ConductScreen

ConductScreen

Automated Morphological Phenotyping of Developmental Neurotoxicity in Human Neural Organoids

A combinatorial in vitro/in silico NAM for quantifying how prenatal substance exposure alters neurodevelopmental morphology. No animals required.

1,407

Images Analyzed

64

Organoids Tracked

4

Cell Lines Compared

23/27

Significant Differences

Four Analysis Modules

A multi-modal combinatorial NAM from single-image morphology to electrophysiology and longitudinal dynamics.

Morphology Module

Operational

2D Morphology from Brightfield Microscopy

Automated segmentation and feature extraction from standard brightfield images. No fluorescent labels required.

  • 16 features + 3 QC metrics per image
  • 9 morphology, 3 intensity, 4 texture features
  • Deterministic pipeline at 130 ms/image
  • Cross-lab reproducibility validated

Electrophysiology Module

Phase 2

MEA/Calcium Electrophysiology

Multi-electrode array and calcium imaging analysis to correlate morphological phenotypes with functional network activity changes from substance exposure.

  • Burst rate and spike frequency
  • Network synchrony metrics
  • Connectivity mapping
  • Functional-morphological correlation

Boundary Module

Phase 2

3D Confocal Neurite Topology

Skeletonization and graph-based analysis of neurite morphology from 3D confocal stacks.

  • Skeleton total length and branch length
  • Branch point and terminal tip counts
  • Topological complexity (Betti numbers)
  • Sholl analysis intersection profile

Temporal Module

Prototype

Time-Series Growth Dynamics

Longitudinal tracking of organoid growth, shape evolution, and phenotypic divergence over culture days.

  • Growth curves with exponential/logistic fits
  • Doubling time estimation
  • Trajectory clustering
  • Dose-response morphological curves

Key Finding

Our automated pipeline detects statistically significant morphological differences between wildtype and disease-model organoids across 23 of 27 feature-clone comparisons (BH-FDR q < 0.05, per-organoid aggregated N=16 per group), achieving an 85.2% discrimination rate with effect sizes |r| = 0.46–1.0.

Critically, the features that distinguish disease from wildtype — area, boundary integrity, tissue texture, and intensity heterogeneity — are the same endpoints reported as affected by prenatal substance exposure across 20+ published studies spanning 9 substance classes. This establishes both analytical validity and biological plausibility for the target context of use.

Beyond Software: A Complete Testing System

ConductScreen is the intelligence layer of a complete, integrated developmental neurotoxicity testing system.

Culture Chip

Hardware

Standard Protocol

Media kit

Any Microscope

Brightfield

ConductScreen

Analysis

Phase 3 delivers the full system. No organoid expertise needed.

Learn more about Culture Chip
Existing Manufacturing

Culture Chip builds on ConductScience's existing microfluidic chip manufacturing platform — organ-on-a-chip devices already in production for labs worldwide.

Regulatory Readiness

ConductScreen is designed for regulatory qualification under OECD Test Guideline 426. Explore our evidence:

Built on Real, Published Data

All analysis is performed on data from Zenodo 10301912 (Schröter et al., Scientific Data 2024) — 1,407 brightfield microscopy images from 64 organoids across 4 iPSC-derived cell lines (1 wildtype + 3 disease models).

No synthetic or simulated data. Every result shown on this site is derived from real experimental images.

NIH Complement-ARIE Reduction to Practice Challenge — Phase 1

A combinatorial in vitro/in silico NAM for automated morphological phenotyping of developmental neurotoxicity in human neural organoids