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

How to avoid spurious cluster validation? A methodological investigation on simulated and fMRI data.

Ulrich Möller1, Marc Ligges, Petra Georgiewa

  • 1Clinic for Child and Adolescents Psychiatry, Friedrich Schiller University Jena, Germany.

Neuroimage
|December 17, 2002
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

[Beyond NICE: Updated Systematic Review on the Current Evidence of Using Puberty Blocking Pharmacological Agents and Cross-Sex-Hormones in Minors with Gender Dysphoria].

Zeitschrift fur Kinder- und Jugendpsychiatrie und Psychotherapie·2024
Same author

Cross-Sectional Investigation of Brain Volume in Dyslexia.

Frontiers in neurology·2022
Same author

Documentation and Communication of Psychooncological Findings in an Interdisciplinary Breast Cancer Center.

Breast care (Basel, Switzerland)·2018
Same author

Influence of imputation strategies on the identification of brain functional connectivity networks.

Journal of neuroscience methods·2018
Same author

Evaluation of a Novel Parent-Rated Scale for Selective Mutism.

Assessment·2018
Same author

Tracking the Reorganization of Module Structure in Time-Varying Weighted Brain Functional Connectivity Networks.

International journal of neural systems·2018

Optimizing clustering and cluster validation is crucial for accurate exploratory fMRI data analysis. Insufficient optimization can lead to misleading results, while sufficient optimization ensures reliable characterization of brain activity patterns.

Area of Science:

  • Neuroimaging
  • Data Science
  • Computational Neuroscience

Background:

  • Exploratory functional Magnetic Resonance Imaging (fMRI) data analysis often employs clustering and cluster validation techniques.
  • Previous research focused on identifying optimal validity functions for characterizing data structure.
  • The current study evaluates the impact of optimization within the clustering process itself.

Purpose of the Study:

  • To evaluate a common two-stage approach for exploratory fMRI data analysis involving clustering and cluster validation.
  • To investigate the influence of optimizing the sequence of partitions against the objective function.
  • To determine the conditions under which clustering and validation yield reliable fMRI data interpretations.

Main Methods:

  • The study optimized sequences of data partitions (clustering) based on objective functions.

Related Experiment Videos

  • Evaluated three clustering algorithms (hard and fuzzy) and three cluster validity functions.
  • Tested the approach on Gaussian clusters, simulated fMRI data, and real fMRI data.

    • Main Results:

    • Insufficient optimization of partitions can lead to spurious cluster validation and misinterpretation of fMRI data.
    • Sufficient optimization for each cluster number provides a reliable data characterization and enables valid function evaluation.
    • Findings were consistent across different clustering algorithms, validity functions, and data types.

    Conclusions:

    • Proper optimization of the clustering sequence is essential for accurate exploratory fMRI analysis.
    • Improved clustering tools should incorporate robust optimization strategies for reliable data interpretation.
    • This work provides a basis for enhancing clustering methodologies in neuroimaging research.