SYNTHETIC DEMONSTRATION DATA · NO REAL SUBJECT DATA · RESEARCH PROTOTYPE

Connectomic Anomaly Explorer

Matrix → metrics → visualization → export. Upper triangle is used; diagonal excluded; zeros are no anomaly; NaNs excluded.
Subject-level overview

Integrated clinical-connectomic timeline

Clinical events and anomaly trajectories on one calendar axis
Events are separated into colored swimlanes to prevent overlap. Scan dates extend vertically through the chart, aligning the clinical course with Global Anomaly Score, hyperconnective burden, hypoconnective burden, and D-CAS. Use the legend above to show or hide layers and switch the domain display mode.
Interval summaries
Calculated between adjacent scans: G-CAS change, hyper/hypo change, domain shifts, and associated clinical events.
Connectomic information layer
AI connectomic assistant
Explains the subject's longitudinal connectomic profile and relevant functional-domain context while avoiding unsupported causal or prognostic claims.
Select a subject and generate an explanatory longitudinal report.
Generate a report to see an at-a-glance explanation of the current connectomic phenotype and longitudinal course.
Generate a report to see interval-by-interval interpretation.
Generate a report to see the most informative functional domains.
Ask about the quantitative trajectory, a specific interval, or what an affected functional domain broadly represents. Each answer is based on the selected subject's longitudinal connectomic measurements.
The assistant provides explanatory context and is not a diagnostic or prognostic system.
Weighted neighborhood network
Signed weighted adjacency matrix
Selected parcel metrics
Current neighborhood metrics
Global metrics and validation
Download current case outputs
Top 10 parcel burden plot
Domain D-CAS bar plot
Domain metrics table
DomainD-CASHyperHypoCBI
Select domains
This tab uses D-CAS calculated at the domain level. Selecting domains lets you inspect their summary metrics and seed the strongest parcels in those domains.
Selected domain summary
Selected-domain D-CAS bar plot
Top parcels in selected domains
RankParcelClinicalDomainStrengthHyperHypo
Domain path analysis
Path cost is 1 / positive edge weight, so stronger hyper edges are easier/shorter.
The direct lost edge is negative. The drawn path uses positive edges only.
Path result
Upload raw signed anomaly matrix CSV

CSV format: first column = row parcel labels; remaining headers = matching parcel labels; values are signed anomaly scores. Positive = hyperconnective; negative = hypoconnective; zero = no anomaly; blank/NaN = excluded.

Definitions

G-CAS

Global Anomaly Score: total weighted anomaly burden across all abnormal upper-triangle edges, equal to hyper burden plus hypo burden.

D-CAS

Domain-Specific Anomaly Score. An abnormal edge contributes to a domain when at least one endpoint parcel belongs to that domain.

Degree vs strength burden

Degree counts abnormal edges. Strength sums absolute anomaly magnitudes.

CBI

Connectivity Balance Index = Hyper Burden / Hypo Burden.

Path analysis

Starts from a high-burden domain parcel, finds its strongest lost edge, then searches for a positive-edge-only route to that same target.

Neighborhood explorer

Seed parcels plus their strongest hyper/hypo neighbors. The matrix shows all internal signed values among visible nodes.

Curved edges and white halos reduce apparent red/blue overlap. Purple/gold overlay marks hyper-only path results.