fMRI volume classification using a 3D convolutional neural network robust to shifted and scaled neuronal activations

Vu, Hanh; Kim, Hyun-Chul; Jung, Myung-Pyo; Lee, Jong Hwan

doi:10.1016/j.neuroimage.2020.117328

Cited by 18 publications

(18 citation statements)

References 81 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous studies have employed some visualizations to build an interpretable brain decoding model in fMRI analysis (Vu et al, 2020;X. Wang et al, 2020;Yin et al, 2020).…”

Section: Visualization Of Attention Mask On the Hcp Datasetmentioning

confidence: 99%

“…The second challenge is the researchers' requirement for a higher degree of accountability of the model, which is the core of the feasibility and reproducibility of brain decoding (Lindsay, 2020). Deep learning is regarded as a black‐box model, and recent efforts have been made to develop an interpretable brain decoding model through feature ranking (Li & Fan, 2019), visualizing the convolutional kernels (Vu, Kim, Jung, & Lee, 2020), guided back‐propagation (X. Wang et al, 2020), and so on. Improved DNN interpretability in fMRI analysis could lead to more accountable usage, better algorithm maintenance and improvement, and more open science (Tjoa & Guan, 2021).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Attention module improves both performance and interpretability of four‐dimensional functional magnetic resonance imaging decoding neural network

Jiang

Wang

Shi

et al. 2022

Human Brain Mapping

View full text Add to dashboard Cite

show abstract

“…Previous studies have employed some visualizations to build an interpretable brain decoding model in fMRI analysis (Vu et al, 2020;X. Wang et al, 2020;Yin et al, 2020).…”

Section: Visualization Of Attention Mask On the Hcp Datasetmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Attention module improves both performance and interpretability of four‐dimensional functional magnetic resonance imaging decoding neural network

Jiang

Wang

Shi

et al. 2022

Human Brain Mapping

View full text Add to dashboard Cite

show abstract

“…SP was shown to give better performance than average pooling and max pooling in recent publications [18][19][20][21]. Recently, strided convolution (SC) is commonly used, which also can shrink the FMs [22,23]. Nevertheless, SC could be considered a simple pooling method, which always outputs the region's fixedposition value [24].…”

Section: Improvement I: N-conv Stochastic Poolingmentioning

confidence: 99%

PSSPNN: PatchShuffle Stochastic Pooling Neural Network for an Explainable Diagnosis of COVID-19 with Multiple-Way Data Augmentation

Wang

Zhang

Cheng

et al. 2021

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

Aim. COVID-19 has caused large death tolls all over the world. Accurate diagnosis is of significant importance for early treatment. Methods. In this study, we proposed a novel PSSPNN model for classification between COVID-19, secondary pulmonary tuberculosis, community-captured pneumonia, and healthy subjects. PSSPNN entails five improvements: we first proposed the n-conv stochastic pooling module. Second, a novel stochastic pooling neural network was proposed. Third, PatchShuffle was introduced as a regularization term. Fourth, an improved multiple-way data augmentation was used. Fifth, Grad-CAM was utilized to interpret our AI model. Results. The 10 runs with random seed on the test set showed our algorithm achieved a microaveraged F1 score of 95.79%. Moreover, our method is better than nine state-of-the-art approaches. Conclusion. This proposed PSSPNN will help assist radiologists to make diagnosis more quickly and accurately on COVID-19 cases.

show abstract

“…Thus, there was the potential for circular analysis even though the data for the CCA model training and the test data were separated by an LOOCV framework. In order to mitigate this potential issue and the weak form of cross‐validation (i.e., LOOCV) (Varoquaux et al, 2017; Vu et al, 2020), we also evaluated our analysis pipeline using the task fMRI (tfMRI) data available from the HCP dataset. Of the seven tasks in the HCP, the WM and GB tasks were adopted because we thought that the between‐subject heterogeneity compared to within‐subject homogeneity in their neuronal activations may be substantial for the working memory performance and gambling outcomes compared to more straightforward tasks such as motor tasks.…”

Section: Methodsmentioning

confidence: 99%

Mixed‐effects multilevel analysis followed by canonical correlation analysis is an effective fMRI tool for the investigation of idiosyncrasies

Kim

Lustig

et al. 2021

Human Brain Mapping

Self Cite

View full text Add to dashboard Cite

We report that regions‐of‐interest (ROIs) associated with idiosyncratic individual behavior can be identified from functional magnetic resonance imaging (fMRI) data using statistical approaches that explicitly model individual variability in neuronal activations, such as mixed‐effects multilevel analysis (MEMA). We also show that the relationship between neuronal activation in fMRI and behavioral data can be modeled using canonical correlation analysis (CCA). A real‐world dataset for the neuronal response to nicotine use was acquired using a custom‐made MRI‐compatible apparatus for the smoking of electronic cigarettes (e‐cigarettes). Nineteen participants smoked e‐cigarettes in an MRI scanner using the apparatus with two experimental conditions: e‐cigarettes with nicotine (ECIG) and sham e‐cigarettes without nicotine (SCIG) and subjective ratings were collected. The right insula was identified in the ECIG condition from the χ2‐test of the MEMA but not from the t‐test, and the corresponding activations were significantly associated with the similarity scores (r = −.52, p = .041, confidence interval [CI] = [−0.78, −0.17]) and the urge‐to‐smoke scores (r = .73, p <.001, CI = [0.52, 0.88]). From the contrast between the two conditions (i.e., ECIG > SCIG), the right orbitofrontal cortex was identified from the χ2‐tests, and the corresponding neuronal activations showed a statistically meaningful association with similarity (r = −.58, p = .01, CI = [−0.84, −0.17]) and the urge to smoke (r = .34, p = .15, CI = [0.09, 0.56]). The validity of our analysis pipeline (i.e., MEMA followed by CCA) was further evaluated using the fMRI and behavioral data acquired from the working memory and gambling tasks available from the Human Connectome Project.

show abstract

fMRI volume classification using a 3D convolutional neural network robust to shifted and scaled neuronal activations

Cited by 18 publications

References 81 publications

Attention module improves both performance and interpretability of four‐dimensional functional magnetic resonance imaging decoding neural network

Attention module improves both performance and interpretability of four‐dimensional functional magnetic resonance imaging decoding neural network

PSSPNN: PatchShuffle Stochastic Pooling Neural Network for an Explainable Diagnosis of COVID-19 with Multiple-Way Data Augmentation

Mixed‐effects multilevel analysis followed by canonical correlation analysis is an effective fMRI tool for the investigation of idiosyncrasies

Contact Info

Product

Resources

About