clopidogrel only, both aspirin and clopidogrel, and no AP. Multivariable regression models were conducted to compare the differences in TEG components. RESULTS: A total of 202 patients were included, 139 with AIS and 63 with ICH. Forty-eight (24%) patients were taking aspirin alone, 12 (6%) were taking clopidogrel, 16 (8%) dual AP, and 126 (62%) no AP. Dual AP use was associated with prolonged mean R (time to initiate clotting) of 5.5 minutes as compared to no AP use (4.6 minutes, P = .04). Additionally, mean maximal amplitude (MA; final clot strength) and angle (rate of clot formation) were decreased in the dual AP group (MA = 59.3 mm, angle = 57.8°) as compared to the no AP group (MA = 64.5 mm, angle = 64.5°; P = .04 and P = .01, respectively). Patients on single AP therapy (either aspirin or clopidogrel) did not differ from those on no AP therapy in any TEG parameters measured. CONCLUSION: Dual AP therapy is associated with a detectable coagulopathy which may have implications in the management of patients with AIS and hemorrhagic stroke. The effects of single AP therapy may not be demonstrated by TEG.

Cocaine Use is Associated with More Rapid Clot Formation and Weaker Clot Strength in Acute Stroke Patients.

INTRODUCTION: 1.1.Cocaine use is a known risk factor for stroke and has been associated with worse outcomes. Cocaine may cause an altered coagulable state by a number of different proposed mechanisms, including platelet activation, endothelial injury, and tissue factor expression. This study analyzes the effect of cocaine use on Thrombelastography (TEG) in acute stroke patients. PATIENT AND METHODS: 1.2.Patients presenting with Acute Ischemic Stroke (AIS) and spontaneous Intracerebral Hemorrhage (ICH) to a single academic center between 2009 and 2014 were prospectively enrolled. Blood was collected for TEG analysis at the time of presentation. Patient demographics and baseline TEG values were compared between two groups: cocaine and non-cocaine users. Multivariable Quantile regression models were used to compare the median TEG components between groups after controlling for the effect of confounders. RESULTS: 1.3.91 patients were included, 53 with AIS and 38 with ICH. 8 (8.8%) patients were positive for cocaine, 4 (50%) with AIS, and 4 (50%) with ICH. There were no significant differences in age, blood pressure, platelet count, or PT/PTT between the cocaine positive and cocaine negative group. Following multivariable analysis, and adjusting for factors known to influence TEG including stroke subtype, cocaine use was associated with shortened median R time (time to initiate clotting) of 3.8 minutes compared to 4.8 minutes in non-cocaine users (p=0.04). Delta (thrombin burst) was also earlier among cocaine users (0.4 minutes) compared with non-cocaine users (0.5 min, p=0.04). The median MA and G (measurements of final clot strength) were reduced in cocaine users (MA=62.5 mm, G=7.8 dynes/cm2) compared to non-cocaine users (MA=66.5 mm, G=10.1 dynes/cm2; p=0.047, p=0.04, respectively). CONCLUSION: 1.4.Cocaine users demonstrate more rapid clot formation but reduced overall clot strength based on admission TEG values.

Thrombelastography does not predict clinical response to rtPA for acute ischemic stroke.

Thrombelastography (TEG) measures coagulation in venous blood. We hypothesized that TEG, by reflecting clot subtype and ex vivo fibrinolysis, might predict fibrinolytic response to tPA as reflected by rapid clinical improvement or hemorrhagic transformation of the infarct. 171 acute ischemic stroke patients treated with tPA were prospectively enrolled. Venous blood for TEG was drawn before and 10 min after tPA bolus. We measured rapid clinical improvement (RCI = 8 point improvement on NIHSS or total NIHSS of 0, 1 at 36 h), Hemorrhagic transformation (HT = any blood on imaging within 36 h), and hyperdense middle cerebral artery sign (HDMCA = biomarker for erythrocyte-rich clot). Multivariable regression models compared TEG parameters after adjusting for potential confounders. No differences in pre- or post-tPA TEG were found between patients with or without RCI. Also, there was no correlation between TEG and HDMCA. Clotting was slightly prolonged in patients with HT (p = 0.046). We failed to find a robust association between TEG and clinical response to tPA. It is likely that arterial clot lysis is determined by factors unrelated to coagulation status as measured by TEG in the venous circulation. It is unlikely that TEG will be useful to predict clinical response to tPA, but may help predict bleeding.

Clot Strength as Measured by Thrombelastography Correlates with Platelet Reactivity in Stroke Patients.

OBJECTIVE: Platelet reactivity may be important in the management of patients with stroke. However, degree of platelet reactivity has not been correlated with Thrombelastography (TEG(®)) parameters in stroke. We sought to detect a correlation between TEG(®) values and clot platelet reactivity in ex vivo clots of stroke patients. METHODS: We collected venous blood from 40 patients with stroke. TEG(®) measurements were carried out and residual clots were fixed in 10% formalin immediately following completion of TEG(®). The formalin specimens were embedded in paraffin blocks, cut at 4 micrometers, and stained with CD 61 (immunohistochemical stain used to detect platelets) with appropriate controls. Under light microscopy, three pathologists blinded to TEG(®) results independently graded CD61 intensity (how aggregated/intense the CD61 stained) into a low and high group, as a proposed measurement representing the platelet reactivity of the clot. We compared pre-tPA-TEG(®) values among groups with different CD 61 intensities. RESULTS: After adjusting for confounding factors, we found statistically significant correlation between CD61 staining and several TEG(®) parameters (Delta and CD61 staining intensity (p=0.047); Angle and CD61 staining intensity grade (p=0.04); and G and CD61 staining intensity grade (p=0.04)). CONCLUSIONS: Clot strength on TEG(®) as measured by Delta, Angle, and G correlates with a clot with greater platelet reactivity.

Iodinated contrast does not alter clotting dynamics in acute ischemic stroke as measured by thromboelastography.

BACKGROUND AND PURPOSE: Iodinated contrast agents used for computed tomography angiography (CTA) may alter fibrin fiber characteristics and decrease fibrinolysis by tissue plasminogen activator (tPA). Thromboelastography (TEG) measures the dynamics of coagulation and correlates with thrombolysis in acute ischemic stroke patients. We hypothesized that receiving CTA before tPA will not impair thrombolysis as measured by TEG. METHODS: Acute ischemic stroke patients receiving 0.9 mg/kg tPA <4.5 hours of symptom onset were prospectively enrolled. For CTA, 350 mg/dL of iohexol or 320 mg/dL of iodixanol at a dose of 2.2 mL/kg was administered. TEG was measured before tPA and 10 minutes after tPA bolus. CTA timing was left to the discretion of the treating physician. RESULTS: Of 136 acute ischemic stroke patients who received tPA, 47 had CTA before tPA bolus, and 42 had either CTA after tPA and post-tPA TEG draw or no CTA (noncontrast group). Median change in clot lysis (LY30) after tPA was 95.3% in the contrast group versus 95.0% in the noncontrast group (P=0.74). Thus, tPA-induced thrombolysis did not differ between contrast and noncontrast groups. Additionally, there was no effect of contrast on any pre-tPA TEG value. CONCLUSIONS: Our data do not support an effect of iodinated contrast agents on clot formation or tPA activity.

GABAA receptor activation in the lateral septum reduces the expression of conditioned defeat and increases aggression in Syrian hamsters.

Exposure to social stressors can cause profound changes in an individual's physiology and behavior. In Syrian hamsters, even a single social defeat results in conditioned defeat, which includes an abolishment of territorial aggression and the emergence of high levels of submissive behavior. The purpose of the current study was to determine whether the lateral septum (LS) is a component of the putative neural circuit underlying conditioned defeat. Experiment 1 explored the possibility that plasticity in the LS is necessary for the induction of conditioned defeat. Infusions of the protein synthesis inhibitor, anisomycin, prior to defeat training, however, failed to alter conditioned defeat during testing on the following day, suggesting that synaptic plasticity in the LS is not critical for defeat-induced suppression of aggression. Experiment 2 tested whether the LS is necessary for the expression of conditioned defeat. Infusions of the GABA(A) agonist muscimol into the LS prior to testing significantly increased aggression and decreased submission in previously defeated animals suggesting that the LS is an important component of the neural circuit mediating the expression of both aggression and submission in conditioned defeat. Experiment 3 examined whether the effects of muscimol on aggression were dependent on prior social defeat. Non-defeated animals receiving muscimol infusions prior to testing with a non-aggressive intruder displayed significantly more aggression than did hamsters receiving control injections. Thus, these data suggest that the activation of GABA(A) receptors in the LS increases aggression regardless of whether or not a hamster has previously experienced social defeat.