Transformer attention-based neural network for cognitive score estimation from sMRI data.

Comput Biol Med

College of Artificial Intelligence (CUIT Shuangliu Industrial College), Chengdu University of Information Technology, Chengdu, 610225, China; National Intelligent Society Comprehensive Governance Experimental Base (CUIT Shuangliu Industrial College), Chengdu University of Information Technology, Che

Published: September 2025


Category Ranking

98%

Total Visits

921

Avg Visit Duration

2 minutes

Citations

20

Article Abstract

Accurately predicting cognitive scores based on structural MRI holds significant clinical value for understanding the pathological stages of dementia and forecasting Alzheimer's disease (AD). Some existing deep learning methods often depend on anatomical priors, overlooking individual-specific structural differences during AD progression. To address these limitations, this work proposes a deep neural network that incorporates Transformer attention to jointly predict multiple cognitive scores, including ADAS, CDRSB, and MMSE. The architecture first employs a 3D convolutional neural network backbone to encode sMRI, capturing preliminary local structural information. Then an improved Transformer attention block integrated with 3D positional encoding and 3D convolutional layer to adaptively capture discriminative imaging features across the brain, thereby focusing on key cognitive-related regions effectively. Finally, an attention-aware regression network enables the joint prediction of multiple clinical scores. Experimental results demonstrate that our method outperforms some existing traditional and deep learning methods based on the ADNI dataset. Further qualitative analysis reveals that the dementia-related brain regions identified by the model hold important biological significance, effectively enhancing the performance of cognitive score prediction. Our code is publicly available at: https://github.com/lshsx/CTA_MRI.

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.compbiomed.2025.110579DOI Listing

Publication Analysis

Top Keywords

neural network
12
cognitive score
8
cognitive scores
8
deep learning
8
learning methods
8
transformer attention
8
transformer attention-based
4
attention-based neural
4
network
4
cognitive
4

Similar Publications

The widespread dissemination of fake news presents a critical challenge to the integrity of digital information and erodes public trust. This urgent problem necessitates the development of sophisticated and reliable automated detection mechanisms. This study addresses this gap by proposing a robust fake news detection framework centred on a transformer-based architecture.

View Article and Find Full Text PDF

The increasing dependence on cloud computing as a cornerstone of modern technological infrastructures has introduced significant challenges in resource management. Traditional load-balancing techniques often prove inadequate in addressing cloud environments' dynamic and complex nature, resulting in suboptimal resource utilization and heightened operational costs. This paper presents a novel smart load-balancing strategy incorporating advanced techniques to mitigate these limitations.

View Article and Find Full Text PDF

Knowledge tracing can reveal students' level of knowledge in relation to their learning performance. Recently, plenty of machine learning algorithms have been proposed to exploit to implement knowledge tracing and have achieved promising outcomes. However, most of the previous approaches were unable to cope with long sequence time-series prediction, which is more valuable than short sequence prediction that is extensively utilized in current knowledge-tracing studies.

View Article and Find Full Text PDF

Accurate prediction of time-varying dynamic parameters during the milling process is a prerequisite for chatter-free cutting of thin-walled parts. In this paper, a matrix iterative prediction method based on weighted parameters is proposed for the time-varying structural modes during the milling of thin-walled blade structures. The thin-walled blade finite element model is established based on the 4-node plate element, and the time-varying dynamic parameters of the workpiece during the cutting process can be obtained by modifying the thickness of the nodes through the constructed mesh element finite element model It is not necessary to re-divide the mesh elements of the thin-walled parts at each cutting position, thus improving the calculation efficiency of the dynamic parameters of the workpiece.

View Article and Find Full Text PDF

At present, significant progress has been made in the research of image encryption, but there are still some issues that need to be explored in key space, password generation and security verification, encryption schemes, and other aspects. Aiming at this, a digital image encryption algorithm was developed in this paper. This algorithm integrates six-dimensional cellular neural network with generalized chaos to generate pseudo-random numbers to generate the plaintext-related ciphers.

View Article and Find Full Text PDF