Multi-Path and Group-Loss-Based Network for Speech Emotion Recognition in Multi-Domain Datasets.

Speech emotion recognition (SER) is a natural method of recognizing individual emotions in everyday life. To distribute SER models to real-world applications, some key challenges must be overcome, such as the lack of datasets tagged with emotion labels and the weak generalization of the SER model for an unseen target domain. This study proposes a multi-path and group-loss-based network (MPGLN) for SER to support multi-domain adaptation. The proposed model includes a bidirectional long short-term memory-based temporal feature generator and a transferred feature extractor from the pre-trained VGG-like audio classification model (VGGish), and it learns simultaneously based on multiple losses according to the association of emotion labels in the discrete and dimensional models. For the evaluation of the MPGLN SER as applied to multi-cultural domain datasets, the Korean Emotional Speech Database (KESD), including KESDy18 and KESDy19, is constructed, and the English-speaking Interactive Emotional Dyadic Motion Capture database (IEMOCAP) is used. The evaluation of multi-domain adaptation and domain generalization showed 3.7% and 3.5% improvements, respectively, of the F1 score when comparing the performance of MPGLN SER with a baseline SER model that uses a temporal feature generator. We show that the MPGLN SER efficiently supports multi-domain adaptation and reinforces model generalization.

Sensor Data Acquisition and Multimodal Sensor Fusion for Human Activity Recognition Using Deep Learning.

In this paper, we perform a systematic study about the on-body sensor positioning and data acquisition details for Human Activity Recognition (HAR) systems. We build a testbed that consists of eight body-worn Inertial Measurement Units (IMU) sensors and an Android mobile device for activity data collection. We develop a Long Short-Term Memory (LSTM) network framework to support training of a deep learning model on human activity data, which is acquired in both real-world and controlled environments. From the experiment results, we identify that activity data with sampling rate as low as 10 Hz from four sensors at both sides of wrists, right ankle, and waist is sufficient in recognizing Activities of Daily Living (ADLs) including eating and driving activity. We adopt a two-level ensemble model to combine class-probabilities of multiple sensor modalities, and demonstrate that a classifier-level sensor fusion technique can improve the classification performance. By analyzing the accuracy of each sensor on different types of activity, we elaborate custom weights for multimodal sensor fusion that reflect the characteristic of individual activities.

Factors affecting clinical outcomes in treating patients with grade 1 degenerative spondylolisthesis using interspinous soft stabilization with a tension band system: a minimum 5-year follow-up.

STUDY DESIGN: Retrospective clinical study. OBJECTIVE: To explore the factors influencing the clinical outcomes and motion-preserving stabilization after interspinous soft stabilization (ISS) with a tension band system for grade 1 degenerative spondylolisthesis (DS). SUMMARY OF BACKGROUND DATA: Despite increasing recognition of the benefits of dynamic stabilization systems for treating lumbar degenerative disorders, the factors affecting the clinical and radiological outcomes of these systems have rarely been identified. METHODS: Sixty-five patients (mean age, 60.3 years) who underwent ISS with a tension band system between 2002 and 2004 were analyzed. The mean follow-up period was 72.5 months. The patients were divided according to the postsurgical clinical improvements into the optimal (n = 44) and suboptimal groups (n = 21), and the radiological intergroup differences were analyzed. Multiple linear regression analysis was performed to determine the impact of the radiological factors on the clinical outcomes. RESULTS: Significant intergroup differences were observed on the follow-up clinical examination. Radiologically, total lumbar lordosis (TLL) and segmental lumbar lordosis (SLL) were significantly improved only in the optimal group, resulting in significant intergroup differences in TLL (P = 0.023), SLL (P = 0.001), and the L1 tilt (P = 0.002). All these measures were closely associated with postoperative segmental lumbar lordosis, which also was the most influential radiological variable for the clinical parameters. CONCLUSION: In the patients with grade 1 DS, the back pain relief and functional improvement following ISS were affected by the improvements in the sagittal spinal alignment through the achievement of segmental lumbar lordosis. ISS can be an alternative treatment to fusion surgery for grade 1 DS in patients who do not require fixation or reduction.

Spinopelvic alignment after interspinous soft stabilization with a tension band system in grade 1 degenerative lumbar spondylolisthesis.

STUDY DESIGN: Retrospective clinical study. OBJECTIVE: The purpose of this study was to examine the changes in spinopelvic alignment after interspinous soft stabilization (ISS) with a tension band system and to identify the lumbosacral parameters related to those changes and to determine their impact on the clinical outcomes compared with posterior lumbar interbody fusion (PLIF) in patients with low-grade degenerative spondylolisthesis (DS). SUMMARY OF BACKGROUND DATA: The sacropelvic morphometric changes after fusion surgery have received much research attention. However, few reports have addressed the issue after use of dynamic or soft stabilization systems. METHODS: From April 2001 to November 2003, 45 patients presenting with grade 1 DS with stenosis underwent either ISS with a tension band system (ISS group) or PLIF with pedicle screw fixation (PLIF group). The mean follow-up period was 76.8 months. Three pelvic parameters, the sacral slope (SS), pelvic tilt (PT), and pelvic incidence, were investigated to address the sacropelvic morphometric change. Clinical outcomes were assessed using the visual analog scale score, the Oswestry Disability Index, and the patient's satisfaction index. RESULTS: Both groups showed significant improvements in all of the clinical outcomes, with no significant differences between groups. In the ISS group, the SS increased and PT decreased, whereas in the PLIF group, the SS decreased and PT increased, resulting in pelvic anteversion and retroversion, respectively, with significant intergroup differences in SS and PT (SS: P = 0.047; PT: P = 0.01). The positive association of lumbar lordosis with SS (r = 0.448) and its negative association with PT (r = -0.674) in the respective groups indicate the influence of changes in lumbar lordosis on pelvic positional changes. Significant correlations between follow-up segmental lumbar lordosis and the visual analog scale score for leg pain (r = -0.685) and Oswestry Disability Index score (r = -0.425) were found in the ISS group alone. CONCLUSION: Segmental lordotic change after ISS with a tension band system was the possible decisive factor in the development of pelvic anteversion while maintaining sagittal lumbar balance; lack of lumbar lordosis led to compensatory pelvic retroversion in the PLIF group. Considering the comparable clinical results with PLIF surgery and the achievement of physiologic sagittal spinopelvic balance, the ISS procedure can be a feasible alternative to fusion surgery in patients with grade 1 DS with stenosis.

Carotid artery-hypoglossal nerve relationships in the neck: an anatomical work.

OBJECTIVE: To review the surgical anatomy of the hypoglossal nerve in the neck, analyse its relationship to surrounding structures and offer landmarks to identify the nerve during carotid endarterectomy. METHOD: The carotid bifurcation, external carotid artery, internal carotid artery, extracranial part of the hypoglossal nerve, occipital artery, sternocleidomastoid artery and surrounding neurovascular structures were dissected and studied on 15 formalin-fixed adult cadaver heads (30 sides and 15 pairs) via a surgical microscope. Landmarks for the hypoglossal nerve and measurements of its distance from the carotid bifurcation are described. The relationship between the sternocleidomastoid artery and the occipital artery is also described. RESULTS: The distance from the carotid bifurcation to the point at which the hypoglossal nerve crosses over the internal carotid artery was variable, ranging from 3.89 to 37.03 mm (mean, 20.95 +/- 7.78 mm). The distance from the bifurcation to the point at which the hypoglossal nerve crosses over the external carotid artery ranged from 2.63 to 29.43 mm (mean, 15.33 +/- 7.86 mm; Table 1). The sternocleidomastoid artery had a very characteristic course and close relationship with the hypoglossal nerve. Ascending for a short distance in a cranial direction, it crossed over the hypoglossal nerve and then descended toward the sternocleidomastoid muscle. The sternocleidomastoid artery originated from the occipital artery (33.4%), the external carotid artery-internal carotid artery junction (30%), the external carotid artery itself (30%) or even the lingual artery (6.6%). CONCLUSION: The relationship between the hypoglossal nerve and the carotid bifurcation is quite variable, and this explains the vulnerability of the nerve during carotid endarterectomy. The sternocleidomastoid artery is a good landmark for identifying the hypoglossal nerve. If there is exact anatomical knowledge about the relationship between the sternocleidomastoid artery and the hypoglossal nerve, the incidence of nerve injuries during carotid endarterectomy can be minimized.

Pure epidural cavernous hemangioma of the cervical spine that presented with an acute sensory deficit caused by hemorrhage.

Pure epidural cavernous hemangioma of the spine without vertebral involvement is rare. Due to the slow growth of this lesion, the most common symptoms are chronic pain, myelopathy, and radiculopathy. In our case, the patient complained of an acute onset sensory deficit of the C4 dermatome. An MRI revealed an epidural mass with an acute hematoma. Here, we report a case of a pure epidural cavernous hemangioma that presented with acute neurologic symptoms caused by intralesional hemorrhage and an acute epidural hematoma, which were demonstrated on the patient's MRI.

Comparative radiologic evaluation of percutaneous endoscopic lumbar discectomy and open microdiscectomy: a matched cohort analysis.

BACKGROUND: Only a few studies comparing percutaneous endoscopic discectomy and open discectomy have been reported in the literature. The purpose of this study was to compare the radiographic changes in patients treated with percutaneous endoscopic lumbar discectomy (PELD) with those of patients treated with open lumbar microdiscectomy (OLM). METHODS: A total of 30 patients who underwent PELD with a minimum three years of follow-up were randomly selected. To compare with the PELD group, 30 patients who underwent OLM during the same period were also randomly selected according to sex, age, and disc level. The clinical outcomes were evaluated by the Macnab criteria. Statistical analysis was performed using independent sample t-test, paired sample t-test, chi-square test, Fisher's exact test, and analysis of variance (ANOVA). RESULTS: The successful clinical outcomes were 96.7% in the PELD group and 93.3% in the OLM group. Among the various radiological parameters, changes of disc height (1.41 +/- 1.19 mm in the PELD group and 2.29 +/- 2.12 mm in the OLM group, p=0.024) and foraminal height (1.26 +/- 0.91 mm in the PELD group and 1.85 +/- 0.92 mm in the OLM group, p=0.017) were significantly different between the two groups. CONCLUSIONS: Although the clinical outcomes were similarly satisfactory in both groups, PELD is a less invasive procedure than open microdiscectomy in s elected cases.

Renal cement embolism during percutaneous vertebroplasty.

Percutaneous vertebroplasty (PVP) is an effective treatment for lesions of the vertebral body that involves a percutaneous injection of polymethylmethacrylate (PMMA). Although PVP is considered to be minimally invasive, complications can occur during the procedure. We encountered a renal embolism of PMMA in a 57-year-old man that occurred during PVP. This rare case of PMMA leakage occurred outside of the anterior cortical fracture site of the L1 vertebral body, and multiple tubular bone cements migrated to the course of the renal vessels via the valveless collateral venous network surrounding the L1 body. Although the authors could not explain the exact cause of the renal cement embolism, we believe that physicians should be aware of the fracture pattern, anatomy of the vertebral venous system, and careful fluoroscopic monitoring to minimize the risks during the PVP.