Deep Learning Models for Multi-Region and Multi-Subject Two-Photon Calcium Imaging Data
Bibliographic record
Abstract
Calcium imaging is an invasive brain imaging method that detects the influx of calcium ions during neural activities such as action potentials, NMDA spikes, etc.It's a popular technique with advantages such as high spatial resolution, preserving spatial compositions, etc.However, due to the noisy and sparse nature of the fluorescent signals, a lengthy and complex preprocessing pipeline is usually required before calcium imaging data can be used for downstream purposes.In recent years, there have been many machine learning and deep learning models working to obtain high-quality reconstructions of neural signals and downstream decoding, such as decoding animal behaviours.However, there are few techniques for conducting analysis on raw calcium fluorescence signals, and there exists a gap for deep learning models that can extract a unified, holistic neural representation from multi-region, multi-subject and multi-task calcium imaging data.We constructed and tested two end-to-end models designed to ingest calcium raw data, one recurrent network-based (2P-DPC) and the other transformer-based (CaPOYO).We employed two multi-region, multi-session, multi-subject and multi-task mouse visual cortex two-photon calcium imaging datasets, and we tested the extracted representation by decoding different behaviours and stimuli.We found the 2P-DPC model to perform less than satisfactory for i downstream decoding tasks.We also found that the transformer-based CaPOYO model is robust and versatile.It can leverage the complexity and noise of calcium data, not only achieving much higher decoding performance than baselines but also being capable of transferring to unseen subjects, sessions, transgenic cre lines, brain areas and even completely unseen datasets.I would like to thank Dr. Blake Richards for giving me the opportunity to pursue my Master's degree under his supervision.Thank you for being an extraordinary mentor, whose vast knowledge, kindness, and understanding have guided me throughout my Master's journey.Your work ethic, dedication to science and the scientific community, and commitment to fostering a healthy and mutually respectful team environment have greatly inspired me.I will carry these values with me in my professional journey.I would like to thank Dr. Shahab Bakhtiari and Dr. J. Quinn Lee for their expertise, patience, and time in supporting me with the 2P-DPC model.A special
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".