Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Nonlinearity-aware based dimensionality reduction and over-sampling for AD/MCI classification from MRI measures.

Peng Cao1, Xiaoli Liu2, Jinzhu Yang2

  • 1Computer Science and Engineering, Northeastern University, Shenyang, China.

Computers in Biology and Medicine
|October 17, 2017
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Spatiotemporal rendering for dynamic sensation reproduction via electrostatic-enhanced vibro-haptic interface.

Innovation (Cambridge (Mass.))·2026
Same author

Asymmetries-induced nonlinear dynamic behaviors enable versatile modulation strategy for insect-scale robotics.

Science advances·2025
Same author

Polymer-constrained excimer enables flexible and self-healable optoelectronic elastomer for mechanical sensor.

Nature communications·2025
Same author

Self-Elevated 3D Helical Oscillator with Addressable Eigenfrequency for Wearable Interface.

Advanced materials (Deerfield Beach, Fla.)·2025
Same author

Robust Textile-Based Electromagnetic Audio Interfaces.

Advanced materials (Deerfield Beach, Fla.)·2025
Same author

Concentric ice-templating of ultracompressible tough hydrogels with bioinspired circumferentially aligned architecture.

Science advances·2025
This summary is machine-generated.

Accurate Alzheimer's disease (AD) diagnosis using brain MRI is crucial. New multi-kernel methods improve diagnostic accuracy by handling complex data relationships and class imbalance, identifying key imaging biomarkers.

Area of Science:

  • Neuroimaging
  • Machine Learning
  • Medical Diagnostics

Background:

  • Alzheimer's disease (AD) poses significant financial and emotional burdens.
  • Early and accurate diagnosis of AD using brain magnetic resonance imaging (MRI) is critical.
  • High dimensionality and imbalanced data are major challenges in computer-aided AD diagnosis.

Purpose of the Study:

  • To develop novel dimensionality reduction and over-sampling methods for AD diagnosis.
  • To address limitations of existing linear methods by capturing complex relationships between MRI features and disease status.
  • To improve the accuracy and reliability of computer-aided AD diagnosis.

Main Methods:

  • Proposed a multi-kernel approach combining Marginal Fisher Analysis with ℓ2,1-norm multi-kernel learning (MKMFA) for sparse region-of-interest selection and dimensionality transformation.
Keywords:
Alzheimer's diseaseFeature selectionManifold learningMulti-kernel learningOver-sampling

Related Experiment Videos

  • Developed a multi-kernel over-sampling (MKOS) method to generate synthetic data in the optimal kernel space, addressing class imbalance.
  • Evaluated models on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset for binary and multi-class classification.

    • Main Results:

    • The proposed multi-kernel methods demonstrated superior performance compared to existing approaches.
    • The methods effectively handled high dimensionality and imbalanced data in AD diagnosis.
    • Identified relevant neuroimaging biomarkers consistent with existing medical knowledge.

    Conclusions:

    • The novel multi-kernel approach significantly enhances the accuracy of computer-aided Alzheimer's disease diagnosis.
    • The methods provide a robust solution for challenges posed by complex MRI data and class imbalance.
    • The identified biomarkers can aid in early AD detection and understanding.