Neurodiagnoses Logo Neurodiagnoses

Interactive Data Lab

NEW

Run custom Python analyses on Neurodiagnoses data

Secure, isolated execution environment with access to our Knowledge Base

FOR RESEARCH USE ONLY - NOT A MEDICAL DEVICE

This analysis platform is for research purposes only and not validated for clinical diagnostic use. Results are intended for scientific investigation and should not be used for medical decision-making.

How it works

1. Upload your CSV or Excel file with patient data below

2. Use AI Co-pilot to ask questions or write Python code manually

3. Code runs in a secure, isolated container with pandas, numpy, matplotlib, and seaborn pre-installed

📁 Data Source

Upload a CSV or Excel file to analyze

No data loaded

🧪 Reproducible Scientific Experiments

Test the Active Risk Hypothesis: Does combining genetic risk (PRS) with causal validation scores improve survival prediction beyond genetics alone?

This experiment trains two Cox proportional hazards models on a synthetic cohort and compares their predictive power.

🧠 Spatial Concordance Hypothesis (Axis 2→3)

Test if molecular markers (Axis 2) preferentially express in atrophied brain regions (Axis 3). This validates the mechanistic coherence of our tridimensional framework.

Uses spatial validation logic to compute concordance scores and t-test for statistical significance.

🧬 Replicación SOTA: Bachmann et al. (Preclinical Screening)

Epic 44

Hypothesis: Plasma biomarker fusion (Aβ42/40 + p-tau217) outperforms single biomarkers for predicting MCI conversion in cognitively healthy (CU) individuals.

Replicates Bachmann et al. finding using NACC CU cohort with longitudinal follow-up (CN→MCI). Compares 3 Logistic Regression models: Aβ only, Tau only, and Fusion (A+T).

🔬 Retrospective Pharmacology Lab (Axis 0c)

Epic 52

Full-Stack Industrialization: Run parameterized case-control studies on NACC data to evaluate the effect of any drug on cognitive progression (MMSE slope).

Uses Axis 0c (Pharmacology) architecture to identify patients on target medications, match controls by demographics/genetics, and compare progression rates. Evolved from Epic 50.

🔬 Retrospective Pharmacology Lab

Configure and run case-control studies using the Axis 0c (Pharmacology) architecture. Compare progression rates between patients on specific medications vs matched controls.

🗺️ Spatial Mechanism Validator (Axis 2 ↔ Axis 3)

Epic 53 v2

Topology & Hallmark Validation: Query the NeuroPro tissue atlas (23K protein-region associations) to validate mechanism hypotheses from literature.

Maps biomarker candidates (Eje 2) to brain expression topology (Eje 3a) and neuropathological hallmarks (NFT, Plaque, CAA). Re-uses Epic 22 (SpatialValidator) infrastructure.

🗺️ Spatial Mechanism Validator

Query the NeuroPro Tissue Atlas (23K protein-region associations) to validate mechanism hypotheses. Maps biomarker candidates to brain expression topology and neuropathological hallmarks (NFT, Amyloid, CAA).

🧠 Spatiotemporal Atrophy Propagation Modeling

Predict personalized trajectories of cortical atrophy propagation using graph-based dynamical modeling (Li et al., arXiv:2511.08847)

🎯 Causal Fusion Hypothesis (Axes 1+2+3a→3b) - REAL DATA

Test if combining genetics (Axis 1), biomarkers (Axis 2), and imaging (Axis 3a) predicts clinical dementia (Axis 3b) better than each axis alone. This validates the core premise of the tridimensional framework.

Uses real NACC data (204K patients). Trains 3 classification models and compares AUC scores.

📊 Validation Results: Genetics (Axis 1) → Molecular Biomarkers (Axis 2)

Pre-loaded analysis from our validation study (SEA-AD cohort, 84 patients)

Loading causal validation data...

🤖 Analysis Co-pilot

Ask questions in natural language

👋 Hello! I'm your Analysis Co-pilot.

You can ask me to:

  • 📊 Create plots: "Show a scatter plot of age vs MoCA score"
  • 📈 Calculate statistics: "What's the average MMSE by diagnosis?"
  • 🔍 Filter data: "Show patients older than 70 with MCI"
  • 🧬 Get diagnosis: "Diagnose patient P001 using Neurodiagnoses"

Note: For diagnosis tasks, I'll use the validated Neurodiagnoses engines.