Quantitative Biology > Neurons and Cognition
[Submitted on 30 Jan 2018 (this version), latest version 4 Dec 2019 (v3)]
Title:Task state decoding and mapping of individual four-dimensional fMRI time series using deep neural network
View PDFAbstract:Decoding specific task states from human fMRI (functional magnetic resonance imaging) signals constitutes a major goal of neuroimaging. Previous studies of task-state classification have focused largely on processed fMRI characteristics (e.g. functional connectivity patterns, activity patterns) or a single fMRI volume. Here, we assembled a novel deep neural network (DNN) for classifying directly from 4-D fMRI time series, retaining both spatial and temporal information of fMRI signals. We trained and tested the DNN classifier by task fMRI data from human connectome project (HCP) S1200 dataset. Reading just a single block of fMRI time series of each task for each individual, the novel classification method identified the seven tasks with 91.75% accuracy in test. Then we extracted task heatmaps of each time series by a guided backpropagation approach, and found that the heatmaps had a high similarity with the group average result of traditional contrast of parameter estimate (COPE). Furthermore, the DNN could be applied to small dataset. The features from the intermediate layer of the DNN were trained and used to train a support vector machine (SVM) to classify four conditions within 300 blocks of task-fMRI scans, data amount comparable to a usual neuroimaging research, and obtained a 95% accuracy. Our reported findings were the first to demonstrate the ability of DNN to classify and extract heatmaps from a 4-D fMRI time series, as well as its ability to be transfer trained by a small dataset. This approach should enable decoding and mapping subject's task status and furthermore help with classification and diagnosis of neuropsychiatric diseases.
Submission history
From: Xiaoxiao Wang [view email][v1] Tue, 30 Jan 2018 05:58:18 UTC (698 KB)
[v2] Tue, 17 Sep 2019 03:27:16 UTC (1,828 KB)
[v3] Wed, 4 Dec 2019 06:40:15 UTC (1,799 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.