Predicting Brain Connectivity from Structure
Author Information
Author(s): Andrew Zalesky, Tabinda Sarwar, Ye Tian, Yuanzhe Liu, B. T. Thomas Yeo, Kotagiri Ramamohanarao
Primary Institution: The University of Melbourne
Hypothesis
Can individual functional connectivity be predicted from structural connectome data?
Conclusion
The study shows that individual predictions of functional connectivity can significantly outperform group-average benchmarks, although the effect sizes are small.
Supporting Evidence
- Predicted individual effects showed a 9.1% improvement in similarity compared to chance levels.
- Statistical tests indicated that predictions significantly outperformed group-average benchmarks.
- Individual predictions were validated using a dataset from the Human Connectome Project.
Takeaway
This study found that we can guess how different parts of the brain connect based on its structure, but the guesses aren't always very accurate.
Methodology
The study used deep learning models to predict functional connectivity from structural connectome data, validated with statistical tests on a dataset of 1,000 healthy adults.
Potential Biases
Potential biases in connectome mapping and the homogeneity of the sample may limit the generalizability of the findings.
Limitations
The effect sizes of the predictions are small and may be overshadowed by group effects.
Participant Demographics
1,000 healthy young adults aged 22-35 years.
Statistical Information
P-Value
0.03
Confidence Interval
95% confidence interval
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website