Ying Gao

Background Stroke is a leading cause of disability worldwide and is frequently accompanied by psychological and functional impairments. Acceptance and Commitment Therapy (ACT), a third-wave behavioral intervention focusing on psychological flexibility, has been increasingly applied in stroke rehabilitation. Although previous reviews have suggested beneficial effects of ACT in stroke patients, whether these benefits persist in the short term after the intervention ends, and the certainty of the evidence, remain unclear. Methods A systematic review and meta-analysis was conducted according to PRISMA guidelines and registered in PROSPERO (CRD420261325108). Randomized controlled trials evaluating ACT in stroke patients were included. Pooled effects were calculated using random-effects models, with separate analyses for post-intervention and follow-up outcomes. Sources of heterogeneity were explored using subgroup analyses and meta-regression. Risk of bias was assessed using the Cochrane Risk of Bias 2 tool, and certainty of evidence was evaluated using the GRADE framework. Results Twenty-two RCTs involving 1,975 participants were included. ACT may be associated with reductions in depressive symptoms (post-intervention: SMD = −1.32, 95% CI: −1.80 to −0.84; follow-up: SMD = −0.76, 95% CI: −1.07 to −0.45) and anxiety symptoms (post-intervention: SMD = −1.36, 95% CI: −1.90 to −0.82; follow-up: SMD = −0.51, 95% CI: −1.42 to 0.40, not significant); depressive symptoms remained statistically significant at follow-up. Heterogeneity was considerable for both outcomes at post-intervention ( I 2 = 93.68% for depression and 93.60% for anxiety), so the pooled estimates represent average effects across studies. ACT may also be associated with improvements in psychological flexibility, quality of life, and sleep quality; however, these outcomes were based on a limited number of studies and showed considerable heterogeneity ( I 2 = 74.68 to 92.26%), so they should be interpreted cautiously. However, no statistically significant effects were observed for activities of daily living, neurological function, or hope. ACT did not appear to increase dropout rates in the available studies ( K = 16; RR = 0.99, 95% CI: 0.58–1.69). Meta-regression analyses indicated that intervention duration may not be a significant moderator of treatment effects. The certainty of evidence ranged from very low to moderate, mainly due to risk of bias and inconsistency. Conclusion Acceptance and Commitment Therapy may be a useful adjunctive intervention for improving psychological outcomes in people with stroke. However, evidence for functional recovery remains insufficient. Future high-quality RCTs with standardized protocols, larger samples, and longer follow-up are needed to confirm the long-term effectiveness of ACT. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/ , identifer CRD420261325108.

To enhance traffic safety, automated vehicles (AVs) are usually programmed to operate more smoothly and follow traffic rules better. Whether these operational characteristics achieve their intended safety benefits, however, also depends on how non-human-like AVs influence nearby human drivers. Using three real-world datasets of Lyft and Waymo driverless AVs and adaptive cruise control-driven vehicles, we assess their safety impacts in car-following scenarios. Unlike previous studies, we align car-following events by similarities in lead-vehicle speed profiles and conduct quasi-experiments. Lyft AVs exhibit smoother and safer car-following, with longer time-headways and higher time-to-collision than human drivers when following aligned manual vehicles. Paradoxically, human drivers behave more riskily, with shorter time-headways and smaller time-to-collision when following aligned Lyft AVs versus manual vehicles, partly due to AV leaders’ greater stability. Similar changes appear in the other two datasets. Our work provides an in-depth understanding of how AVs influence traffic safety. Yueying Chu, Ying Gao and colleagues analyze three real-world datasets of automated vehicles using a quasi-experimental design that aligns lead-vehicle speed profiles in car-following scenarios. They found that automated vehicles drive more smoothly and safely but paradoxically induce riskier car-following behavior in human followers.

برای ارتقای ایمنی ترافیک، وسایل نقلیه خودران (AVs) معمولاً به گونه‌ای برنامه‌ریزی می‌شوند که روان‌تر عمل کنند و بهتر از قوانین ترافیکی پیروی نمایند. با این حال، تحقق مزایای ایمنی مورد نظر این ویژگی‌های عملیاتی، همچنین به این بستگی دارد که وسایل نقلیه خودران غیرانسانی چگونه بر رانندگان انسانی اطراف خود تأثیر می‌گذارند. با استفاده از سه مجموعه‌داده واقعی از وسایل نقلیه خودران Lyft و Waymo و وسایل نقلیه دارای کنترل تطبیقی سرعت، تأثیرات ایمنی آن‌ها را در سناریوهای دنبال‌کردن خودرو ارزیابی می‌کنیم. برخلاف پژوهش‌های پیشین، رویدادهای دنبال‌کردن خودرو را بر اساس شباهت در پروفایل‌های سرعت خودروی پیشرو هم‌تراز می‌کنیم و آزمایش‌های شبه‌تجربی انجام می‌دهیم. وسایل نقلیه خودران Lyft رفتار دنبال‌کردن خودرو را روان‌تر و ایمن‌تر نشان می‌دهند، با فاصله زمانی طولانی‌تر و زمان تا برخورد بالاتر نسبت به رانندگان انسانی هنگام دنبال‌کردن وسایل نقلیه دستی هم‌تراز. به طور متناقض، رانندگان انسانی رفتار پرخطرتر از خود نشان می‌دهند، با فاصله زمانی کوتاه‌تر و زمان تا برخورد کمتر هنگام دنبال‌کردن وسایل نقلیه خودران Lyft هم‌تراز در مقایسه با وسایل نقلیه دستی، که بخشی از آن به دلیل پایداری بیشتر رهبران خودران است. تغییرات مشابهی در دو مجموعه‌داده دیگر نیز مشاهده می‌شود. این پژوهش درک عمیقی از نحوه تأثیرگذاری وسایل نقلیه خودران بر ایمنی ترافیک ارائه می‌دهد. Yueying Chu، Ying Gao و همکاران سه مجموعه‌داده واقعی از وسایل نقلیه خودران را با استفاده از طراحی شبه‌تجربی که پروفایل‌های سرعت خودروی پیشرو را در سناریوهای دنبال‌کردن خودرو هم‌تراز می‌کند، تحلیل کردند. آن‌ها دریافتند که وسایل نقلیه خودران روان‌تر و ایمن‌تر رانندگی می‌کنند اما به طور متناقض رفتار دنبال‌کردن خودرو پرخطرتر را در دنبال‌کنندگان انسانی القا می‌کنند.