Wiki source code of Neurodiagnoses

Last modified by manuelmenendez on 2025/03/03 22:46
Show last authors
1 (% class="jumbotron" %)
2 (((
3 (% class="container" %)
4 (((
5 = //A new tridimensional diagnostic framework for complex CNS diseases// =
6
7 This project is focused on developing a novel nosological and diagnostic framework for complex CNS diseases by using advanced AI techniques and integrating data from neuroimaging, biomarkers, and biomedical ontologies.
8 We aim to create a structured, interpretable, and scalable diagnostic tool.
9 )))
10 )))
11
12 (% class="row" %)
13 (((
14 (% class="col-xs-12 col-sm-8" %)
15 (((
16 = What is this about and what can I find here? =
17
18 = **Overview** =
19
20 The classification and diagnosis of central nervous system (CNS) diseases have long been constrained by traditional, phenotype-based approaches that often fail to capture the complex interplay of pathophysiological mechanisms, molecular biomarkers, and neuroanatomical changes.
21
22 **Neurodiagnoses** redefines this landscape by integrating advanced AI with multi-modal data—including genetics, neuroimaging, biomarkers, and digital health records—to create a more precise, scalable, and data-driven diagnostic system.
23
24 Additionally, **Neurodiagnoses is now expanding into disease prediction and biomarker estimation**, integrating state-of-the-art machine learning models to enhance precision diagnostics and disease progression forecasting.
25
26 On this page, you will find:
27
28 * Detailed descriptions of both the clinical diagnostic tools and the research framework.
29 * Access to our AI models, data processing pipelines, and digital twin simulations.
30 * Collaborative resources for researchers, clinicians, and AI developers.
31 * Guidelines and instructions on how to contribute to and expand the project.
32
33 == **The role of AI-powered annotation** ==
34
35 To enhance standardization, interpretability, and clinical application, the framework integrates an AI-powered annotation system, which:
36
37 * Assigns structured metadata tags to diagnostic features.
38 * Provides real-time contextual explanations for AI-based classifications.
39 * Tracks longitudinal disease progression using timestamped AI annotations.
40 * Improves AI model transparency through interpretability tools (e.g., SHAP analysis).
41 * Facilitates decision-making for clinicians by linking annotations to standardized biomedical ontologies (SNOMED, HPO).
42
43 Neurodiagnoses provides **two complementary AI-driven diagnostic approaches**:
44
45 1. **Probabilistic Diagnosis**
46 * AI assigns probability scores to multiple possible diagnoses based on biomarker, imaging, and clinical data.
47 * Useful for differential diagnosis and treatment decision-making.
48
49 2. **Tridimensional Diagnosis**
50 * Diagnoses are structured based on:
51 - **(1) Etiology** (genetic, autoimmune, metabolic, infectious).
52 - **(2) Molecular Biomarkers** (amyloid-beta, tau, inflammatory markers, EEG patterns).
53 - **(3) Neuroanatomoclinical Correlations** (brain atrophy, connectivity alterations).
54 * This approach enables precise disease subtyping and biologically meaningful classification, particularly useful for tracking progression over time.
55
56 Both systems will be offered for every patient case, allowing clinicians to compare AI-generated probabilistic diagnosis with a structured tridimensional classification.
57
58 == **Disease Prediction and Biomarker Estimation** ==
59
60 Neurodiagnoses is also implementing **biomarker prediction and disease progression modeling**, using advanced machine learning techniques:
61
62 * **Biomarker Prediction:**
63 - Estimation of fluid-based and neuroimaging biomarkers without invasive testing.
64 - Multi-modal machine learning models for predicting molecular and clinical markers.
65
66 * **Disease Progression Modeling:**
67 - AI-driven forecasts for neurodegenerative disease evolution.
68 - Probabilistic disease conversion models (e.g., MCI to AD, Parkinson's prodromal phases).
69 - Survival models and risk stratification for precision medicine applications.
70
71 == **The case of neurodegenerative diseases** ==
72
73 There have been described these 3 diagnostic axes:
74
75 [[Neurodegenerative diseases can be studied and classified in a tridimensional scheme with three axes: anatomic–clinical, molecular, and etiologic. CSF, cerebrospinal fluid; FDG, fluorodeoxyglucose; MRI, magnetic resonance imaging; PET, positron emission tomography.>>image:tridimensional.png||alt="tridimensional view of neurodegenerative diseases"]]
76
77 * (((
78 **Axis 1: Etiology**
79 * //Description//: Focuses on genetic and sporadic causes, identifying risk factors and potential triggers.
80 * //Examples//: APOE ε4 as a genetic risk factor, or cardiovascular health affecting NDD progression.
81 * //Tests//: Genetic testing, lifestyle, and cardiovascular screening.
82 )))
83 * (((
84 **Axis 2: Molecular Markers**
85 * //Description//: Analyzes primary (amyloid-beta, tau) and secondary biomarkers (NFL, GFAP) for tracking disease progression.
86 * //Examples//: CSF amyloid-beta concentrations to confirm Alzheimer’s pathology.
87 * //Tests//: Blood/CSF biomarkers, PET imaging (Tau-PET, Amyloid-PET).
88 )))
89 * (((
90 **Axis 3: Neuroanatomoclinical**
91 * //Description//: Links clinical symptoms to neuroanatomical changes, such as atrophy or functional impairments.
92 * //Examples//: Hippocampal atrophy correlating with memory deficits.
93 * //Tests//: MRI volumetrics, FDG-PET, neuropsychological evaluations.
94 )))
95
96 == **Applications** ==
97
98 This system enhances:
99
100 * **Research**: By stratifying patients, reducing cohort heterogeneity in clinical trials.
101 * **Clinical Practice**: Provides dynamic diagnostic annotations with timestamps for longitudinal tracking.
102
103 == How to Contribute ==
104
105 * Access the `/docs` folder for guidelines.
106 * Use `/code` for the latest AI pipelines.
107 * Share feedback and ideas in the wiki discussion pages.
108 * Join our [[Community on EBRAINS>>https://community.ebrains.eu/_ideas/-OJHTZrpKrrrkx-u0djj/about]]
109 * Join the [[Discussion Forum at GitHub>>https://github.com/Fundacion-de-Neurociencias/neurodiagnoses/discussions]]
110
111 == Key Objectives ==
112
113 * Develop interpretable AI models for diagnosis and progression tracking.
114 * Integrate data from Human Phenotype Ontology (HPO), Gene Ontology (GO), and other biomedical resources.
115 * Foster collaboration among neuroscientists, AI researchers, and clinicians.
116 * Provide a dual diagnostic system:
117 ** Probabilistic Diagnosis – AI assigns multiple traditional possible diagnoses with probability percentages.
118 ** Tridimensional Diagnosis – AI structures diagnoses based on etiology, biomarkers, and neuroanatomical correlations.
119 * Implement disease prediction models for neurodegenerative conditions.
120 * Predict biomarkers from non-invasive data sources.
121
122 == Who has access? ==
123
124 We welcome contributions from the global community. Join us as we transform CNS diagnostics and drive precision medicine forward through a collaborative, open-source approach. Let’s build the future of neurological diagnostics together!
125 )))
126
127 (% class="col-xs-12 col-sm-4" %)
128 (((
129 {{box title="**Contents**"}}
130 {{toc/}}
131 {{/box}}
132
133 == Main contents ==
134
135 * `/docs`: Documentation and contribution guidelines.
136 * `/code`: Machine learning pipelines and scripts.
137 * `/data`: Sample datasets for testing.
138 * `/outputs`: Generated models, visualizations, and reports.
139 * [[Methodology>>url:https://wiki.ebrains.eu/bin/view/Collabs/neurodiagnoses/Methodology/]]
140 * [[Notebooks>>Notebooks]]
141 * [[Results>>url:https://wiki.ebrains.eu/bin/view/Collabs/neurodiagnoses/Results/]]
142 * [[to-do-list>>to-do-list]]
143 )))
144 )))