YASS: Yet Another Spike Sorter
Authors: Jin Hyung Lee, David E. Carlson, Hooshmand Shokri Razaghi, Weichi Yao, Georges A. Goetz, Espen Hagen, Eleanor Batty, E.J. Chichilnisky, Gaute T. Einevoll, Liam Paninski
NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The proposed methods improve on the state-of-the-art in terms of accuracy and stability on both real and biophysically-realistic simulated MEA data. and We evaluate performance to compare several algorithms (detailed in Section 3.1) to our proposed methodology on both synthetic (Section 3.2) and real (Section 3.3) dense MEA recordings. |
| Researcher Affiliation | Academia | 1Columbia University, 2Duke University, 3Stanford University, 4University of Oslo, 5Norwegian University of Life Sciences |
| Pseudocode | Yes | Pseudocode for the flow of the pipeline can be found in Algorithm 1. and Algorithm 1 Pseudocode for the complete proposed pipeline. |
| Open Source Code | Yes | Lastly, YASS is available online at https://github.com/paninski-lab/yass |
| Open Datasets | Yes | First, we used the biophysics-based spike activity generator Vi SAPy [18] to generate multiple 30-channel datasets... and To examine real data, we focused on 30 minutes of extracellular recordings of the peripheral primate retina, obtained ex-vivo using a high-density 512-channel recording array [30]. |
| Dataset Splits | Yes | For each synthetic dataset we evaluate the ability to capture ground truth in addition to the per-cluster stability metrics. and The detection network was trained on the ground truth from a low signal-to-noise level recording. and We evaluate the performance of YASS and competing algorithms using 4 distinct sets of 49 spatially contiguous electrodes. and A gold standard" sort was constructed for this dataset by extensive hand validation of automated techniques, as detailed in Supplemental Section H. |
| Hardware Specification | Yes | The CPU runtime was calculated on a single core of a six-core i7 machine with 32GB of RAM. GPU runtime is given from a Nvidia Titan X within the same machine. |
| Software Dependencies | No | The paper mentions software used for comparison (Kilosort, Spyking Circus, Mountain Sort) and frameworks like Dirichlet Process Gaussian Mixture Model, but does not list specific versions of programming languages, libraries, or other software dependencies used for their own implementation of YASS. |
| Experiment Setup | No | The paper describes the overall pipeline and algorithmic choices in detail (e.g., use of neural networks for detection, DP-GMM for clustering), but it does not specify concrete hyperparameter values (e.g., learning rate, batch size, number of epochs) or system-level training settings for its models. |