Research map / Probabilistic and scientific modeling / Variational and generative sampling
Fast amortized inference of neural activity from calcium imaging data with variational autoencoders
Artur Speiser, Jinyao Yan, Evan W Archer, Lars Buesing, Srinivas C. Turaga, Jakob H Macke
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use a recurrent, non-factorized posterior instead of a factorized posterior (ablation) | DeepSpike on SCDF simulated data · Simulated SCDF data | correlation | No clear effect |
| Use a recurrent, non-factorized posterior instead of a factorized posterior (ablation) | DeepSpike on SCDF simulated data · Simulated SCDF data | posterior probability of true spike trains | Improved |
| Use a recurrent, non-factorized posterior instead of a factorized posterior (ablation) | DeepSpike · SCF-Sim. | correlation: 0.88 ± 0.01 -> 0.89 ± 0.01 | Improved |
| Use a recurrent, non-factorized posterior instead of a factorized posterior (ablation) | DeepSpike · GCaMP6s | correlation: 0.74 ± 0.02 -> 0.72 ± 0.02 | Hurt |
| Use a recurrent, non-factorized posterior instead of a factorized posterior (ablation) | DeepSpike · GCaMP6f | correlation: 0.74 ± 0.02 -> 0.73 ± 0.02 | Hurt |
| Amortize inference by sharing one recognition model across cells (ablation) | DS-NF · 30 held-out simulated SCDF cells | performance | Hurt |
| Use amortized DS-F rather than OASIS | Spike inference on 10K time-step traces · CPU | inference time: 0.07 s vs 0.05 s | No clear effect |
| Use amortized DS-NF rather than MLspike | Spike inference on 10K time-step traces · CPU | inference time: 9.2 s vs 4.6 s | Improved |
| Use DS-F-DEN rather than Foopsi-RR | Dendritic spike inference · Dendritic dataset, Soma | correlation: 0.66 ± 0.02 -> 0.84 ± 0.01 | Improved |
| Use DS-F-DEN rather than Foopsi-RR | Dendritic spike inference · Dendritic dataset, Spine | correlation: 0.60 ± 0.01 -> 0.78 ± 0.01 | Improved |
Closest papers by meaning
- Fast Active Set Methods for Online Spike Inference from Calcium Imaging (NeurIPS 2016 · similarity 0.90)
- SpikeSSL: A Universal Spike Inference Framework with Dynamics-Informed State-Space Layers (NeurIPS 2026 · similarity 0.89)
- Model-based Bayesian inference of neural activity and connectivity from all-optical interrogation of a neural circuit (NeurIPS 2017 · similarity 0.89)
- Scalable Spike Source Localization in Extracellular Recordings using Amortized Variational Inference (NeurIPS 2019 · similarity 0.88)
- Brain-like Variational Inference (NeurIPS 2025 · similarity 0.86)
- Multi-modal Gaussian Process Variational Autoencoders for Neural and Behavioral Data (ICLR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
