Moraros, J. | Search bioRxiv

bioRxiv2026

Patterns and correlates of invasive alien plant richness in China's abandoned croplands

Cropland abandonment is a widespread socio-ecological transition that can generate heterogeneous ecological outcomes, including opportunities for vegetation recovery as well as the accumulation of invasive alien plants (IAPs). Understanding broad-scale correlates of IAP richness in abandoned croplands is challenging because published studies differ in sampling design, spatial extent, and reporting detail. To provide an exploratory national-scale synthesis, we compiled literature-derived occurrence records of 57 IAP species reported from abandoned croplands across China and examined study-level recorded richness using generalized linear mixed models (GLMMs) with province as a random effect. Fourteen socio-economic and environmental predictors were screened for multicollinearity and standardized, and the number of Higher Education Institutions (HEIs) was included as a proxy for potential sampling-effort variation. Asteraceae accounted for 43.9% of recorded species, and more than 70% of taxa were classified as high-risk invaders (Levels 1-2). Recorded richness exhibited a pronounced southeast-northwest gradient, with higher values in economically developed coastal provinces. In the exploratory GLMM, regional GDP (IRR = 1.67, p < 0.05) and mean annual temperature (IRR = 1.49, p < 0.05) were positively associated with recorded richness, whereas other environmental, accessibility, soil, and sampling-effort variables showed no detectable associations. Model diagnostics indicated no evidence of overdispersion, zero inflation, or residual spatial autocorrelation. Given the heterogeneous nature of literature-derived data and the small number of study units, these results should be interpreted as broad-scale correlational patterns rather than causal drivers. The findings highlight regional socio-economic context and climatic conditions as key correlates of recorded IAP richness in abandoned croplands and provide an initial baseline for understanding invasion patterns in post-agricultural landscapes undergoing rapid land-use transitions..

ecology↗

bioRxiv2026

Pre-trained Vision Transformers for Seizure Prediction: A Reproducible Baseline with Event-Based Evaluation and Statistical Validation

BackgroundScalp electroencephalography (EEG) based seizure prediction plays a critical role in improving the quality of life for patients with drug-resistant epilepsy, offering the potential for real-time warnings and timely interventions. Despite its clinical significance and decades of research, the field still lacks an open benchmark with reproducible baselines and deployment-oriented event-level evaluation. Most prior work relies on the small and outdated Childrens Hospital Boston (CHB-MIT) dataset and reports window-level metrics only, leaving the false-alarm burden of a real warning system underspecified. In seizure prediction, the cost of false alarm is significantly high since patients may receive painful electrical stimulation to suppress seizure. Hence, false alarms per hour (FA/h) and partial AUC (pAUC) are the most deployment-relevant metrics, reflecting alarm burden and discriminability in the low-false-alarm operating region that a usable warning system can realistically tolerate. However, few studies have systematically reported such metrics. In addition, vision transformers event-level performance under deployable FA/h constraints remains underexplored, and newer backbones such as MambaVision have yet to be evaluated under this setting. MethodsIn this work, we introduce a reproducible 5-fold benchmark derived from the Temple University Hospital EEG Seizure Corpus (TUSZ) dataset, and evaluate models using a pseudo-real-time event pipeline, reporting event-level sensitivity, false alarms per hour (FA/h) and partial AUC (pAUC). All models are compared to random predictors for statistical validation. We benchmark pre-trained vision transformers (SegFormer and MambaVision) under three EEG-to-image encoding methods, including a self-proposed Temporal-Patchify encoding for SegFormer. ResultsOur proposed Temporal-Patchify encoding method achieves state-of-the-art performance. We achieved 0.61 pAUC, which is 16.2% higher than the baseline Temporal-Tile SegFormer of Parani et al. The false-alarm burden (0.40{+/-}0.28 FA/h) is 44.4% lower than the Temporal-Tile SegFormer baseline while maintaining clinically usable sensitivity (60.7%{+/-}5.0%). We further perform statistical validation against a matched Poisson random predictor, confirming performance exceeds chance. Finally, we report end-to-end inference through-put up to 920 windows/s, confirming MambaVisions fastest inference speed, exceeding SegFormer by over 20%. ConclusionsThis work bridges the gap between seizure prediction algorithms and clinically usable seizure prediction systems in real-world settings. Our findings indicate that pre-trained vision transformers, when coupled with appropriate EEG encoding methods, can achieve robust performance in low-false-alarm operating regimes, which is critical for real-world deployment. This benchmark and evaluation framework may facilitate more clinically meaningful and reproducible seizure prediction research.

bioinformatics↗