SCIENCE@home

ADHD Classification Based on fMRI Spatial-Temporal Features Using Monofractal and Multifractal Dimensions

Mengyunqiu Zhang,Sei-ichiro Kamata

Published 2023 in International Conference on Biomedical Signal and Image Processing

ABSTRACT

Attention Deficit Hyperactivity Disorder (ADHD) is one of the most common mental disorders worldwide. Since it is associated with abnormal fuctional connectivity of brain activity, current researches tend to extract biomarkers and construct objective ADHD diagnosis systems by resting state functional magnetic resonance imaging (rs-fMRI). We select 3 ADHD-related resting state networks (RSNs) to cut noise and redundant information of the whole rs-fMRI. Since there is self-similarity in rs-fMRI over different spatial and temporal scales, the fractal dimension is popular in features extraction. However, the method is usually employed to extract features from only one dimension, while the self-similar features of the other are ignored, leading to loss of hidden information about the structure of time-series. In this paper, we propose an automatic ADHD diagnosis method based on deep learning which extract spatial-temporal features, from rs-fMRI using both monofractal and multifractal dimensions. First of all, we separate the Default Mode Network (DMN), the Auditory Network (AN), and the anterior Salience Network (aSN) from the whole brain rs-fMRI. Then we employ monofractal analysis to extract the 3D spatial features from spatial maps. The extracted features are then fed into a 3D Convolution-based Spatial Feature Mixer (CSRM) to learn discriminative feature against ADHD. We also employ multifractal analysis to extract the temporal features from the time-series of each RSN and feed the features into an Attention-based Temporal Feature Mixer (ATFM). We fuse both outputs from CSRM and ATFM to predict the final result. We use five datasets from ADHD_200 to evaluate the performance of our proposed method, and obtain the state-of-the-art accuracy of 85.1. In addition, we find that the fractal spatial-temporal features helps to describe underlying stochastic properties of ADHD rs-fMRI.

PUBLICATION RECORD

  • Publication year

    2023

  • Venue

    International Conference on Biomedical Signal and Image Processing

  • Publication date

    2023-07-21

  • Fields of study

    Medicine, Computer Science

  • Identifiers
    DOI 10.1145/3613307.3613314
  • External record

    Open on Semantic Scholar

  • Source metadata

    Semantic Scholar

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-23 of 23 references · Page 1 of 1

CITED BY