Two-Way Feature Extraction for Speech Emotion Recognition Using Deep Learning.

Recognizing human emotions by machines is a complex task. Deep learning models attempt to automate this process by rendering machines to exhibit learning capabilities. However, identifying human emotions from speech with good performance is still challenging. With the advent of deep learning algorithms, this problem has been addressed recently. However, most research work in the past focused on feature extraction as only one method for training. In this research, we have explored two different methods of extracting features to address effective speech emotion recognition. Initially, two-way feature extraction is proposed by utilizing super convergence to extract two sets of potential features from the speech data. For the first set of features, principal component analysis (PCA) is applied to obtain the first feature set. Thereafter, a deep neural network (DNN) with dense and dropout layers is implemented. In the second approach, mel-spectrogram images are extracted from audio files, and the 2D images are given as input to the pre-trained VGG-16 model. Extensive experiments and an in-depth comparative analysis over both the feature extraction methods with multiple algorithms and over two datasets are performed in this work. The RAVDESS dataset provided significantly better accuracy than using numeric features on a DNN.

The kickstand rod technique for correction of coronal malalignment in patients with adult spinal deformity: a systematic review and pooled analysis of 97 cases.

Coronal malalignment (CM) has recently gained focus as a key predictor of functional outcomes in patients with adult spinal deformity (ASD). The kickstand rod technique has been described as a novel technique for CM correction using an accessory rod on the convex side of the deformity. This review aimed to evaluate the surgical technique and outcomes of corrective surgery using this technique. The literature search was conducted on three databases (PubMed, EMBASE, and Scopus). After reviewing the search results, six studies were shortlisted for data extraction and pooled analysis. Weighted means for surgical duration, length of stay, amount of coronal correction, and sagittal parameters were calculated. The studies included in the review were published between 2018 and 2023, with a total sample size of 97 patients. The mean age of the study cohort was 61.1 years, with female preponderance. The mean operative time was 333.6 minutes. The mean correction of CM was 5.1 cm (95% confidence interval [CI], 3.6-6.6), the mean sagittal correction was 5.6 cm (95% CI, 4.1-7.1), and the mean change in lumbar lordosis was 17° (95% CI, 10.4-24.1). Preoperative coronal imbalance and mean correction achieved postoperatively were directly related with age. The reoperation rate was 13.2%. The kickstand rod technique compares favorably with conventional techniques such as asymmetric osteotomies in CM management. This technique provides an additional accessory rod that helps increase construct stiffness. Because of limited data, definitive conclusions cannot be drawn from this review; however, this technique is a valuable tool for a surgeon dealing with ASD.

The Psychosocial Impact of COVID-19 on an Adult Indian Population.

BACKGROUND: Coronavirus disease (COVID-19) was a pandemic with many physical, psychological, and socioeconomic effects. COVID-19 caused a global increase in anxiety and depression because of its novelty, high infectivity, varied presentation, and unpredictable mortality. In the face of collapsing healthcare facilities, monetary setbacks, and loneliness because of lockdowns, people were anxious, and this was compounded by media sensationalism. We aimed to study the psychosocial impact of COVID-19 on the adult Indian population. METHODS: An online survey using SurveyMonkey was floated through WhatsApp messages in April 2020, using the 'chain-referral sampling' method. Responses from individuals >18 years were included, and questions included age, sex, occupation, demographics, and socioeconomic conditions. The prevalence of anxiety and depression was assessed using the Generalized Anxiety Disorder (GAD-7) and the Patient Health Questionnaire (PHQ-9) scales. Data was analyzed using IBM SPSS software, and predictors of anxiety and depression were assessed. RESULTS: A total of 2640 responses from individuals between 18 years and 81 years were analyzed, of which 39% were from females and 85% from those <50 years of age. There were students (15.6%), teachers (10.7%), healthcare workers (16.8%), homemakers (9%), and daily wage laborers (4.1%), among others. Nearly 80% lived in cities, 55% had salaried jobs, 37% were working from home, 22% were temporarily unemployed, 10% were feeling work stress, 11% had increased alcohol intake, and 7.5% saw an increase in domestic violence. The income of 50% was adversely affected. Nearly 50% of our respondents had some symptoms of anxiety, and 23% had significant anxiety (GAD ≥5). The presence of anxiety was significantly higher in females, younger adults, city dwellers, healthcare workers, unemployed people, individuals living away from home, those without fixed salaries, those with work stress, and in people whose incomes had been adversely affected by the pandemic. On logistic regression analysis, female sex, younger age, unemployment, lack of salaried jobs, work stress, being a healthcare worker, and media reports were independent predictors of anxiety. About 60% of our respondents had some symptoms of depression, with 26% having significant depression (PHQ-9 ≥5). The presence of depression was significantly higher in females, younger adults, city dwellers, unemployed people, individuals living away from home without fixed salaries, and people with work stress. On logistic regression analysis, younger age, female sex, unemployment, lack of salaried jobs, work stress, and media reports were independent predictors of depression. Among our respondents, 70% used the time during the lockdown to study, 77% caught up with their families, and 56% reconnected with hobbies. Nearly 88% of our respondents had adjusted to their changing circumstances, helped by their religious beliefs and faith, the support of family and friends, good government measures, and the assurance of healthcare. CONCLUSIONS:  Significant anxiety and depression were seen in 23% and 26% of respondents, respectively. Being a healthcare worker was an independent predictor of anxiety. Female sex, younger age, unemployment, work stress, and sensational media reports were independent predictors of both anxiety and depression.

The burden of post-traumatic amputations in a developing country - An epidemiological study from a level I trauma centre.

BACKGROUND: With the rapidly growing population and expanding vehicle density on the roads, there has been an upsurge in road accidents in developing countries. Knowledge about the causes and patterns of trauma-related amputations helps in the formulation of strategies for limb savage, timely management, and effective rehabilitation. OBJECTIVE: To study the epidemiology, demographic profile, and outcomes of post-trauma amputations at a level I tertiary care centre in North India. METHODS: Retrospective evaluation of the amputee data from 1st January 2018 to 31st December 2019, focusing on demographic details, injury mechanisms, surgical delays, hospital stay, and complications. RESULTS: A total of 17,445 trauma cases were seen in our trauma centre during the study period. Of these, 442 patients (2.5%) underwent major limb amputation. The hospital-based prevalence of traumatic limb amputation was 2.5%. The mean age of the amputees was 35.6years (range 1-75), and the majority were males (n = 369, 83.5%). The lower to upper limb involvement ratio was 3:1 (n = 338:105). A road traffic accident was the most common mode of injury (77.4%), followed by machine-cut injuries (16.1%). On-site traumatic amputation was seen in 23.1% (n = 102), while 43.5% had a mangled limb amputated in the hospital (mean MESS score 9.53). Overall, 27% of cases had a vascular injury after trauma, ultimately ending in limb amputation. The in-hospital mortality was 2% (n = 9/442). 43.7% of patients with a single limb amputation were discharged within 48 h. Extended hospital stay was noted in cases with associated fractures in the other limbs (28.5%), head or facial injury (9.9%), and with or without a combination of chest, abdomen, pelvic, or spine injury (7.2%). CONCLUSION: A 2.5% incidence of post-trauma amputation reflects on the severity of injury related to road and industrial accidents which predominantly affect the lower limbs at the peak of productive work life. In the absence of national amputation registries, the results underscore the need to focus on road safety protocols, patient transfer methods, and the up-gradation of local hospitals.