ronniebasak commited on
Commit
a40fd1c
·
verified ·
1 Parent(s): f9b2b55

Upload folder using huggingface_hub

Browse files
README.md ADDED
@@ -0,0 +1,98 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ task_categories:
4
+ - tabular-regression
5
+ language:
6
+ - en
7
+ tags:
8
+ - neuroscience
9
+ - computational-neuroscience
10
+ - acetylcholine
11
+ - neuromodulation
12
+ - sim-to-real
13
+ - jaxley
14
+ - allen-brain-observatory
15
+ - neuropixels
16
+ - biophysical-simulation
17
+ size_categories:
18
+ - 10K<n<100K
19
+ ---
20
+
21
+ # ACh Cortical Circuits: Cross-Cortical Transfer Dataset
22
+
23
+ 15,000 simulated cortical microcircuits with acetylcholine (ACh) modulation + 2,678 real mouse V1 recording epochs from the Allen Brain Observatory.
24
+
25
+ ## Dataset Description
26
+
27
+ This dataset accompanies the paper **"Cross-Cortical Transfer of Cholinergic Modulation Principles via Biophysical Simulation Surrogates"** (CAISc 2026).
28
+
29
+ ### Simulated Circuits (15,000 total)
30
+
31
+ Three mechanistic variants of ACh modulation, each with 5,000 circuits × 11 ACh levels = 55,000 simulations per variant:
32
+
33
+ | Variant | ACh Mechanism | File |
34
+ |---------|--------------|------|
35
+ | **v14a** | Synaptic suppression only (E2 dose-response) | `v14a_circuits.tar.gz` |
36
+ | **v14b** | Depolarization only (OU mean shift) | `v14b_circuits.tar.gz` |
37
+ | **v14c** | Both mechanisms combined | `v14c_circuits.tar.gz` |
38
+
39
+ **Parameters source:** Ramaswamy, Colangelo & Markram (2018), *Frontiers in Neural Circuits* — rat somatosensory cortex (S1).
40
+
41
+ **Simulator:** [Jaxley](https://github.com/jaxleyverse/jaxley) (JAX-based, GPU-accelerated)
42
+
43
+ Each circuit JSON contains:
44
+ - Circuit parameters (n_neurons, connectivity, synaptic conductances, OU noise params)
45
+ - ACh level (0.0 to 1.0 in 0.1 steps)
46
+ - 11 population statistics (firing rate, CV ISI, Fano factor, synchrony index, pairwise correlation, peak frequency, spectral power, etc.)
47
+
48
+ ### Allen Brain Observatory Epochs (2,678 total)
49
+
50
+ Real mouse V1 (area VISp) neural recordings from the Allen Brain Observatory Visual Coding Neuropixels dataset:
51
+ - 2,325 running epochs + 353 stationary epochs
52
+ - 44 sessions, 16 with both behavioral states
53
+ - Quality filtered (presence_ratio > 0.95, ISI violations < 0.5)
54
+ - Same 11 population statistics as simulated data
55
+
56
+ File: `allen_epochs.tar.gz`
57
+
58
+ ### Model Checkpoints
59
+
60
+ 12 trained surrogate models (3 architectures × 3 ACh variants + 3 baselines):
61
+ - CircuitTransformer (222K params)
62
+ - CircuitMLP (5.6K params)
63
+ - XGBoost
64
+
65
+ File: `checkpoints.tar.gz`
66
+
67
+ ## Key Results
68
+
69
+ | Statistic | Model S (syn) | Model D (depol) | Model SD (both) |
70
+ |-----------|:---:|:---:|:---:|
71
+ | Pairwise correlation | **75.0%** | 25.0% | 25.0% |
72
+ | Mean firing rate | 12.5% | **87.5%** | **87.5%** |
73
+ | Spectral power | **68.8%** | **68.8%** | **68.8%** |
74
+
75
+ **Key finding:** Synaptic suppression captures correlation structure; depolarization captures rate modulation. The mechanisms are dissociable in sim-to-real transfer.
76
+
77
+ ## Citation
78
+
79
+ ```bibtex
80
+ @inproceedings{neuropeek2026crosstransfer,
81
+ title={Cross-Cortical Transfer of Cholinergic Modulation Principles via Biophysical Simulation Surrogates},
82
+ author={NeuroPeek AI and Basak, Sohan},
83
+ booktitle={1st Conference for AI Scientists (CAISc 2026)},
84
+ year={2026}
85
+ }
86
+ ```
87
+
88
+ ## License
89
+
90
+ MIT
91
+
92
+ ## Acknowledgments
93
+
94
+ - Simulation parameters from [Ramaswamy et al. 2018](https://www.frontiersin.org/articles/10.3389/fncir.2018.00077/full)
95
+ - Real data from [Allen Brain Observatory](https://portal.brain-map.org/)
96
+ - Simulation engine: [Jaxley](https://github.com/jaxleyverse/jaxley) (Deistler et al., Nature Methods 2025)
97
+ - Compute: [Modal](https://modal.com/) cloud GPUs (~$230 total)
98
+ - Built with NeuroPeek AI Agents
data/allen_epochs.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d99eee4a8befbe52a3e92dc86cee7fc2822c498d36168eef6212b8c05cc8ae5d
3
+ size 271882
data/checkpoints.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1260837f4ffbf6237dd33f4469afb202ea30e06522a0c27d2193aad8c72fb6a0
3
+ size 280007
data/v14a_circuits.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f8d4ee0852161529a538c5b62c5f1295ad1aac4e5cf156f226ce28ea71e7bda7
3
+ size 2654515
data/v14b_circuits.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb417cecfc244f04a73afe303bed65a4c82b39a9c0b819e24c814a4bdbeeaee4
3
+ size 2388494
data/v14c_circuits.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:013e6a054306b3413f5740a0265f3e752661694efff4405e07938c7e3f14e676
3
+ size 2461428