Skip to content

Latest commit

 

History

History
327 lines (220 loc) · 8.81 KB

File metadata and controls

327 lines (220 loc) · 8.81 KB

🧠 Project Overview (Refined Plan)

1. Goal: Morphometric Analysis via Deformation Fields

We start with images already registered in MNI space.

  • Nonlinear registration produces a deformation (warp) field
  • For each voxel, this encodes:
    • local expansion (inflation)
    • local contraction (shrinkage)

👉 This deformation serves as a proxy for brain morphology

Clinical motivation:

  • Neurodegenerative diseases (e.g. Parkinson’s, Alzheimer’s)
  • Expect atrophy patterns (systematic shrinkage)
  • Deformation fields reveal where structural changes occur

2. From Voxels to Brain Regions

Problem:

  • Voxel-level data is too large (~2M voxels)
  • Morphological changes are not independent, but structured across regions

Solution:

  • Apply brain parcellation
  • Reduce dimensionality to region-level representation

Typical scale:

  • ~1k – 10k parcels (explored experimentally)

👉 Each parcel aggregates multiple voxels → first compression step


3. Brain Partitioning Strategy

  1. Split into macro compartments:

    • Left cortex
    • Right cortex
    • Subcortical regions
  2. Perform micro-parcellation within each compartment

  3. Recombine into full-brain parcel set

⚠️ Important:

  • Test multiple resolutions (low / medium / high)
  • Avoid jumping directly to maximum resolution

4. Parcel-Level Feature Construction (Node Features)

Instead of reducing each parcel to a single summary statistic such as the mean, we represent each parcel by the distribution of log-Jacobian determinant values measured across all voxels in that parcel.

Node definition:

For each parcel:

  1. Extract all voxel-wise log-Jacobian determinant values from the nonlinear warp
  2. Sort the values and interpolate them onto a common 15-point quantile grid
  3. Use that quantile representation as the parcel distribution P(x)

👉 Result:

  • Each parcel is described by its full morphometric profile, not just by one number

🎯 Purpose:

  • Preserve the shape of the deformation pattern
  • Distinguish uniform abnormalities from heterogeneous or patchy abnormalities
  • Retain richer information for downstream network construction

5. Edge Definitions (3 Types)

Goal:

Build subject-level parcel graphs using multiple definitions of pairwise morphometric similarity so that downstream validation does not depend on a single edge metric.

For each pair of parcels A and B, compute edges using:

  1. Wasserstein distance
  2. KL divergence (Gaussian version)
  3. Median + IQR

Edge construction details

For each pair of parcels A and B, compute the symmetric Wasserstein-1 distance between their common quantile representations. The resulting graph is undirected: $w_{AB} = w_{BA}$.

Similarity convention

  • For Wasserstein distance, convert distance into a similarity score before adding the edge to the graph. The current transforms are $\exp(-W)$ and $1/(1+W)$.
  • For KL divergence (Gaussian version), use the divergence-derived pairwise comparison defined on Gaussian parcel summaries.
  • For Median + IQR, use robust summary-statistic comparisons rather than full-density matching.

👉 Result:

  • Nodes = parcels
  • Edge weights = pairwise parcel relationships under three alternative morphometric definitions

6. Graph Construction

For each subject and for each edge definition:

  • build a graph using the corresponding parcel-to-parcel edge weights
  • start with a fully dense graph
  • later, evaluate 2-3 different thresholding strategies

Recommended workflow:

  1. construct the full dense weighted graph
  2. compute graph statistics on the dense version first
  3. then test a small set of thresholded variants for sensitivity analysis

This makes it possible to compare how stable downstream results are with respect to sparsification.


7. Graph Metric

For each graph, compute the weighted degree of every parcel:

$$ \mathrm{WeightedDegree}_i = \frac{\sum_{j \ne i} w_{ij}}{N-1} $$

where:

  • $w_{ij}$ is the edge weight between parcel $i$ and parcel $j$
  • $N$ is the number of nodes in the graph
  • the diagonal self-similarity is excluded

👉 Primary graph-level readout:

  • weighted degree = mean similarity to all other nodes

Interpretation of weighted degree

  • High weighted degree means a parcel is morphometrically similar to many other parcels.
  • Low weighted degree means a parcel is unusual relative to the subject's overall parcel population. It is an anomaly score, not a direct or specific measurement of atrophy.
  • Diseased parcels are expected to be more atrophied and to show deformation patterns that deviate from the rest of the brain.
  • Therefore, diseased parcels are hypothesized to show lower weighted degree than healthy parcels.
  • Weighted degree depends on graph composition. Large compartments such as cortex and CSF contribute many comparison nodes, so anatomical compartment and parcel counts must be considered during group-level interpretation.

8. Datasets for Validation

Validation will be carried out on the following datasets:

  • OASIS
  • Epilepsy dataset
  • PPMI
  • Psychiatric datasets

These datasets provide the test bed for comparing graph-derived parcel metrics between healthy and diseased groups.


9. Validation Procedure

For each subject-level graph:

  1. compute the weighted degree for every parcel
  2. split subjects into:
    • Healthy
    • Diseased (137 subjects)
  3. for each parcel:
    • compare weighted degree between healthy and diseased groups
    • run a t-test
    • compute effect size

Repeat this analysis for:

  • all parcels (approximately 90k)
  • all three edge metrics

👉 Core validation question:

  • which parcels show the strongest group differences in weighted degree, and how consistent are those differences across edge definitions?

🧪 Practical Study Design

Primary analysis

  • Build one graph per subject for each of the three edge definitions
  • Compute parcel-wise weighted degree
  • Perform parcel-wise healthy vs diseased comparisons
  • Record both significance and effect size

Sensitivity analysis

  • Start from fully dense graphs
  • Later test 2-3 threshold choices
  • Re-run the weighted-degree validation after thresholding

Scale

  • Perform the analysis across the full parcellation
  • Target scale: approximately 90k parcels

🕰️ Legacy Methods

The following methods remain part of the project as legacy approaches from earlier exploration, but they are no longer the primary validation pathway:

A. Leiden-Based Graph Clustering

  • Sparse weighted graph construction
  • Leiden community detection

Historical purpose:

  • identify communities of parcels with coordinated deformation patterns

Current status:

  • retained for comparison and historical reference
  • not the main validation endpoint

B. Mini-Batch Clustering

  • Mini-batch K-means on parcel-level features
  • optional dimensionality reduction before clustering

Historical purpose:

  • fast scalable clustering baseline

Current status:

  • retained as a legacy baseline
  • not the main analysis for the current project description

🚫 What to Avoid

  • Treating one edge definition as definitive before validation
  • Discarding the dense-graph analysis too early
  • Relying only on clustering outputs when the current endpoint is parcel-wise statistical validation
  • Gradient-based / deep learning approaches (for now)

👉 The current emphasis is on validating parcel-wise graph statistics across multiple edge definitions and datasets.


📊 Evaluation Metrics

Statistical:

  • t-statistics
  • effect sizes
  • parcel-wise healthy vs diseased differences

Graph-derived:

  • weighted degree per parcel
  • sensitivity to threshold choice
  • consistency across the three edge definitions

Practical:

  • runtime
  • peak RAM
  • scalability to approximately 90k parcels

🎯 Final Output Goal

  • Build parcel graphs under three alternative edge definitions
  • Compute parcel-wise weighted degree for each subject
  • Identify parcels whose weighted degree differs between healthy and diseased groups
  • Quantify how robust those findings are across datasets and threshold choices

🧭 Final Strategy Summary

Core idea:

Represent each parcel through morphometric summaries derived from the warp, construct dense subject-level graphs using three alternative edge definitions, compute parcel-wise weighted degree, and validate group differences between healthy and diseased subjects.


Active analysis path:

  1. Wasserstein-based graph
  2. Gaussian KL-based graph
  3. Median + IQR-based graph
  4. Weighted-degree validation across parcels and datasets

Legacy methods kept in scope:

  1. Leiden community detection → legacy graph method
  2. Mini-batch K-means → legacy clustering baseline