Wiki source code of Methodology

Version 12.2 by manuelmenendez on 2025/02/09 09:54
Show last authors
1 ==== **Overview** ====
2
3 This project develops a **tridimensional diagnostic framework** for **CNS diseases**, incorporating **AI-powered annotation tools** to improve **interpretability, standardization, and clinical utility**. The methodology integrates **multi-modal data**, including **genetic, neuroimaging, neurophysiological, and biomarker datasets**, and applies **machine learning models** to generate **structured, explainable diagnostic outputs**.
4
5 === **Workflow** ===
6
7 1. (((
8 **We Use GitHub to [[Store and develop AI models, scripts, and annotation pipelines.>>https://github.com/Fundacion-de-Neurociencias/neurodiagnoses/discussions]]**
9
10 * Create a **GitHub repository** for AI scripts and models.
11 * Use **GitHub Projects** to manage research milestones.
12 )))
13 1. (((
14 **We Use EBRAINS for Data & Collaboration**
15
16 * Store **biomarker and neuroimaging data** in **EBRAINS Buckets**.
17 * Run **Jupyter Notebooks** in **EBRAINS Lab** to test AI models.
18 * Use **EBRAINS Wiki** for structured documentation and research discussion.
19 )))
20
21 ----
22
23 === **1. Data Integration** ===
24
25 == Overview ==
26
27
28 Neurodiagnoses integrates clinical data via the **EBRAINS Medical Informatics Platform (MIP)**. MIP federates decentralized clinical data, allowing Neurodiagnoses to securely access and process sensitive information for AI-based diagnostics.
29
30 == How It Works ==
31
32
33 1. (((
34 **Authentication & API Access:**
35
36 * Users must have an **EBRAINS account**.
37 * Neurodiagnoses uses **secure API endpoints** to fetch clinical data (e.g., from the **Federation for Dementia**).
38 )))
39 1. (((
40 **Data Mapping & Harmonization:**
41
42 * Retrieved data is **normalized** and converted to standard formats (.csv, .json).
43 * Data from **multiple sources** is harmonized to ensure consistency for AI processing.
44 )))
45 1. (((
46 **Security & Compliance:**
47
48 * All data access is **logged and monitored**.
49 * Data remains on **MIP servers** using **federated learning techniques** when possible.
50 * Access is granted only after signing a **Data Usage Agreement (DUA)**.
51 )))
52
53 == Implementation Steps ==
54
55
56 1. Clone the repository.
57 1. Configure your **EBRAINS API credentials** in mip_integration.py.
58 1. Run the script to **download and harmonize clinical data**.
59 1. Process the data for **AI model training**.
60
61 For more detailed instructions, please refer to the **[[MIP Documentation>>url:https://mip.ebrains.eu/]]**.
62
63 ----
64
65 = Data Processing & Integration with Clinica.Run =
66
67
68 == Overview ==
69
70
71 Neurodiagnoses now supports **Clinica.Run**, an open-source neuroimaging platform designed for **multimodal data processing and reproducible neuroscience workflows**.
72
73 == How It Works ==
74
75
76 1. (((
77 **Neuroimaging Preprocessing:**
78
79 * MRI, PET, EEG data is preprocessed using **Clinica.Run pipelines**.
80 * Supports **longitudinal and cross-sectional analyses**.
81 )))
82 1. (((
83 **Automated Biomarker Extraction:**
84
85 * Standardized extraction of **volumetric, metabolic, and functional biomarkers**.
86 * Integration with machine learning models in Neurodiagnoses.
87 )))
88 1. (((
89 **Data Security & Compliance:**
90
91 * Clinica.Run operates in **compliance with GDPR and HIPAA**.
92 * Neuroimaging data remains **within the original storage environment**.
93 )))
94
95 == Implementation Steps ==
96
97
98 1. Install **Clinica.Run** dependencies.
99 1. Configure your **Clinica.Run pipeline** in clinica_run_config.json.
100 1. Run the pipeline for **preprocessing and biomarker extraction**.
101 1. Use processed neuroimaging data for **AI-driven diagnostics** in Neurodiagnoses.
102
103 For further information, refer to **[[Clinica.Run Documentation>>url:https://clinica.run/]]**.
104
105 ==== ====
106
107 ==== **Data Sources** ====
108
109 [[List of potential sources of databases>>https://github.com/Fundacion-de-Neurociencias/neurodiagnoses/blob/main/data/sources/list_of_potential_databases]]
110
111 **Biomedical Ontologies & Databases:**
112
113 * **Human Phenotype Ontology (HPO)** for symptom annotation.
114 * **Gene Ontology (GO)** for molecular and cellular processes.
115
116 **Dimensionality Reduction and Interpretability:**
117
118 * **Evaluate interpretability** using metrics like the **Area Under the Interpretability Curve (AUIC)**.
119 * **Leverage [[DEIBO>>https://github.com/Mellandd/DEIBO]] (Data-driven Embedding Interpretation Based on Ontologies)** to connect model dimensions to ontology concepts.
120
121 **Neuroimaging & EEG/MEG Data:**
122
123 * **MRI volumetric measures** for brain atrophy tracking.
124 * **EEG functional connectivity patterns** (AI-Mind).
125
126 **Clinical & Biomarker Data:**
127
128 * **CSF biomarkers** (Amyloid-beta, Tau, Neurofilament Light).
129 * **Sleep monitoring and actigraphy data** (ADIS).
130
131 **Federated Learning Integration:**
132
133 * **Secure multi-center data harmonization** (PROMINENT).
134
135 ----
136
137 ==== **Annotation System for Multi-Modal Data** ====
138
139 To ensure **structured integration of diverse datasets**, **Neurodiagnoses** will implement an **AI-driven annotation system**, which will:
140
141 * **Assign standardized metadata tags** to diagnostic features.
142 * **Provide contextual explanations** for AI-based classifications.
143 * **Track temporal disease progression annotations** to identify long-term trends.
144
145 ----
146
147 === **2. AI-Based Analysis** ===
148
149 ==== **Machine Learning & Deep Learning Models** ====
150
151 **Risk Prediction Models:**
152
153 * **LETHE’s cognitive risk prediction model** integrated into the annotation framework.
154
155 **Biomarker Classification & Probabilistic Imputation:**
156
157 * **KNN Imputer** and **Bayesian models** used for handling **missing biomarker data**.
158
159 **Neuroimaging Feature Extraction:**
160
161 * **MRI & EEG data** annotated with **neuroanatomical feature labels**.
162
163 ==== **AI-Powered Annotation System** ====
164
165 * Uses **SHAP-based interpretability tools** to explain model decisions.
166 * Generates **automated clinical annotations** in structured reports.
167 * Links findings to **standardized medical ontologies** (e.g., **SNOMED, HPO**).
168
169 ----
170
171 === **3. Diagnostic Framework & Clinical Decision Support** ===
172
173 ==== **Tridimensional Diagnostic Axes** ====
174
175 **Axis 1: Etiology (Pathogenic Mechanisms)**
176
177 * Classification based on **genetic markers, cellular pathways, and environmental risk factors**.
178 * **AI-assisted annotation** provides **causal interpretations** for clinical use.
179
180 **Axis 2: Molecular Markers & Biomarkers**
181
182 * **Integration of CSF, blood, and neuroimaging biomarkers**.
183 * **Structured annotation** highlights **biological pathways linked to diagnosis**.
184
185 **Axis 3: Neuroanatomoclinical Correlations**
186
187 * **MRI and EEG data** provide anatomical and functional insights.
188 * **AI-generated progression maps** annotate **brain structure-function relationships**.
189
190 ----
191
192 === **4. Computational Workflow & Annotation Pipelines** ===
193
194 ==== **Data Processing Steps** ====
195
196 **Data Ingestion:**
197
198 * **Harmonized datasets** stored in **EBRAINS Bucket**.
199 * **Preprocessing pipelines** clean and standardize data.
200
201 **Feature Engineering:**
202
203 * **AI models** extract **clinically relevant patterns** from **EEG, MRI, and biomarkers**.
204
205 **AI-Generated Annotations:**
206
207 * **Automated tagging** of diagnostic features in **structured reports**.
208 * **Explainability modules (SHAP, LIME)** ensure transparency in predictions.
209
210 **Clinical Decision Support Integration:**
211
212 * **AI-annotated findings** fed into **interactive dashboards**.
213 * **Clinicians can adjust, validate, and modify annotations**.
214
215 ----
216
217 === **5. Validation & Real-World Testing** ===
218
219 ==== **Prospective Clinical Study** ====
220
221 * **Multi-center validation** of AI-based **annotations & risk stratifications**.
222 * **Benchmarking against clinician-based diagnoses**.
223 * **Real-world testing** of AI-powered **structured reporting**.
224
225 ==== **Quality Assurance & Explainability** ====
226
227 * **Annotations linked to structured knowledge graphs** for improved transparency.
228 * **Interactive annotation editor** allows clinicians to validate AI outputs.
229
230 ----
231
232 === **6. Collaborative Development** ===
233
234 The project is **open to contributions** from **researchers, clinicians, and developers**.
235
236 **Key tools include:**
237
238 * **Jupyter Notebooks**: For data analysis and pipeline development.
239 ** Example: **probabilistic imputation**
240 * **Wiki Pages**: For documenting methods and results.
241 * **Drive and Bucket**: For sharing code, data, and outputs.
242 * **Collaboration with related projects**:
243 ** Example: **Beyond the hype: AI in dementia – from early risk detection to disease treatment**
244
245 ----
246
247 === **7. Tools and Technologies** ===
248
249 ==== **Programming Languages:** ====
250
251 * **Python** for AI and data processing.
252
253 ==== **Frameworks:** ====
254
255 * **TensorFlow** and **PyTorch** for machine learning.
256 * **Flask** or **FastAPI** for backend services.
257
258 ==== **Visualization:** ====
259
260 * **Plotly** and **Matplotlib** for interactive and static visualizations.
261
262 ==== **EBRAINS Services:** ====
263
264 * **Collaboratory Lab** for running Notebooks.
265 * **Buckets** for storing large datasets.
266
267 ----
268
269 === **Why This Matters** ===
270
271 * The annotation system ensures that AI-generated insights are structured, interpretable, and clinically meaningful.
272 * It enables real-time tracking of disease progression across the three diagnostic axes.
273 * It facilitates integration with electronic health records and decision-support tools, improving AI adoption in clinical workflows.