Tourette syndrome results from a complex interaction between social–environmental factors, multiple genetic abnormalities, and neurotransmitter disturbances. This study is a double-blinded, randomized controlled trial using probiotics Lactobacillus plantarum PS128 as an intervention to examine if probiotics improve symptoms of children with Tourette syndrome. This study enrolled children aged 5 to 18 years old who fulfilled DSM-V diagnostic criteria for Tourette syndrome. Patients were assessed before initiating the trial, at one month, and at two months after randomization. The primary outcome was evaluated by Yale Global Tic Severity Scale (YGTSS), and the secondary outcome studied the possible comorbidities in these children. The results revealed no significant difference in improvement in YGTSS between the control group and the PS128 group. As for secondary endpoints, an analysis of Conners’ Continuous Performance Test (CPT) showed improvement in commission and detectability in the PS128 group. In conclusion, although probiotics may not have tic-reducing effects in children with Tourette syndrome, it may have benefits on comorbidities such as attention deficit and hyperactivity disorder (ADHD). Further studies are needed to clarify the effects of probiotics on the comorbidities of Tourette syndrome children.

Background: There is still controversy regarding the association between primary headaches and obstructive sleep apnea. We explored the relationship between tension-type headache (TTH) and obstructive sleep apnea (OSA) using a large nationwide population-based data set in Taiwan. Methods: We identified 4759 patients diagnosed with OSA from the Taiwan Longitudinal Health Insurance Database, based on polysomnography, as the OSA group. We then randomly selected 19036 subjects without OSA, matched by sex and age, to serve as the non-OSA group. The multivariate Cox proportional hazards model with matching for age and sex was used to assess the possible associations between TTH and OSA among the patients. Results: The prevalence of TTH was 10.2% among OSA patients and 7.7% among non-OSA patients (p < 0.001). The multivariate Cox proportional hazards model revealed patients with OSA were more likely to have TTH (hazard ratio, 1.18; 95% CI, 1.06–1.31) (p = 0.003) than patients in the non-OSA group. Conclusion: Patients with OSA had a higher likelihood of developing TTH than patients in the non-OSA group. Further studies of physiological patterns between OSA and TTH are needed to confirm the study findings.

Background: The temporality of household income level with overweight/obesity in children has not been extensively studied. Little research has been conducted to determine the impact of household income on the risk of childhood overweight/obesity over time. This population-based cohort study aimed to investigate the impact of household income on the risk of overweight/obesity over time among preschool-aged children in Taiwan. Methods: From 2009 to 2018, we recruited 1,482 preschool-aged children ( ≦ 7 y of age) from low-income households and selected age- and sex-matched controls from non-low-income households for comparison; All participants were selected from those who consistently participated in the Taipei Child Development Screening Program and were monitored for overweight/obesity using body mass index (BMI) until December 31, 2018. Low-income households were defined as those with an average monthly disposable income < 60% of the minimum standard of living expense in Taiwan. The primary outcome was childhood overweight or obesity in study participants, defined as BMI (kg/m²) ≥ 85th percentile or ≥ 95th percentile, respectively. The generalized estimating equations (GEE) model was used to determine the impact of low-income households on the risk of overweight/obesity in study participants. Results: Over 21,450 person-years of follow-up, 1,782 participants developed overweight /obesity, including 452 (30.5%) and 1,330 (22.4%) children from low- and non-low-income households, respectively. The GEE model showed that the first group had a significantly higher risk of becoming overweight/obese than the other during the follow-up period (adjusted odds ratio [aOR] = 1.44, 95% CI: 1.29–1.60). Moreover, children of foreign mothers had a higher risk of becoming overweight/obese than those of Taiwanese mothers during the follow-up period (aOR = 1.51, 95% CI: 1.24–1.8). The subgroup analysis revealed a significant association between low-income households and an increased risk of overweight/obesity in children aged 2–7 years (P =.01). However, this association was not observed in children aged 0–1 years (P >.999). Conclusions: During the follow-up period, there was a notable correlation between low-income households and an increased risk of preschool-aged children developing overweight or obesity. Implementing health promotion initiatives aimed at reducing overweight and obesity in this demographic is crucial.

Deep learning (DL) algorithms are used to diagnose diabetic retinopathy (DR). However, most of these algorithms have been trained using global data or data from patients of a single region. Using different model architectures (e.g., Inception-v3, ResNet101, and DenseNet121), we assessed the necessity of modifying the algorithms for universal society screening. We used the open-source dataset from the Kaggle Diabetic Retinopathy Detection competition to develop a model for the detection of DR severity. We used a local dataset from Taipei City Hospital to verify the necessity of model localization and validated the three aforementioned models with local datasets. The experimental results revealed that Inception-v3 outperformed ResNet101 and DenseNet121 with a foreign global dataset, whereas DenseNet121 outperformed Inception-v3 and ResNet101 with the local dataset. The quadratic weighted kappa score ($\kappa$) was used to evaluate the model performance. All models had 5–8% higher κ for the local dataset than for the foreign dataset. Confusion matrix analysis revealed that, compared with the local ophthalmologists’ diagnoses, the severity predicted by the three models was overestimated. Thus, DL algorithms using artificial intelligence based on global data must be locally modified to ensure the applicability of a well-trained model to make diagnoses in clinical environments.