ORIGINAL RESEARCH article
Front. Neurol.
Sec. Artificial Intelligence in Neurology
NLF-YOLO:A Lightweight Neuro-Fusion Detector for Emotion-Related Brain Tumor Imaging
Linxi Li 1
Yang Li 2
Zhicong Ding 1
Zhengnan Yin 1
Wengang Che 1
1. School of Data Science and Engineering, Kunming City College, Kunming, China
2. Xi'an University of Architecture and Technology School of Civil Engineering, Xi'an, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Abstract
Emotion recognition serves as a crucial approach for exploring emotional states in individuals and groups. Recent evidence suggests that brain tumors can influence emotional regulation and affective processing by disrupting functional interactions among key cortical regions such as the prefrontal cortex, thereby altering emotional responses and perception. To effectively integrate medical imaging data with neurobiological signals, this study introduces NLF-YOLO, a lightweight cognitive-neuroscience-oriented brain tumor recognition framework that is designed to identify tumor regions associated with emotion regulation. The framework incorporates three core modules—EBSConv, B2LKD, and ConvMixer— where EBSConv achieves efficient feature extraction through the combination of depthwise separable and pointwise convolutions, making it particularly suitable for representing subtle structural details in brain tumor imaging and facilitating the identification of tumor regions associated with emotion regulation. B2LKD performs implicit distillation via low-dimensional bottleneck reconstruction and multi-scale attention guidance, thereby enhancing multi-scale feature representation of complex brain areas without relying on an explicit teacher network. The ConvMixer module adopts a design that integrates convolution and skip connections to improve sensitivity to fine-grained lesion characteristics, enhancing both the precision and efficiency of feature extraction in brain tumor imaging. Experimental results demonstrate that NLF-YOLO surpasses existing medical image analysis models on brain tumor datasets, achieving 94.3% mAP0.5 and 73.1% mAP0.5:0.95, with a highly efficient training throughput of 294 FPS, measured on the training set with a batch size of 16. The results indicate that NLF-YOLO provides an efficient and precise tool for brain tumor imaging feature representation, facilitating the exploration of how brain lesions affect emotion-regulation–related regions, while also meeting the real-time processing requirements of clinical applications.
Summary
Keywords
brain tumor, Medical Image Analysis, Neuroscience, sentiment analysis, Sentiment perception, YOLO
Received
27 October 2025
Accepted
23 December 2025
Copyright
© 2025 Li, Li, Ding, Yin and Che. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Linxi Li
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.