-
Predicting Human Brain States with Transformer
Authors:
Yifei Sun,
Mariano Cabezas,
Jiah Lee,
Chenyu Wang,
Wei Zhang,
Fernando Calamante,
Jinglei Lv
Abstract:
The human brain is a complex and highly dynamic system, and our current knowledge of its functional mechanism is still very limited. Fortunately, with functional magnetic resonance imaging (fMRI), we can observe blood oxygen level-dependent (BOLD) changes, reflecting neural activity, to infer brain states and dynamics. In this paper, we ask the question of whether the brain states rep-resented by…
▽ More
The human brain is a complex and highly dynamic system, and our current knowledge of its functional mechanism is still very limited. Fortunately, with functional magnetic resonance imaging (fMRI), we can observe blood oxygen level-dependent (BOLD) changes, reflecting neural activity, to infer brain states and dynamics. In this paper, we ask the question of whether the brain states rep-resented by the regional brain fMRI can be predicted. Due to the success of self-attention and the transformer architecture in sequential auto-regression problems (e.g., language modelling or music generation), we explore the possi-bility of the use of transformers to predict human brain resting states based on the large-scale high-quality fMRI data from the human connectome project (HCP). Current results have shown that our model can accurately predict the brain states up to 5.04s with the previous 21.6s. Furthermore, even though the prediction error accumulates for the prediction of a longer time period, the gen-erated fMRI brain states reflect the architecture of functional connectome. These promising initial results demonstrate the possibility of developing gen-erative models for fMRI data using self-attention that learns the functional or-ganization of the human brain. Our code is available at: https://github.com/syf0122/brain_state_pred.
△ Less
Submitted 10 December, 2024;
originally announced December 2024.
-
Learning from pseudo-labels: deep networks improve consistency in longitudinal brain volume estimation
Authors:
Geng Zhan,
Dongang Wang,
Mariano Cabezas,
Lei Bai,
Kain Kyle,
Wanli Ouyang,
Michael Barnett,
Chenyu Wang
Abstract:
Brain atrophy is an important biomarker for monitoring neurodegeneration and disease progression in conditions such as multiple sclerosis (MS). An accurate and robust quantitative measurement of brain volume change is paramount for translational research and clinical applications. This paper presents a deep learning based method, DeepBVC, for longitudinal brain volume change measurement using 3D T…
▽ More
Brain atrophy is an important biomarker for monitoring neurodegeneration and disease progression in conditions such as multiple sclerosis (MS). An accurate and robust quantitative measurement of brain volume change is paramount for translational research and clinical applications. This paper presents a deep learning based method, DeepBVC, for longitudinal brain volume change measurement using 3D T1-weighted MRI scans. Trained with the intermediate outputs from SIENA, DeepBVC is designed to take into account the variance caused by different scanners and acquisition protocols. Compared with SIENA, DeepBVC demonstrates higher consistency in terms of volume change estimation across multiple time points in MS subjects; and greater stability and superior performance in scan-rescan experiments. Moreover, the results also show that DeepBVC is insensitive to acquisition variance in terms of imaging contrast, voxel resolution, random bias field and signal-to-noise ratio. Measurement robustness, automation and processing speed suggest a broad potential of DeepBVC in both research and clinical quantitative neuroimaging applications.
△ Less
Submitted 8 February, 2023;
originally announced February 2023.
-
Survival prediction using ensemble tumor segmentation and transfer learning
Authors:
Mariano Cabezas,
Sergi Valverde,
Sandra González-Villà,
Albert Clérigues,
Mostafa Salem,
Kaisar Kushibar,
Jose Bernal,
Arnau Oliver,
Xavier Lladó
Abstract:
Segmenting tumors and their subregions is a challenging task as demonstrated by the annual BraTS challenge. Moreover, predicting the survival of the patient using mainly imaging features, while being a desirable outcome to evaluate the treatment of the patient, it is also a difficult task. In this paper, we present a cascaded pipeline to segment the tumor and its subregions and then we use these r…
▽ More
Segmenting tumors and their subregions is a challenging task as demonstrated by the annual BraTS challenge. Moreover, predicting the survival of the patient using mainly imaging features, while being a desirable outcome to evaluate the treatment of the patient, it is also a difficult task. In this paper, we present a cascaded pipeline to segment the tumor and its subregions and then we use these results and other clinical features together with image features coming from a pretrained VGG-16 network to predict the survival of the patient. Preliminary results with the training and validation dataset show a promising start in terms of segmentation, while the prediction values could be improved with further testing on the feature extraction part of the network.
△ Less
Submitted 4 October, 2018;
originally announced October 2018.
