NeuroInsight AI

Agentic clinical intelligence for Alzheimer's MRI analysis — a research-driven Flutter platform exploring 3D neuroimaging, explainability and case-aware RAG.

Flutter Dart Agentic AI Case-Aware RAG 3D MRI Explainable AI
View on GitHub
NeuroInsight AI dashboard NeuroInsight AI multi-planar MRI viewer NeuroInsight AI 3D brain visualization NeuroInsight AI explainability heatmap NeuroInsight AI case-aware assistant NeuroInsight AI structured clinical report

Overview

Traditional AI medical-imaging demos stop at Image → Model → Prediction. NeuroInsight AI explores a much broader clinical workflow: image-quality assessment, brain-structure segmentation, quantitative biomarkers, explainability, evidence retrieval, case-aware reasoning and structured reporting — all around a single MRI study.

The platform is built as a research and prototype product experience, designed to be visual, explainable, evidence-aware and human-centered, with every AI conclusion intended to remain under expert clinical review rather than replace it.

Purpose — Why NeuroInsight AI Exists

Reading an Alzheimer's MRI is slow, highly specialist-dependent, and a plain "98% probability" score gives a clinician nothing to reason with. NeuroInsight AI exists to make that judgment faster and more transparent — pairing a prediction with the anatomical regions, measurements and research evidence that support it, so a radiologist can verify the AI's reasoning instead of just trusting a number.

Animated Agentic Workflow

1
MRI Study
A patient's MRI study enters the pipeline.
2
Imaging Agent
Study inspection and image-quality assessment.
3
Segmentation Agent
Brain-structure segmentation across planes.
4
Measurement Agent
Volumetric and morphometric biomarkers.
5
Clinical Findings Agent
Synthesizes patterns and abnormalities.
6
RAG & Reasoning Agents
Retrieves evidence and reasons over the case.
7
Report Agent
Structured report, handed to human review.

Key Capabilities

Multi-Planar MRI Viewer — axial, coronal, sagittal and 3D inspection views.
3D Brain Visualization — interactive rendering of affected anatomical regions.
Quantitative Biomarkers — hippocampal, ventricular and cortical volume measurements.
Explainable AI — attention/heatmap overlays showing which regions drove a finding.
Case-Aware RAG — connects the active case with research and clinical evidence.
Structured Reporting — findings, measurements, explanations and evidence in one review-ready report.

Benefits

Faster Triage

Agents pre-process, measure and summarize a study before a radiologist opens it.

Human-in-the-Loop

Every finding stays explainable and reviewable — never an unchecked black box.

Evidence, Not Guesses

Case-aware RAG surfaces the research behind a finding automatically.

Consistent Reporting

Every case follows the same structured report format, easing review across patients.

Technology Stack

Flutter Dart Web Windows Agentic AI Orchestration Retrieval-Augmented Generation 3D Visualization
Explore the Repository