Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
CREIMBO: Cross-Regional Ensemble Interactions in Multi-view Brain Observations
Authors: Noga Mudrik, Ryan Ly, Oliver Ruebel, Adam Charles
ICLR 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate CREIMBO s ability to recover true components in synthetic data, and uncover meaningful brain dynamics in human high-density electrode recordings, including cross-subject neural mechanisms as well as intervs. intra-region dynamical motifs. Furthermore, using mouse whole-brain recordings, we show CREIMBO s ability to discover dynamical interactions that capture task and behavioral variables and meaningfully align with the biological importance of the brain areas they represent. |
| Researcher Affiliation | Academia | Noga Mudrik Biomedical Engineering, Kavli NDI, CIS The Johns Hopkins University Baltimore, MD, USA. EMAIL Ryan Ly Scientific Data Division Lawrence Berkeley National Laboratory Berkeley, CA, USA. EMAIL Oliver Rübel Scientific Data Division Lawrence Berkeley National Laboratory Berkeley, CA, USA. EMAIL Adam S. Charles Biomedical Engineering, Kavli NDI, CIS The Johns Hopkins University Baltimore, MD, USA. EMAIL |
| Pseudocode | Yes | Algorithm 1 CREIMBO training |
| Open Source Code | Yes | The code is available on Git Hub at this link. |
| Open Datasets | Yes | The human neural data used in this study Kyzar et al. (2024b;a) is publicly available via DANDI Archive at https://dandiarchive.org/dandiset/000469/0.240123.1806. The mice neural data (Chen et al., 2024a; 2023) is publicly available via DANDI Archive at https: //dandiarchive.org/dandiset/000363?search=mesoscale&pos=2. |
| Dataset Splits | No | The paper mentions splitting 'trials' into 'four equal-duration time windows' for feature extraction for a prediction task, but it does not provide explicit training, validation, or test dataset splits for evaluating CREIMBO's core learning algorithm or the logistic regression model in the typical machine learning sense. For example, it does not specify how the trials were split for training and testing the logistic regression model. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used, such as GPU models, CPU specifications, or memory amounts. |
| Software Dependencies | No | The paper mentions several software components like 'Sci Py s implementation of the linear_sum_assignment problem', 'SSM Python package', 'SPGL1', and 'Spiral-Tap package'. However, it does not provide specific version numbers for these software dependencies, which is required for reproducibility. |
| Experiment Setup | Yes | The full set of parameters for the synthetic experiment are available in Table 1. The full set of parameters for the human data experiment are available in Table 2. The full list of parameters is shown in 3, and an example exploration of a random session is shown in (Fig.28,A-C show the firing rate, trial start-end times, and distribution of trial durations, respectively). |