Machine unlearning in brain-inspired neural network paradigms.

Machine unlearning, which is crucial for data privacy and regulatory compliance, involves the selective removal of specific information from a machine learning model. This study focuses on implementing machine unlearning in Spiking Neuron Models (SNMs) that closely mimic biological neural network behaviors, aiming to enhance both flexibility and ethical compliance of AI models. We introduce a novel hybrid approach for machine unlearning in SNMs, which combines selective synaptic retraining, synaptic pruning, and adaptive neuron thresholding. This methodology is designed to effectively eliminate targeted information while preserving the overall integrity and performance of the neural network. Extensive experiments were conducted on various computer vision datasets to assess the impact of machine unlearning on critical performance metrics such as accuracy, precision, recall, and ROC AUC. Our findings indicate that the hybrid approach not only maintains but in some cases enhances the neural network's performance post-unlearning. The results confirm the practicality and efficiency of our approach, underscoring its applicability in real-world AI systems.

Prediction of anemia using facial images and deep learning technology in the emergency department.

Background: According to the WHO, anemia is a highly prevalent disease, especially for patients in the emergency department. The pathophysiological mechanism by which anemia can affect facial characteristics, such as membrane pallor, has been proven to detect anemia with the help of deep learning technology. The quick prediction method for the patient in the emergency department is important to screen the anemic state and judge the necessity of blood transfusion treatment. Method: We trained a deep learning system to predict anemia using videos of 316 patients. All the videos were taken with the same portable pad in the ambient environment of the emergency department. The video extraction and face recognition methods were used to highlight the facial area for analysis. Accuracy and area under the curve were used to assess the performance of the machine learning system at the image level and the patient level. Results: Three tasks were applied for performance evaluation. The objective of Task 1 was to predict patients' anemic states [hemoglobin (Hb) <13 g/dl in men and Hb <12 g/dl in women]. The accuracy of the image level was 82.37%, the area under the curve (AUC) of the image level was 0.84, the accuracy of the patient level was 84.02%, the sensitivity of the patient level was 92.59%, and the specificity of the patient level was 69.23%. The objective of Task 2 was to predict mild anemia (Hb <9 g/dl). The accuracy of the image level was 68.37%, the AUC of the image level was 0.69, the accuracy of the patient level was 70.58%, the sensitivity was 73.52%, and the specificity was 67.64%. The aim of task 3 was to predict severe anemia (Hb <7 g/dl). The accuracy of the image level was 74.01%, the AUC of the image level was 0.82, the accuracy of the patient level was 68.42%, the sensitivity was 61.53%, and the specificity was 83.33%. Conclusion: The machine learning system could quickly and accurately predict the anemia of patients in the emergency department and aid in the treatment decision for urgent blood transfusion. It offers great clinical value and practical significance in expediting diagnosis, improving medical resource allocation, and providing appropriate treatment in the future.

[Application progress of smart glasses for triage during mass casualty incident].

In mass casualty incidents (MCI), the number of casualties can far exceed the capacity of medical emergency units to treat and transport in a very short period of time. A rapid MCI triage according to the severity of their injuries, can not only effectively use limited medical resources, but also improve the survival rate of injured patients. With the emergence of artificial intelligence (AI) and augmented reality (AR), smart glasses have been developed and used in different scenarios, and have achieved remarkable results in the medical field. This article focuses on the role and advantages of smart glasses in the triage of MCI, while proposing the problems in the application of smart glasses. At the same time, we elaborate on the development status of smart glasses in the triage, and discuss the application trend and development direction of smart glasses in the triage of pre-hospital injuries.

[EFFECTIVENESS OF Ilizarov TECHNOLOGY FOR INFECTED FOREARM NONUNION].

OBJECTIVE: To explore the effectiveness and method of Ilizarov technology for the treatment of infected forearm nonunion. METHODS: Between January 2004 and March 2014, 19 patients with infected forearm nonunion were treated, including 12 males and 7 females with a mean age of 37.4 years (range, 18-62 years). The injury causes included traffic accident in 11 patients, falling from height in 4 patients, and machine twist injury in 4 patients. The patients had received surgical treatment for 1-5 times (mean, 2.7 times). Bone defects located at the radius in 10 cases, at the ulna in 7 cases, and at the radius and ulna in 2 cases. The mean time of chronic infection was 8.3 months (range, 4-16 months). The mean length of the bone defects after debridement was 3.54 cm (range, 2.2-7.5 cm). Under the guidance of C-arm fluoroscope, the Orthofix unilateral external fixator was used to fix. Distraction was performed at 7-10 days after operation, and X-ray film was taken regularly to detect the osteogenesis. RESULTS: The mean external fixation time was 6.5 months (range, 3-12 months), and the mean external fixation index was 1.72 months/cm (range, 1.14-2.15 months/cm). All patients were followed up for 35.4 months on average (range, 24-55 months). The bone union time was 3-11 months (mean, 6 months); and no recurrence of infection was observed. At last follow-up, the mean wrist range of motion (ROM) were 52.78° (range, 42-55°) in flexion and 46.53° (range, 40-60°) in extension; the mean elbow ROM were 139.23° (range, 130-150°) in flexion and 3.57° (range, 0-20°) in extension; and the mean forearm ROM were 76.68° (range, 68-90°) in pronation and 81.75° (range, 72-90°) in supination. CONCLUSIONS: Ilizarov technology for infected forearm nonunion can acquire satisfactory clinical results. Radical debridement is the key to control bone infection.

[APPLICATION OF POSTEROLATERAL APPROACH FOR TREATMENT OF BIMALLEOLAR AND TRIMALLEOLAR FRACTURES].

OBJECTIVE: To evaluate the advantage of single posterolateral approach for exposing the fragments of lateral malleolus and posterior malleolus in treating bimalleolar and trimalleolar fractures. METHODS: Between January 2013 and June 2015, 47 patients with ankle fractures were treated. Of 47 cases, 16 were male and 31 were female with an average age of 47.7 years (range, 25-65 years); the locations were the left side in 21 cases and the right side in 26 cases. According to types system of Lauge-Hansen, 11 cases were rated as pronation-external rotation and 36 cases as supination-external rotation. There were 9 cases of bimalleolar fractures and 38 cases trimalleolar fractures. CT examination showed that posterior malleolus fracture involved more than 35% of the joint surface in 11 cases, and 9 cases had comminuted posterior malleolus fracture. Of 47 cases, 44 had fresh fractures with a mean disease course of 4 days (range, 8 hours to 7 days), and 3 had old fracture with the disease course of 43, 58, and 62 days respectively. Posterior malleolus fractures were fixed with T-type plates in 12 cases, one-third tubular plates in 10 cases, and cannulated screws in 25 cases. RESULTS: The operation time was 60-100 minutes (mean, 80 minutes); the bleeding volume was 50-100 mL (mean, 72 mL). Primary healing of incision was obtained in all patients, and no postoperative complications of infection, cutaneous necrosis, deep venous thrombosis, and sural nerve injury occurred. The CT examinations after operation showed anatomical reduction of fractures. All patients were followed up from 12 to 20 months (mean, 16 months). No fixation failure or reduction loss was found at last follow-up. All patients could walk normally. The active dorsiflexion of the affected ankles was within 5° less than normal side in 43 cases (91.5%) and 5-10° less than normal side in 4 cases (8.5%). The passive dorsiflexion of the affected ankles was within 5° less than normal side in 44 cases (93.6%) and 5-10° less than normal side in 3 cases (6.4%). According to the Olerud-Molander scoring system, the results were excellent in 40 cases, good in 5 cases, and fair in 2 cases; and the excellent and good rate was 95.7% at last follow-up. CONCLUSIONS: The single posterolateral approach could simultaneously expose the fragments of lateral malleolus and posterior malleolus, with less complication of the incision, short operation time, and clear exposure of posterior malleolus. The surgery could be performed under direct vision. It especially is suitable for bimalleolar and trimalleolar fractures with lateral malleolus and posterior malleolus.