Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Early-stage detection of cognitive impairment by hybrid quantum-classical algorithm using resting-state functional MRI time-series

Junggu Choi, Tak Hur, Daniel K. Park, Na-Young Shin, Seung-Koo Lee, Hakbae Lee, Sanghoon Han

TL;DR

The study tackles early cognitive impairment detection by applying a hybrid quantum-classical algorithm to resting-state fMRI ROI time-series, combining a classical 1D CNN with a four-qubit QCNN that uses amplitude encoding and a SU(4) PQC for quantum convolution and pooling. Evaluated on ADNI rs-fMRI data across 116 ROIs, the hybrid model consistently outperformed a classical CNN baseline, with balanced accuracy rising from $0.523$ to $0.581$ as more QCNNs were added, under 5-fold cross-validation. Nine ROIs emerged as particularly informative for discrimination, and seed-based functional connectivity validated their relevance to cognitive decline, aligning with prior literature. The work demonstrates that QCNN-based quantum-classical hybrids can achieve better performance with fewer parameters on raw time-series data, informing future multi-modal studies and hardware-based validation for clinical neuroimaging.

Abstract

Following the recent development of quantum machine learning techniques, the literature has reported several quantum machine learning algorithms for disease detection. This study explores the application of a hybrid quantum-classical algorithm for classifying region-of-interest time-series data obtained from resting-state functional magnetic resonance imaging in patients with early-stage cognitive impairment based on the importance of cognitive decline for dementia or aging. Classical one-dimensional convolutional layers are used together with quantum convolutional neural networks in our hybrid algorithm. In the classical simulation, the proposed hybrid algorithms showed higher balanced accuracies than classical convolutional neural networks under the similar training conditions. Moreover, a total of nine brain regions (left precentral gyrus, right superior temporal gyrus, left rolandic operculum, right rolandic operculum, left parahippocampus, right hippocampus, left medial frontal gyrus, right cerebellum crus, and cerebellar vermis) among 116 brain regions were found to be relatively effective brain regions for the classification based on the model performances. The associations of the selected nine regions with cognitive decline, as found in previous studies, were additionally validated through seed-based functional connectivity analysis. We confirmed both the improvement of model performance with the quantum convolutional neural network and neuroscientific validities of brain regions from our hybrid quantum-classical model.

Early-stage detection of cognitive impairment by hybrid quantum-classical algorithm using resting-state functional MRI time-series

TL;DR

The study tackles early cognitive impairment detection by applying a hybrid quantum-classical algorithm to resting-state fMRI ROI time-series, combining a classical 1D CNN with a four-qubit QCNN that uses amplitude encoding and a SU(4) PQC for quantum convolution and pooling. Evaluated on ADNI rs-fMRI data across 116 ROIs, the hybrid model consistently outperformed a classical CNN baseline, with balanced accuracy rising from to as more QCNNs were added, under 5-fold cross-validation. Nine ROIs emerged as particularly informative for discrimination, and seed-based functional connectivity validated their relevance to cognitive decline, aligning with prior literature. The work demonstrates that QCNN-based quantum-classical hybrids can achieve better performance with fewer parameters on raw time-series data, informing future multi-modal studies and hardware-based validation for clinical neuroimaging.

Abstract

Following the recent development of quantum machine learning techniques, the literature has reported several quantum machine learning algorithms for disease detection. This study explores the application of a hybrid quantum-classical algorithm for classifying region-of-interest time-series data obtained from resting-state functional magnetic resonance imaging in patients with early-stage cognitive impairment based on the importance of cognitive decline for dementia or aging. Classical one-dimensional convolutional layers are used together with quantum convolutional neural networks in our hybrid algorithm. In the classical simulation, the proposed hybrid algorithms showed higher balanced accuracies than classical convolutional neural networks under the similar training conditions. Moreover, a total of nine brain regions (left precentral gyrus, right superior temporal gyrus, left rolandic operculum, right rolandic operculum, left parahippocampus, right hippocampus, left medial frontal gyrus, right cerebellum crus, and cerebellar vermis) among 116 brain regions were found to be relatively effective brain regions for the classification based on the model performances. The associations of the selected nine regions with cognitive decline, as found in previous studies, were additionally validated through seed-based functional connectivity analysis. We confirmed both the improvement of model performance with the quantum convolutional neural network and neuroscientific validities of brain regions from our hybrid quantum-classical model.
Paper Structure (19 sections, 2 equations, 10 figures, 12 tables)

This paper contains 19 sections, 2 equations, 10 figures, 12 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Data source
  4. MRI acquisition and preprocessing
  5. Hybrid quantum-classical classification model
  6. Model architecture
  7. Simulation of hybrid quantum-classical model
  8. Additional analysis : seed-based functional connectivitiy (SBFC)
  9. Tools
  10. Results
  11. Classification performance in the hybrid and baseline models
  12. Comparisons of model performances in 116 brain regions
  13. SBFC with selected ROIs
  14. Discussion
  15. Conclusion
  16. ...and 4 more sections

Figures (10)

  • Figure 1: The research scheme of this study.
  • Figure 2: The parameterized quantum circuit (PQC) used in the QCNN (the PQC a indicates the PQC for the quantum convolutional layer and the PQC b represents the PQC for the quantum pooling layer).
  • Figure 3: Hybrid model architecture example with a single QCNN
  • Figure 4: The classification performance and the number of parameter changes of the hybrid model with the number of the QCNN (blue line with circles: the averaged balanced accuracy / red line with diamonds: the number of parameters).
  • Figure 5: The different brain regions with statistically significant differences in the SBFC maps between healthy and early-MCI groups (seed : ROI 1, 84, 18, 17, 39, and 38).
  • ...and 5 more figures