siddhant-rajhans commited on
Commit
51415ee
·
1 Parent(s): ab204cc

Update README with full feature list, live inference docs, deployment guide

Browse files
Files changed (1) hide show
  1. README.md +59 -6
README.md CHANGED
@@ -1,13 +1,19 @@
1
  # CortexLab Dashboard
2
 
3
- Interactive analysis dashboard for [CortexLab](https://github.com/siddhant-rajhans/cortexlab) - multimodal fMRI brain encoding toolkit.
 
 
4
 
5
  ## Pages
6
 
7
- - **Brain Alignment Benchmark** - Compare AI model representations against brain responses (RSA, CKA, Procrustes)
8
- - **Cognitive Load Scorer** - Visualize cognitive demand across visual, auditory, language, and executive dimensions
9
- - **Temporal Dynamics** - Peak latency, lag correlations, sustained vs transient response decomposition
10
- - **ROI Connectivity** - Correlation matrices, network clustering, degree centrality, graph visualization
 
 
 
 
11
 
12
  ## Quick Start
13
 
@@ -16,12 +22,59 @@ pip install -r requirements.txt
16
  streamlit run Home.py
17
  ```
18
 
19
- Runs on **synthetic data** by default - no GPU or real fMRI data required.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
 
21
  ## Links
22
 
23
  - [CortexLab Library](https://github.com/siddhant-rajhans/cortexlab)
24
  - [CortexLab on HuggingFace](https://huggingface.co/SID2000/cortexlab)
 
25
 
26
  ## License
27
 
 
1
  # CortexLab Dashboard
2
 
3
+ Futuristic interactive analysis dashboard for [CortexLab](https://github.com/siddhant-rajhans/cortexlab) - multimodal fMRI brain encoding toolkit built on Meta's TRIBE v2.
4
+
5
+ Glassmorphism dark theme with 3D brain visualization, real-time inference, and research-grade analysis tools.
6
 
7
  ## Pages
8
 
9
+ | Page | Description |
10
+ |---|---|
11
+ | **Brain Alignment Benchmark** | Score AI models against brain responses with RSA, CKA, Procrustes + permutation tests, bootstrap CIs, FDR correction, noise ceiling, RDM visualization |
12
+ | **Cognitive Load Scorer** | Predict cognitive demand across 4 dimensions with confidence bands, comparison mode, per-ROI breakdown |
13
+ | **Temporal Dynamics** | Raw timecourses, peak latency hierarchy, lag correlation with null bands, cross-ROI lag matrix, sustained/transient decomposition |
14
+ | **ROI Connectivity** | Partial correlation, dendrogram, modularity, degree/betweenness centrality, edge weight distribution, network graph |
15
+ | **3D Brain Viewer** | Interactive rotatable fsaverage brain with activation overlays, publication-quality 4-panel views, ROI highlighting, sulcal depth blending |
16
+ | **Live Inference** | Real-time brain prediction from webcam, screen capture, or video file with live-updating 3D brain, cognitive load timeline, and metrics |
17
 
18
  ## Quick Start
19
 
 
22
  streamlit run Home.py
23
  ```
24
 
25
+ Runs on **biologically realistic synthetic data** by default (HRF convolution, modality-specific ROI activation, spatial smoothing). No GPU or real fMRI data required.
26
+
27
+ ## Live Inference (Local Only)
28
+
29
+ For real-time brain prediction from webcam, screen, or video:
30
+
31
+ ```bash
32
+ # Install optional capture dependencies
33
+ pip install opencv-python mss Pillow
34
+
35
+ # For real TRIBE v2 inference (needs GPU):
36
+ pip install -e ../cortexlab[analysis]
37
+
38
+ # Start dashboard
39
+ streamlit run Home.py
40
+ # Navigate to Live Inference page
41
+ ```
42
+
43
+ Without CortexLab installed, live inference runs in **simulation mode** - predictions are generated from image statistics (brightness, contrast, color) mapped to brain ROIs.
44
+
45
+ ## Features
46
+
47
+ - **Futuristic UI**: Glassmorphism dark theme, neon accents, gradient headings, glowing metric cards, animated borders
48
+ - **3D Brain Hero**: Rotatable fsaverage brain mesh on the home page
49
+ - **Biologically Realistic Data**: HRF-convolved synthetic data with modality-specific activation patterns
50
+ - **Statistical Rigor**: Permutation tests, bootstrap CIs, FDR correction, noise ceiling estimation
51
+ - **Cross-Page State**: ROI selections carry between pages, shared session predictions
52
+ - **File Upload**: Upload .npy predictions from real CortexLab runs
53
+ - **CSV/JSON Export**: Download results from every analysis page
54
+ - **Methodology Docs**: Every page has an expandable methodology section with references
55
+
56
+ ## Deployment
57
+
58
+ ### HuggingFace Spaces
59
+
60
+ Live at: [huggingface.co/spaces/SID2000/cortexlab-dashboard](https://huggingface.co/spaces/SID2000/cortexlab-dashboard)
61
+
62
+ Docker-based deployment. Live inference page shows simulation mode (no webcam/GPU access in Spaces).
63
+
64
+ ### Local
65
+
66
+ ```bash
67
+ git clone https://github.com/siddhant-rajhans/cortexlab-dashboard.git
68
+ cd cortexlab-dashboard
69
+ pip install -r requirements.txt
70
+ streamlit run Home.py
71
+ ```
72
 
73
  ## Links
74
 
75
  - [CortexLab Library](https://github.com/siddhant-rajhans/cortexlab)
76
  - [CortexLab on HuggingFace](https://huggingface.co/SID2000/cortexlab)
77
+ - [Live Demo](https://huggingface.co/spaces/SID2000/cortexlab-dashboard)
78
 
79
  ## License
80