Update README.md
Browse files
README.md
CHANGED
|
@@ -21,7 +21,7 @@ tags:
|
|
| 21 |
|
| 22 |
This repository contains **subject-specific NSD voxelwise encoding-model fits and the variance-partitioning fits reported in the associated study** from:
|
| 23 |
|
| 24 |
-
**Diaz, D. M., & Henderson, M. M. (2026).
|
| 25 |
|
| 26 |
**DOI:** [10.32470/0416gfsq](https://doi.org/10.32470/0416gfsq)<br>
|
| 27 |
**arXiv:** [2607.19316](https://arxiv.org/abs/2607.19316)<br>
|
|
@@ -67,7 +67,8 @@ The four pretrained models were trained on complementary versions of the same so
|
|
| 67 |
|
| 68 |
| Component | Status |
|
| 69 |
| --- | --- |
|
| 70 |
-
|
|
|
|
|
| 71 |
| Variance-partitioning fits reported in the paper | Available |
|
| 72 |
| Model card | Available |
|
| 73 |
| Feature-extraction / model-fitting code | Forthcoming |
|
|
@@ -84,14 +85,18 @@ baseline/
|
|
| 84 |
fovea-gaze/
|
| 85 |
periph/
|
| 86 |
periph-nf/
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
variance-partitioning/
|
| 88 |
βββ fovea-gaze-vs-periph/
|
| 89 |
βββ periph-vs-periph-nf/
|
| 90 |
```
|
| 91 |
|
| 92 |
-
Each
|
| 93 |
|
| 94 |
-
|
| 95 |
|
| 96 |
```text
|
| 97 |
fovea-gaze/
|
|
@@ -99,6 +104,14 @@ fovea-gaze/
|
|
| 99 |
βββ NSD_S2_resnet18-Fovea-Gaze_concat.npy
|
| 100 |
βββ ...
|
| 101 |
βββ NSD_S8_resnet18-Fovea-Gaze_concat.npy
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 102 |
```
|
| 103 |
|
| 104 |
## Architecture
|
|
@@ -117,13 +130,13 @@ Voxelwise encoding models were fit using fMRI responses from the [**Natural Scen
|
|
| 117 |
|
| 118 |
For each participant, the **1,000 images shared across all NSD participants** were reserved as the final held-out evaluation set. The remaining approximately **9,000 participant-specific images** were used for model fitting and regularization selection. Within this fitting set, a nested holdout partition was used to select the ridge penalty independently for each voxel.
|
| 119 |
|
| 120 |
-
The same intact `224 Γ 224` NSD images were presented to each of the
|
| 121 |
|
| 122 |
Encoding performance was evaluated on the held-out 1,000-image set using voxelwise coefficient of determination (`RΒ²`) and Pearson correlation (`corr`). These held-out metrics are stored directly in each released encoding-model fit.
|
| 123 |
|
| 124 |
## Encoding-Model Procedure
|
| 125 |
|
| 126 |
-
The encoding models were fit separately for each **NSD subject (S1βS8)** and each
|
| 127 |
|
| 128 |
NSD stimuli were presented to the pretrained ResNet-18 models as intact `224 Γ 224` images. The Baseline, Fovea-Gaze, Periph, and Periph-NF transformations were used during SimCLR pretraining and were **not reapplied to NSD images** during encoding-model evaluation.
|
| 129 |
|
|
@@ -166,7 +179,7 @@ Performance on the final held-out set was quantified with voxelwise coefficient
|
|
| 166 |
Each `.npy` file contains a Python dictionary with the fitted encoding model and associated metadata:
|
| 167 |
|
| 168 |
- `subject` β NSD subject number
|
| 169 |
-
- `model` β visual model
|
| 170 |
- `features_file_list` β feature files used for model fitting
|
| 171 |
- `lambdas` β candidate ridge penalties
|
| 172 |
- `weights` β fitted voxelwise ridge-regression weights
|
|
@@ -190,6 +203,11 @@ The released `.npy` files retain the internal model identifiers used during the
|
|
| 190 |
| Fovea-Gaze | `resnet18-FoveaGaze` |
|
| 191 |
| Periph | `resnet18-PeriphNonTTM` |
|
| 192 |
| Periph-NF | `resnet18-PeriphTTM` |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 193 |
|
| 194 |
These internal identifiers are preserved in fields such as `model`, `model1`, `model2`, and the keys of the variance-partitioning dictionaries. They reflect the naming convention used during model development and do not indicate additional model conditions.
|
| 195 |
|
|
@@ -304,7 +322,9 @@ voxel_nc = fit["voxel_nc"]
|
|
| 304 |
|
| 305 |
## Scope of the Released Artifacts
|
| 306 |
|
| 307 |
-
The repository contains **
|
|
|
|
|
|
|
| 308 |
|
| 309 |
The released files contain fitted voxelwise weights, held-out prediction metrics, ridge-penalty information, and voxel metadata used in the study. The variance-partitioning artifacts additionally contain the model-A-only, model-B-only, and combined fits used to compute unique variance.
|
| 310 |
|
|
|
|
| 21 |
|
| 22 |
This repository contains **subject-specific NSD voxelwise encoding-model fits and the variance-partitioning fits reported in the associated study** from:
|
| 23 |
|
| 24 |
+
**Diaz, D. M., & Henderson, M. M. (2026). *Eccentricity-Constrained CNN Training Reveals Adaptive Information Coding Around the Visual Field.* Proceedings of the 9th Conference on Cognitive Computational Neuroscience.**
|
| 25 |
|
| 26 |
**DOI:** [10.32470/0416gfsq](https://doi.org/10.32470/0416gfsq)<br>
|
| 27 |
**arXiv:** [2607.19316](https://arxiv.org/abs/2607.19316)<br>
|
|
|
|
| 67 |
|
| 68 |
| Component | Status |
|
| 69 |
| --- | --- |
|
| 70 |
+
| VEDB subject-specific encoding-model fits | Available |
|
| 71 |
+
| Reference-model encoding fits (STL-10, ImageNet-100, ImageNet-1K) | Available |
|
| 72 |
| Variance-partitioning fits reported in the paper | Available |
|
| 73 |
| Model card | Available |
|
| 74 |
| Feature-extraction / model-fitting code | Forthcoming |
|
|
|
|
| 85 |
fovea-gaze/
|
| 86 |
periph/
|
| 87 |
periph-nf/
|
| 88 |
+
reference-models/
|
| 89 |
+
βββ stl10/
|
| 90 |
+
βββ imagenet-100/
|
| 91 |
+
βββ imagenet-1k/
|
| 92 |
variance-partitioning/
|
| 93 |
βββ fovea-gaze-vs-periph/
|
| 94 |
βββ periph-vs-periph-nf/
|
| 95 |
```
|
| 96 |
|
| 97 |
+
Each encoding-model folder contains voxelwise fits for **NSD subjects S1βS8**. The four VEDB folders contain fits derived from the Baseline, Fovea-Gaze, Periph, and Periph-NF models. The `reference-models/` directory contains the corresponding fits derived from the STL-10, ImageNet-100, and ImageNet-1K reference models. The `variance-partitioning/` directory contains the two model comparisons reported in Figure 4CβD of the paper, with one fit for each of the same eight NSD participants.
|
| 98 |
|
| 99 |
+
Examples:
|
| 100 |
|
| 101 |
```text
|
| 102 |
fovea-gaze/
|
|
|
|
| 104 |
βββ NSD_S2_resnet18-Fovea-Gaze_concat.npy
|
| 105 |
βββ ...
|
| 106 |
βββ NSD_S8_resnet18-Fovea-Gaze_concat.npy
|
| 107 |
+
|
| 108 |
+
reference-models/
|
| 109 |
+
βββ stl10/
|
| 110 |
+
β βββ NSD_S1_resnet18-simclr-stl10_concat.npy
|
| 111 |
+
βββ imagenet-100/
|
| 112 |
+
β βββ NSD_S1_resnet18-simclr-imagenet100_concat.npy
|
| 113 |
+
βββ imagenet-1k/
|
| 114 |
+
βββ NSD_S1_resnet18-simclr-imagenet1k_concat.npy
|
| 115 |
```
|
| 116 |
|
| 117 |
## Architecture
|
|
|
|
| 130 |
|
| 131 |
For each participant, the **1,000 images shared across all NSD participants** were reserved as the final held-out evaluation set. The remaining approximately **9,000 participant-specific images** were used for model fitting and regularization selection. Within this fitting set, a nested holdout partition was used to select the ridge penalty independently for each voxel.
|
| 132 |
|
| 133 |
+
The same intact `224 Γ 224` NSD images were presented to each of the seven pretrained ResNet-18 models. The Baseline, Fovea-Gaze, Periph, and Periph-NF transformations were applied only during VEDB SimCLR pretraining and were **not reapplied to NSD stimuli** during feature extraction. The same NSD feature-extraction procedure was used for the STL-10, ImageNet-100, and ImageNet-1K reference models.
|
| 134 |
|
| 135 |
Encoding performance was evaluated on the held-out 1,000-image set using voxelwise coefficient of determination (`RΒ²`) and Pearson correlation (`corr`). These held-out metrics are stored directly in each released encoding-model fit.
|
| 136 |
|
| 137 |
## Encoding-Model Procedure
|
| 138 |
|
| 139 |
+
The encoding models were fit separately for each **NSD subject (S1βS8)** and each of the seven pretrained visual models.
|
| 140 |
|
| 141 |
NSD stimuli were presented to the pretrained ResNet-18 models as intact `224 Γ 224` images. The Baseline, Fovea-Gaze, Periph, and Periph-NF transformations were used during SimCLR pretraining and were **not reapplied to NSD images** during encoding-model evaluation.
|
| 142 |
|
|
|
|
| 179 |
Each `.npy` file contains a Python dictionary with the fitted encoding model and associated metadata:
|
| 180 |
|
| 181 |
- `subject` β NSD subject number
|
| 182 |
+
- `model` β visual model identifier
|
| 183 |
- `features_file_list` β feature files used for model fitting
|
| 184 |
- `lambdas` β candidate ridge penalties
|
| 185 |
- `weights` β fitted voxelwise ridge-regression weights
|
|
|
|
| 203 |
| Fovea-Gaze | `resnet18-FoveaGaze` |
|
| 204 |
| Periph | `resnet18-PeriphNonTTM` |
|
| 205 |
| Periph-NF | `resnet18-PeriphTTM` |
|
| 206 |
+
| STL-10 | `resnet18-pretrained-simclr` |
|
| 207 |
+
| ImageNet-100 | `resnet18-simclr-imgnet100` |
|
| 208 |
+
| ImageNet-1K | `resnet18-simclr-imgnet1k` |
|
| 209 |
+
|
| 210 |
+
The internal identifier `resnet18-pretrained-simclr` corresponds to the STL-10 SimCLR ResNet-18 checkpoint obtained from the external [Spijkervet/SimCLR](https://github.com/Spijkervet/SimCLR) release.
|
| 211 |
|
| 212 |
These internal identifiers are preserved in fields such as `model`, `model1`, `model2`, and the keys of the variance-partitioning dictionaries. They reflect the naming convention used during model development and do not indicate additional model conditions.
|
| 213 |
|
|
|
|
| 322 |
|
| 323 |
## Scope of the Released Artifacts
|
| 324 |
|
| 325 |
+
The repository contains **56 subject-specific encoding-model fits** across seven pretrained visual models: **32 fits from the four VEDB-pretrained models** (4 models Γ 8 NSD participants) and **24 fits from the three non-egocentric reference models** (3 models Γ 8 NSD participants).
|
| 326 |
+
|
| 327 |
+
The repository additionally contains **16 variance-partitioning fits** corresponding to the two variance-partition analyses reported in the paper (2 model comparisons Γ 8 participants), for a total of **72 fitted `.npy` artifacts**.
|
| 328 |
|
| 329 |
The released files contain fitted voxelwise weights, held-out prediction metrics, ridge-penalty information, and voxel metadata used in the study. The variance-partitioning artifacts additionally contain the model-A-only, model-B-only, and combined fits used to compute unique variance.
|
| 330 |
|