Updated README.
Browse files
README.md
CHANGED
|
@@ -33,9 +33,14 @@ The data is organized into a highly efficient, indexed HDF5 format, allowing for
|
|
| 33 |
| Sessions | 14 sessions of registered neural activity |
|
| 34 |
| Stimuli | 3 categories (`Clip`, `Monet2`, `Trippy`) identified by unique condition hashes |
|
| 35 |
| Neural Data | Calcium traces from thousands of neurons across V1, LM, RL, AL |
|
| 36 |
-
|
|
|
|
|
| 37 |
| Eye Tracking | Pupil center coordinates (x, y) for gaze analysis |
|
| 38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
---
|
| 40 |
|
| 41 |
## 🗂️ Internal HDF5 Structure
|
|
|
|
| 33 |
| Sessions | 14 sessions of registered neural activity |
|
| 34 |
| Stimuli | 3 categories (`Clip`, `Monet2`, `Trippy`) identified by unique condition hashes |
|
| 35 |
| Neural Data | Calcium traces from thousands of neurons across V1, LM, RL, AL |
|
| 36 |
+
| Treadmill | Running speed (cm/s) synchronized to stimulus frames |
|
| 37 |
+
| Pupil Size | Major and minor radius of the fitted pupil ellipse (in pixels) |
|
| 38 |
| Eye Tracking | Pupil center coordinates (x, y) for gaze analysis |
|
| 39 |
|
| 40 |
+
### Temporal alignment & sampling
|
| 41 |
+
|
| 42 |
+
All data streams — neural responses, pupil, and treadmill — were independently interpolated to 30 Hz to match the stimulus frame rate, aligning every signal to a common stimulus clock. The full 30 Hz timeseries was then uniformly downsampled by a factor of 4, yielding a final sampling rate of **7.5 Hz** for all stored signals. As a result, `responses`, `treadmill`, `pupil`, and `stim_times` share an identical time axis within every trial, with each sample corresponding to a stimulus frame spaced ~133 ms apart.
|
| 43 |
+
|
| 44 |
---
|
| 45 |
|
| 46 |
## 🗂️ Internal HDF5 Structure
|